AI Camerais in the Security-Surveillance Technology Industry

AI Camera

aicamera

Artificial Intelligence (AI) is transforming industries and creating new opportunities at the speed of light. From AI-generated video to its integration into major companies' products, the potential for to drive innovation and transformation is truly vast. In this article, we’ve gathered top AI trends by analyzing forecasts from leading technology companies and research firms, as well as our work with AI projects. AI By the way - if you’d like to stay ahead of the AI curve, discover cutting-edge projects, and expand your skills through free mini-courses, sign up for my weekly Each edition features the latest AI breakthroughs, tools, and tips to help you apply this transformative technology. It’s absolutely free! .’ newsletter ‘AI Hunters For now, let’s explore the top AI trends and their implications for the future. 1. AI Video Unlocks Creativity in Film, Ads and Games AI-generated video is gaining traction in the film and entertainment industry, with companies like Warner Bros. leveraging AI to analyze movie trailers and predict box office success. More than that, AI is being used in advertising, enabling personalized ad creation based on user data. AI is also finding its way into video game development, with companies like Electronic Arts using AI to generate game assets and environments. Generative AI is already democratizing content creation and changing the power and economics of video streaming. 2. In-Demand AI Skills Propel the Jobs Market The demand for AI skills is on the rise, with LinkedIn's 2020 Emerging Jobs Report identifying AI specialists as the top emerging jobs in the United States, experiencing a 74% annual growth rate. These skills are sought after across diverse industries, including healthcare, finance, and retail. Furthermore, AI skills are becoming increasingly important for marketing professionals, given the prevalence of AI-generated content in content marketing strategies. 3. Tech Heavyweights Pave the Way Leading companies such as Meta, Google, Amazon, Apple, and IBM are at the forefront of integrating AI into their products, spanning from virtual assistants to image recognition software. Moreover, AI is being integrated into customer relationship management (CRM) systems by companies like Salesforce and Oracle, enhancing customer data analysis and improving experiences. AI is currently finding its way into Internet of Things (IoT) devices, with companies like Intel using AI to enhance device performance and user experience. 4. All Spotlight on Virtual Influencers The emergence of AI-generated influencers on social media platforms like Instagram is reshaping digital marketing. Companies like Brud are creating virtual influencers such as Lil Miquela. AI-generated content is also making inroads into journalism, with The Washington Post using AI to write news articles. Additionally, e-commerce players like Amazon are leveraging AI to generate product descriptions and recommendations. AI-generated content is forever changing the way marketing is done, allowing for personalized content at a scale that was hardly conceivable before. 5. AI Touches Everyday Lives With the development of AI-driven cars by companies like Tesla and Google, AI's impact actively strikes root in the physical world. Wearable technology, including smartwatches and fitness trackers, is harnessing AI to enhance user experience and provide more accurate data. AI is also being integrated into smart homes, with companies like Amazon incorporating AI into the Alexa virtual assistant system. 6. Tailored AI Optimizes Impact Customization of enterprise AI is proving to be a game-changer, offering significant cost savings and improved accuracy and efficiency. Companies like IBM report that customized AI solutions can reduce costs by up to 50%. Customization can also lead to enhanced customer experiences, as personalized content and recommendations can be tailored to individual user preferences and behaviors. 7. Multimodal AI: A Universal Language In healthcare, multimodal AI is enhancing patient outcomes through AI-powered medical imaging and diagnosis systems. Plus, AI natural language processing systems are becoming adept at understanding and responding to both text and speech. In computer vision, AI applications can analyze and interpret images and videos, opening new frontiers in various industries. 8. Regulators are Fostering Responsible AI Innovation Regulating AI is a growing area of interest, with the European Union establishing guidelines for the ethical use of AI that emphasize transparency, accountability, and data protection. The United States has also set up a National AI Initiative Office to coordinate AI research and development and address ethnic concerns. Companies like IBM and Salesforce are already developing AI governance frameworks as the first major step towards the future, where responsible AI use will be ensured. 9. AI Gets Full Tech Stack Integration As AI is being harnessed across various applications, including chatbots, virtual assistants, and recommendation systems, it is playing a pivotal role in improving cybersecurity through real-time response to cyber threats. Companies like Bank of America are already leveraging AI in finance to automate financial processes and enhance decision-making. 10. Deepfake Brings Discomfort A deepfake image of an explosion near the Pentagon, purportedly related to the arrest of former President Donald Trump, caused uproar on social media. In 2023, deepfake technology became a pervasive tool, reshaping public discourse and causing confusion through exploitative uses, including fraud and manipulation of public figures — like Indian PM Modi, which already highlighted serious threats to democracy, privacy, and global harmony. As the 2024 US presidential elections approach, deepfakes are a growing concern for political campaigns, as it is increasingly being used to spread disinformation and propaganda. Wrapping up, AI's pervasive influence is reshaping industries and creating new opportunities. The potential for AI to drive innovation and transformation is vast, and its impact is already being felt across numerous sectors. However, as AI continues to evolve, the need for responsible regulation and governance becomes increasingly paramount. The future of AI is bright, and its potential to improve living quality and create real sources of value is immense. P.S. Check out my previous articles on AI at HackerNoon: How to Build Your Personal GPTs: From Zero to AI Hero GPT-4 Turbo: The Most Monumental Update Since ChatGPT's Debut! Essential Insights from 'State of AI 2023' GPT-4V Unveiled: From Detecting Emotions to Ordering Food - You Won't Believe What Else It Can Do! ChatGPT Now Speaks, Listens, and Understands: All You Need to Know Fine-Tuning for GPT-3.5 Turbo: AI Game Changer The Rise of AI-Generated Films: Lights, Camera, Algorithm! NFT Marketing Guide - The Most Complete and Detailed Playbook 2023

Top 10 AI Trends of 2024: How AI Transforms Everything 

From self-driving cars and facial recognition to AI surveillance and GANs, computer vision tech has been the poster child of the AI industry in recent years. With such a collaborative global data science community, the advancements have come both from research teams, big tech, and computer vision startups alike. For all startups, Product Hunt has been a great platform to help get the word out there about your projects and get feedback from the online community. In this article, we’ll cover some of the most-upvoted startups on Product Hunt that have created computer vision software. Most-Upvoted Computer Vision Startups on Product Hunt We looked at the pages on Product Hunt and have compiled a list of the products that have historically received the most upvotes. Many of these products have been featured on large publications and some of these apps have received upwards of 50,000  downloads. computer vision You can check out the list of the following companies Product hunt pages . here 1. Vay Vay is a fitness coach app that uses human motion analysis to track overall physical performance. With the app, users can receive real-time feedback on their exercises, as if they had their own personal trainer. Vay uses computer vision algorithms to track and digitize human movements, transforming movement patterns into data that can be analyzed in a variety of ways. The product has received 80 upvotes on Product Hunt and you can read more about it on . Vay’s website 2. Origin From Shaper Tools, Origin is a computer-vision-powered woodworking tool that was named one of the top tools in 2020 by Popular Woodworking Magazine. Origin gives craftsmen AI-powered automatic cut correction and precise depth control with the help of a specialized ShaperTape that you apply to your workstation. Image via shapertools.com The ShaperTape is used to help the computer vision model detect your workstation and track your movement around it. has received 130 upvotes on Product Hunt. Origin 3. Camma.ai Camma.ai is a computer vision startup that has created an augmented reality note-taking software that allows you to place notes on real objects through your iPhone’s camera. Users take a photo of an object and type a note in their iPhone. When the Camma app sees the object, your note will pop up. Unfortunately for Android users, this app is only available on iOS. Camma.ai has received 143 upvotes and can be downloaded via the . App Store 4. Roboflow Organize Roboflow is a machine learning platform that aims to streamline data preprocessing and the training of computer vision models. With a multitude of annotation formats and file types, it can take an exorbitant amount of time to get images ready for training. Roboflow aims to be the essential architecture between the annotation and training phases that computer vision startups desperately need to improve the and lessen the time spent writing boilerplate code. data annotation process Roboflow has received 215 upvotes and is available for a . free trial 5. Logo Crunch From brandmark.io, Logo Crunch is a computer vision tool that optimizes logos to make them clear and legible at lower resolutions. The tool is suitable for those with no graphic design experience and is completely free to use. Logo Crunch has received 283 upvotes on Product Hunt. You can try it for free here. 6. Nanit For parents everywhere, Nanit has become an essential part of their daily lives. Nanit is a computer vision startup that produces advanced baby monitor cameras which you can view via a mobile app. Nanit’s AI camera can track your baby’s sleeping patterns and records them via the app. The app can give you live alerts telling you when your baby has woken up and if they were able to go back to sleep on their own. has received 362 upvotes on Product Hunt. Nanit 7. MDacne MDacne is a godsend for those prone to acne. With the help of professional dermatologists, MD acne created a computer vision model to scan images of the user’s face to provide a full analysis of their acne and skin condition within just a few seconds. This step is completely free of charge. The computer vision model powering the app has been trained on around 1 million photos and analyzes skin color around the affected area, as well as the size and number of pimples. Next, if users wish, they can order an acne treatment kit from the same company, which is customized to suit their skin condition determined by the app. is FDA approved and has received 428 upvotes on Product Hunt. MDacne 8. Automatic GIF Creator – gifs.com From gifs.com, Automatic GIF Creator is a GIF generator that uses AI to help users make memes. The company also used computer vision models to . You can use the tool completely for free. programmatically generate over 1 million GIFs has received 490 upvotes on Product Hunt. Automatic GIF Creator 9. Supersense Supersense has been developed by alumni from the at MIT. It is a mobile scanner app made for the blind or the visually-impaired. It is powered by AI technology that can scan your surroundings, text, and documents. Using object detection and recognition algorithms, Supersense uses the mobile phone’s camera to identify what the user is pointing at and dictates what it is. In the video review below, you can see how the app works on textbooks, US currency, food products, and even objects in the user’s immediate surroundings. Media Lab The app works as the user’s personal dictation assistant, saying aloud everything the user points the camera at. has received 768 upvotes on Product Hunt and is available on Android and iOS devices. Supersense 10. Facebook Computer Vision Tags While this isn’t exactly a computer vision startup, we wanted to include it in this list because it is one of Product Hunt’s most-upvoted computer vision software of all time. It won the badge for #1 Product of the Day when it was first posted in 2017. Facebook Computer Vision Tags is an that shows how Facebook’s AI models have annotated your images. Many people aren’t aware that Facebook uses computer vision models to automatically add labels to images uploaded on their platform. By using this plugin, can see the specific tags added to your photos. See an example of this below: open-source Google Chrome plugin Image via Github The original creator said their goal for this project was to make people more aware of the information that is constantly being extracted from the images we post on Facebook every day. has received 805 upvotes. Facebook Computer Vision Tags This list gives us a glimpse into the world of possibilities for computer vision startups. We hope that one of these products inspires you, maybe even to start building a new product yourself. Also published on: https://lionbridge.ai/articles/10-most-upvoted-computer-vision-startups-on-producthunt/

10 Computer Vision Startups on Product Hunt with the Most Upvotes

Edge AI starts with edge computing. Also called edge processing, edge computing is a network technology that positions servers locally near devices. This helps to reduce system processing load and resolve data transmission delays. These processes are performed at the location where the sensor or device generates the data, also called the edge. Note: This article was written by Tatsuo Kurita Developments in edge computing mean that edge AI is becoming more important. This is true across a variety of industries, particularly when it comes to processing latency and data privacy. In this article we’ll look at the impact of Edge AI, why it’s important, and common use cases for it. What is edge AI? Edge AI refers to AI algorithms that process locally on hardware devices, and can process data without a connection. This means operations such as data creation can occur without streaming or storing data in the cloud. This is important because there are an increasing number of cases where device data can’t be handled via the cloud. Factory robots and cars, for example, need high-speed processing with minimal latency. To achieve these goals, edge computing can generate data through deep learning on the cloud to develop deductive and predictive models at the data origin point, i.e. the device itself (the edge). We can see an example of this at work in factory robots. AI technology can be used here to visualize and assess vast amounts of multimodal data from surveillance cameras and sensors at speeds humans can’t process. We can also use it to detect faulty data on production lines that humans might miss. These kinds of IoT structures can store vast amounts of data generated from production lines and carry out analysis with machine learning. They are also at the heart of the deductive and predictive models that improve the smartification of factories. Edge AI, the Internet of Things, and 5G: Edge AI is often talked about in relation to the Internet of Things (IoT) and 5G networks. The term IoT refers to devices connected to each other through the internet, and includes smartphones, robotics, and electronic devices. As a platform that performs analysis with AI, edge AI can collect and store the vast amount of data generated by IoT, making it possible to use clouds with scalable characteristics. This allows for improved data processing and infrastructural flexibility. 5G networks can enhance the above-mentioned processes because their three major features — ultra-high speed, massive simultaneous connections, and ultra-low latency — clearly surpass that of 4G. 5G is indispensable for the development of IoT and edge AI, because when IoT devices transmit data, data volume swells and impacts transfer speed. Drops in transfer speed can create latency, which is the biggest issue when it comes to real-time processing. Why are edge computing and edge AI important? There are an increasing number of cases in which device data can’t be handled via the cloud. This is often the case with factory robots and cars, which require high-speed processing because of issues that can arise when increased data flow creates latency. For example, imagine a self-driving car suffering from cloud latency while detecting objects on the road, or operating the brakes or steering wheel. Any slowdown in data processing will result in a slower response from the vehicle. If the slowdown is such that the vehicle does not respond in time, this could result in an accident. Lives are literally at risk. For these IoT devices, a real-time response is a necessity. This means the ability for devices to analyze and assess images/data on the spot without relying on cloud AI. By entrusting edge devices with information processing usually entrusted to the cloud, we can achieve real-time processing without transmission latency. In addition, by limiting cloud data transmissions to only vital information, it is possible to reduce data volume and minimize communication interruptions. Edge AI Use Cases The edge AI market is chiefly comprised of two areas: industrial machinery, and consumer devices. We’re seeing progress with demonstration tests in areas including controlling and optimizing equipment, and automating skilled labor techniques. Progress is also being made with consumer devices that have cameras with AI that automatically recognize photographic subjects. Because the number of devices is larger than industrial machines, the consumer device market is expected to rise drastically from 2021 onwards. We’ve put some common use cases for edge AI below: Self-Driving Cars Self-driving cars are the most anticipated area of applied edge computing. There are many cases where self-driving cars have to make instantaneous assessments of a situation, and this requires real-time data processing. In December of 2019, revisions to the Road Traffic Act and Road Transportation Vehicle Law in Japan made it easier to get level 3 self-driving cars on the road. These include the safety standards that autonomous vehicles are held to, and the areas in which they can operate. As a result, car manufacturers are working on self-driving cars that adhere to these standards. Toyota, for example, is already testing full automation (level 4) with the TRI-P4. Autonomous Drones There’s been an increase of news about drones losing control and going missing while on remote flight experiments. This has even resulted in accidents. Depending on where the drone lands, a crash can be catastrophic. With autonomous drones, the pilot is not actively involved in the drone’s flight. They monitor the operation remotely, and only pilot the drone when absolutely necessary. The best known example of this is Amazon Prime Air, a drone delivery service which is developing self-piloting drones to deliver packages. Facial Recognition Facial recognition systems are a development in surveillance cameras, which can learn to recognize people by their faces. In November 2019, WDS Co., Ltd began supplying Eeye, an AI camera module that analyzes facial features in real-time through edge AI computing processes. Eeye recognizes faces quickly and accurately, and is suited for marketing tools that target characteristics such as gender and age, and face identification for unlocking devices. Smartphones This edge AI device is the one we’re all most familiar with. Siri and Google Assistant are good examples of edge AI on smartphones, as the technology drives their vocal user interfaces. With on-device AI, processing happens on the device (edge) side, meaning there is no need to deliver device data to the cloud. This helps secure privacy and reduce traffic. The Future of Edge AI Edge AI is growing, and we’ve seen big investments in the technology. Companies like Konduit AI are making it a key part of their AI strategy in Southeast Asia. Another example was in January of 2020, when it was reported that Apple invested 200 million dollars to acquire the Seattle-based AI enterprise, Xnor.ai. Xnor.ai’s AI tech processes data on the user’s smartphone with edge processing. With built-in AI on the smartphone itself, we’ll likely see advancements in voice processing, facial recognition technology, and enhanced privacy. According to the “2019 AI Business Aggregate Survey” published by Fuji Keizai Group, the edge AI computing market in Japan had a forecast market size of 11 billion yen in the 2018 fiscal year. The survey predicts the market to expand to 66.4 billion yen in the 2030 fiscal year. And with the spread of 5G, we’ll also likely see decreasing costs and increasing demand for edge AI services across the world. Author Profile From self-employed field engineer to PHP programmer, Tatsuo Kurita is now a UX director working mainly as a technical director to support corporate products. His expertise covers a wide range of areas, including certification in applied information technology, information security management, mental health management grade II, HTML, general deep learning, and AI implementation. Previously published on: https://lionbridge.ai/articles/what-is-edge-ai-computing/

A Simple Introduction to Edge AI

Our world will become filled with actors — virtual characters that know us better than we know ourselves Update: MagicLeap announced a virtual character demo: In China there are online celebrities who make over $10,000,000 a year broadcasting their lives via handheld cameras to paying subscribers. These streams will soon be popping out of the tiny restrictions of the mobile screen and appearing as holograms, dancing in your living room. The stage is then set to make take these mass broadcasts and make them individually tailored to each paying subscriber — with each performance being customised by A.I. The use of Extended Reality together with A.I. mixes together to make a brilliant new show, broadcast 24/7 with you as the star, surrounded by your favorite celebrities. Welcome to the future of TV. Japan is firmly ahead in this future with the fictional character, Hatsune Miku, already being adored by fans and appearing all around the world in a holographic form: The target audience of these Mixed Reality broadcasts is bypassing the classical Otaku — the geek who sits alone and only interacts with girls in computer games — to turn the heads of all youtube viewers of all ages. The future of celebrity broadcasting, and hence influencer marketing, will be a customised, ultra relevant, personal stream. For example we can imagine youtube celebs like a virtual Beyonce coming into your house, sitting on your sofa and looking across to talk to you, asking you details about your day. This scenario has been cleverly explored in the recent film “Her”. It will create a new form of broadcast entertainment where the stage is your living room and the virtualized actors are communicating with you in your intimate spaces. The first stage of this will be fixed hologram recordings, locked to a spot in your house via Augmented Reality.For example we can imagine youtube celebs like a virtual Beyonce coming into your house, sitting on your sofa and looking across to talk to you.The paradigm of entertainment will shift from being a distraction to being a companion. Entertainment will shift from being a distraction to being a companion. Virtual Avatars are already capable of fantastic emotional expression ( note the virtual actor below is being driven by a human performance): When these lifelike avatars are driven by A.I. the show will never really end. And at the end of each day the bell d’jour will be in your bedroom acting our bedroom stuff: This next stage of streaming — where A.I is being used to manipulate generative characters will be the most interesting. This kind of entertainment will be created in house and broadcast out of it. your The celebrity will supply the avatar and then each fan will enjoy a unique experience, depending on their environment and individual reactions. Because of this everyone will be constantly streaming experiences with their favourite characters incase something unique happens. This kind of bi-directional mass media has never been seen before and is perfectly suited to the Mixed Reality platform. It could soon become the new way of influencer marketing — with popular stars being used to sell everything on a very personal level. Today we already get glimpses of this future in developer demos: It is not crazy to think in a couple of years people will live alongside these virtual avatars, using them for entertainment, companionship and more. Combining A.I. Avatars with an algorithmic gaze will allow them to understand the context of beds, chairs, doors and help them live convincingly alongside their loyal fans. Magic Leap seems to be at the forefront of this idea with hardware and software available to the public to make this happen today: Amazon also recognised the importance of characters in the MR space and has pushed ahead with its Sumerian Project: Although still young, Amazon’s Sumerian paves the way for virtual actors that can appear in VR or AR spaces and can be communicated via Alexa’s natural spoken language. These Mixed Reality Avatars will have a view of the same room you are looking at ( either via your mobile camera, or via another camera you might install in your home for permanent A.I., which Amazon also sells): DEEPLENS This algorithmic gaze will give virtual avatars total contextual understanding of your home and life. It is likely that these ‘bot-eye’ cameras will be installed in every room in every home for your personal AR avatars to see through. Amazon is ahead here again, as it already sells Alexa units of varying sizes — intended one for each room. It’s not beyond the imagination to assume users will be happy buying Echo Dots with cameras on for the use of home avatar systems. For example you could put an A.I. camera in your kitchen, to watch the fridge and stove, and another by your front door to watch you come in. This would be enough for the AI driven M.R. avatar to know what you have in left in your fridge, to remind you to do the shopping, to watch what you have eaten, to tease you about your diet and even teach you to cook. And it can watch the stove whilst you are out of the room. It should do all this naturally. When the kettle is boiling the algorithmic gaze will see the steam, and the Avatar will notify me. An A.I. camera placed on your front door will ensure your friendly avatar is always there to greet you will a big smile when you come back home. Maybe these avatars will take the shape of dogs, cats, or branded cartoon characters. And lets not forget there are funded companies working on Mixed Reality technology that does not require the user to wear or hold any devices to see this reality, holographic projections are possible: Lightfields Mixed Reality, with the combination of Computer Vision, A.I. and digital overlays will allow everyday human behaviours to be offloaded from your mind to another system without needing complex infrastructure to connect devices or expensive robotics to interact with humans and systems. In the near future it is very likely we will live and work alongside intelligent robots, but only in the virtual sense. Next consider that these M.R. avatars could not only be celebrities — they could be whoever you fancy, or they could be long distance friends, deceased relatives or even clones of yourself. This technology is here today, nascent, and possible. It is one of the areas Ninjar is exploring as future use cases for M.R. We believe the utilisation of M.R. Avatars will be for more than just entertainment, they will be customer service representatives and skilled tradesman too. For example if you want to see how a new BMW looks in your driveway, you can just place one there via AR. ( We’ve done plenty of work on this, as you can see here: ) ninjar.com/work But A.R. is not the best sales pitch by itself, not compared to the experience of visiting a showroom and having a brand employee pitch emotionally to you about the upgrade. M.R. avatars will become virtual salespeople. Part of our vision at is that when you drop an A.R. car on your driveway you should expect a M.R. salesperson to come along for the ride. They will know what you are looking at and respond correctly — if you move your phone for a closeup of the alloys you can expect them to say “Hey, Good choice on those wheels — I can give them to you for a free upgrade…” ninjar.com One argument we hear to the idea of Mixed Reality avatars is “But they can’t do anything real — they can’t touch anything… They can’t pick up a hammer and knock down a wall!”. This is true, but then again — when was the last time you picked up a hammer or used any tool to permanently change the real world? For most of us we live in world where the physical manipulation of hardware is outdated or automated. Food is delivered pre-cooked and in disposable trays. Apartments come serviced and need little maintenance. Home based mechanical repair is a thing of the past, replaced with planned obsolescence, appliance trade-ins and annual upgrade plans. It is almost like modern life has left a void in the homes where real world tools, and robots, are no longer essential. Conclusion Mixed Reality avatars, combined with ubiquitous A.I. cameras, might be our main access to computation in the near future. They will see what we see, yet they will have the superhuman intelligence, memory and connections from the cloud. They will become so useful we will miss them when they are not there. They will look astonishingly real: Extra Notes… But will this be an authentic experience that people will want? We’ve learnt from mobile apps and social networks that desire is easy to manipulate. Social App designers use a technique called F.O.M.O. — Fear of Missing Out — to drive retention. This psychological trick keeps us checking social feeds for fear we might miss a post from a friend. This hijacking of user’s minds to increase retention will be strengthened further by the natural behaviours of the realistic M.R. avatars and the consideration that they will be a channel for all our important messages, friendships and potential sexual opportunities. For example, in the world of Mixed Reality avatars, if I want to send you a message to ask you something, I can ‘send’ my avatar — my clone — over to your house to ask. And you can answer back and I will see your avatar, your clone sitting next to me. This natural interface might make today’s chat apps look ridiculously basic and is probably part of the reason why Facebook invested a hefty sum to buy Oculus. They realize the future of Social is MR. This idea of a virtual space filled with believable avatars that hold and mediate the connections between us and the rest of humanity is and should be scary. This new interface will become the truth of our news, our friends, potential dates and all product sales and will become a powerful platform. The combination of an Information Technology, and a Social Network, along with a emotion display via realistic humans is uniquely powerful. There are other, possibly negative, effects of surrounding yourself with MR avatars. Consider the fact that when your phone is near you, you become stupider. The findings of a study in Social Psychology publication found that the “mere presence” of a mobile — even when it is not being used — influenced simple mental performance. It was noted that when a user’s phone was in view the person did 20% worse at a given task. Imagine how lazy your mind would become when surrounded by A.I. driven avatars. Did you enjoy our post? You can comment and join the chat on Medium We’d love to hear from you. Come find us on , and Twitter Facebook LinkedIn

The Evolution of Mixed Reality Avatars

Around half a year ago I decided to start decorating my new house. Before getting started, I did lots of research on a variety of different topics relating to interior decoration, such as how to choose a consistent color scheme, which measurements to make and how to make them, and how to choose the right furniture. However, my preparations made me realize that no matter how well prepared you are, you're always going to run into many unexpected challenges. Before rushing to the furniture store, I listed all the different pieces of furniture that I wanted to place in my living room, including a sofa, tea table, potted plants, dining table, and carpet, and determined the expected dimensions, colors, and styles of these various items of furniture. However, when I finally got to the furniture store, the dizzying variety of choices had me confused, and I found it very difficult to imagine how the different choices of furniture would actually look like in actual living room. At that moment a thought came to my mind: wouldn't it be great if there was an app that allows users to upload images of their home and then freely select different furniture products to see how they'll look like in their home? Such an app would surely save users wishing to decorate their home lots of time and unnecessary trouble, and reduce the risks of users being dissatisfied with the final decoration result. That's when the idea of developing an app by myself came to my mind. My initial idea was to design an app that people could use to help them quickly satisfy their home decoration needs by allowing them see what furniture would look like in their homes. The basic way the app works is that users first upload one or multiple images of a room they want to decorate, and then set a reference parameter, such as the distance between the floor and the ceiling. Armed with this information, the app would then automatically calculate the parameters of other areas in the room. Then, users can upload images of furniture they like into a virtual shopping cart. When uploading such images, users need to specify the dimensions of the furniture. From the editing screen, users can drag and drop furniture from the shopping cart onto the image of the room to preview the effect. But then a problem arises: images of furniture dragged and dropped into the room look pasted on and do not blend naturally with their surroundings. By a stroke of luck, I happened to discover when looking for a solution for the aforementioned problem. This development kit provides the ability to integrate virtual objects realistically into the real world, which is exactly what my app needs. With its plane detection capability, my app will be able to detect the real planes in a home and allow users to place virtual furniture based on these planes; and with its hit test capability, users can interact with virtual furniture to change their position and orientation in a natural manner. HMS Core AR Engine Next, I'd like to briefly introduce the two capabilities this development kit offers. AR Engine tracks the illumination, planes, images, objects, surfaces, and other environmental information, to allow apps to integrate virtual objects into the physical world and look and behave like they would if they were real. Its plane detection capability identifies feature points in groups on horizontal and vertical planes, as well as the boundaries of the planes, ensuring that your app can place virtual objects on them. In addition, the kit continuously tracks the location and orientation of devices relative to their surrounding environment, and establishes a unified geometric space between the virtual world and the physical world. The kit uses its hit test capability to map a point of interest that users tap on the screen to a point of interest in the real environment, from where a ray will be emitted pointing to the location of the device camera, and return the intersecting point between the ray and the plane. In this way, users can interact with any virtual object on their device screen. Functions and Features Plane detection: Both horizontal and vertical planes are supported. Accuracy: The margin of error is around 2.5 cm when the target plane is 1 m away from the camera. Texture recognition delay: < 1s Supports polygon fitting and plane merging. Development Procedure (Plane Detection) Create a object. This example demonstrates how to use the world AR scenario of AR Engine. WorldActivity Public class WorldActivity extends BaseActivity{
Protected void onCreate (Bundle saveInstanceState) {
Initialize DisplayRotationManager.
  mDisplayRotationManager = new DisplayRotationManager(this);
Initialize WorldRenderManager.
     mWorldRenderManager = new WorldRenderManager(this,this);
}
// Create a gesture processor.
 Private void initGestureDetector(){
mGestureDetector = new GestureDetector(this,new GestureDetector.SimpleOnGestureListener()){
  }
 }
 mSurfaceView.setOnTouchListener(new View.OnTouchListener()){
  public Boolean onTouch(View v,MotionEvent event){
   return mGestureDetector.onTouchEvent(event);
  }
 }
// Create ARWorldTrackingConfig in the onResume lifecycle.
 protected void onResume(){
  mArSession = new ARSession(this.getApplicationContext());
  mConfig = new ARWorldTrackingConfig(mArSession);
  …
}
// Initialize a refresh configuration class.
private void refreshConfig(int lightingMode){
// Set the focus.
 mConfig.setFocusMode(ARConfigBase.FocusMode.AUTO_FOCUS);
 mArSession.configure(mConfig);
}
} 2) Initialize the class, which manages rendering related to world scenarios, including label rendering and virtual object rendering. WorldRenderManager public class WorldRenderManager implements GLSurfaceView.Renderr{
// Initialize a class for frame drawing.
 Public void onDrawFrame(GL10 unused){
// Set the openGL textureId for storing the camera preview stream data.
mSession.setCameraTextureName(mTextureDisplay.getExternalTextureId());
// Update the calculation result of AR Engine. You are advised to call this API when your app needs to obtain the latest data.
ARFrame arFrame = mSession.update();
// Obtains the camera specifications of the current frame.
ARCamera arCamera = arFrame.getCamera();
// Returns a projection matrix used for coordinate calculation, which can be used for the transformation from the camera coordinate system to the clip coordinate system.
arCamera.getProjectionMatrix(projectionMatrix, PROJ_MATRIX_OFFSET, PROJ_MATRIX_NEAR, PROJ_MATRIX_FAR);
Session.getAllTrackables(ARPlane.class)
...
 }
} 3) Initialize the class, which provides properties of the virtual object and the necessary methods for rendering the virtual object. VirtualObject Public class VirtualObject{
} 4) class to draw virtual objects based on specified parameters. Initialize the tDisplay Objec Public class ObjectDisplay{
} Development Procedure (Hit Test) Initialize the class, which manages rendering related to world scenarios, including label rendering and virtual object rendering. WorldRenderManager public class WorldRenderManager implementsGLSurfaceView.Renderer{
// Pass the context.
 public WorldRenderManager(Activity activity, Context context) {
  mActivity = activity;
       mContext = context;
  …
}
// Set ARSession, which updates and obtains the latest data in OnDrawFrame.
public void setArSession(ARSession arSession) {
        if (arSession == null) {
            LogUtil.error(TAG, "setSession error, arSession is null!");
            return;
        }
        mSession = arSession;
 }
// Set ARWorldTrackingConfig to obtain the configuration mode.
public void setArWorldTrackingConfig(ARWorldTrackingConfig arConfig) {
        if (arConfig == null) {
LogUtil.error(TAG, "setArWorldTrackingConfig error, arConfig is null!");
            return;
        }
        mArWorldTrackingConfig = arConfig;
    }
// Implement the onDrawFrame() method.
@Override
public void onDrawFrame(GL10 unused) {
 mSession.setCameraTextureName(mTextureDisplay.getExternalTextureId());
ARFrame arFrame = mSession.update();
ARCamera arCamera = arFrame.getCamera();
...
}
// Output the hit result.
private ARHitResult hitTest4Result(ARFrame frame, ARCamera camera, MotionEvent event) {
   ARHitResult hitResult = null;
   List<ARHitResult> hitTestResults = frame.hitTest(event);
// Determine whether the hit point is within the plane polygon.
ARHitResult hitResultTemp = hitTestResults.get(i);
            if (hitResultTemp == null) {
                continue;
            }
ARTrackable trackable = hitResultTemp.getTrackable();
// Determine whether the point cloud is tapped and whether the point faces the camera.
  boolean isPointHitJudge = trackable instanceof ARPoint
&& ((ARPoint) trackable).getOrientationMode() == ARPoint.OrientationMode.ESTIMATED_SURFACE_NORMAL;
// Select points on the plane preferentially.
if (isPlanHitJudge || isPointHitJudge) {
      hitResult = hitResultTemp;
      if (trackable instanceof ARPlane) {
                break;
             }
      }
return hitResult;
}
} 2) Create a object. This example demonstrates how to use the world AR scenario of AR Engine. WorldActivity public class WorldActivity extends BaseActivity {
 private ARSession mArSession;
private GLSurfaceView mSurfaceView;
private ARWorldTrackingConfig mConfig;
@Override
protected void onCreate(Bundle savedInstanceState) {
  LogUtil.info(TAG, "onCreate");
       super.onCreate(savedInstanceState);
       setContentView(R.layout.world_java_activity_main);
 mWorldRenderManager = new WorldRenderManager(this, this);
mWorldRenderManager.setDisplayRotationManage(mDisplayRotationManager);
mWorldRenderManager.setQueuedSingleTaps(mQueuedSingleTaps)  
}
@Override
protected void onResume() {
 if (!PermissionManager.hasPermission(this)) {
            this.finish();
        }
        errorMessage = null;
        if (mArSession == null) {
            try {
                if (!arEngineAbilityCheck()) {
                    finish();
                    return;
                }
                mArSession = new ARSession(this.getApplicationContext());
                mConfig = new ARWorldTrackingConfig(mArSession);
                refreshConfig(ARConfigBase.LIGHT_MODE_ENVIRONMENT_LIGHTING | ARConfigBase.LIGHT_MODE_ENVIRONMENT_TEXTURE);
            } catch (Exception capturedException) {
                setMessageWhenError(capturedException);
            }
            if (errorMessage != null) {
                stopArSession();
                return;
            }
}

@Override
    protected void onPause() {
        LogUtil.info(TAG, "onPause start.");
        super.onPause();
        if (mArSession != null) {
            mDisplayRotationManager.unregisterDisplayListener();
            mSurfaceView.onPause();
            mArSession.pause();
        }
        LogUtil.info(TAG, "onPause end.");
    }
@Override
    protected void onDestroy() {
        LogUtil.info(TAG, "onDestroy start.");
        if (mArSession != null) {
            mArSession.stop();
            mArSession = null;
        }
        if (mWorldRenderManager != null) {
            mWorldRenderManager.releaseARAnchor();
        }
        super.onDestroy();
        LogUtil.info(TAG, "onDestroy end.");
    }
...
} Summary If you've ever done any interior decorating, I'm sure you've wanted the ability to see what furniture would look like in your home without having to purchase them first. After all, most furniture isn't cheap and delivery and assembly can be quite a hassle. That's why apps that allow users to place and view virtual furniture in their real homes are truly life-changing technologies. HMS Core AR Engine can help greatly streamline the development of such apps. With its plane detection and hit test capabilities, the development kit enables your app to accurately detect planes in the real world, and then blend virtual objects naturally into the real world. In addition to virtual home decoration, this powerful kit also has a broad range of other applications. For example, you can leverage its capabilities to develop an AR video game, an AR-based teaching app that allows students to view historical artifacts in 3D, or an e-commerce app with a virtual try-on feature.

I Decorated My House Using AR: Here's How I Did It

The whole proceeding was a delirium, a madness.

The pile of paper currency

2024 has been a challenging year so far for Apple’s embattled stock, but could the launch of the iPhone 16 and arrival of generative AI capabilities for iOS help the smartphone giant recapture its status as a market leader? Apple’s well-documented problems have been playing out on Wall Street throughout the year, and the stock finished the first quarter more than 7.5% lower after Barclays issued a downgrade owing to a lack of innovation and slowing sales in China. "The iPhone 15 has been lackluster and we believe iPhone 16 should be the same," in a client note in January. warned Barclays analyst Tim Long Owing in part to geopolitical tensions between the United States and China, Apple fell to fifth place in the nation’s smartphone market as shipments to 10 million units, with Huawei claiming a 17% market share at the top of the pile. declined by 25% It’s a similar story among global markets for Apple. Once an undisputed leader in terms of smartphone sales, Korean firm Samsung claimed the top spot for sales volumes worldwide in Q1 2024, , outstripping Apple’s 48.7 million units. shipping 60 million units This indicates that 2024 is shaping up to be a key battleground in determining Apple’s future, and the company that’s typically at the forefront of innovation will be well aware that the iPhone 16 must be a hit to recapture its former glories. As global competition continues to grow, the generative AI boom will likely be seen as the solution for Apple’s difficult start to the year, but will it be enough to recapture the Silicon Valley firm’s lost market share? Embracing the Generative AI Boom The expectation of Apple’s emergence into the world of generative AI is growing. Stock market analysts Morningstar has even sought to for iPhone revenue in 2025 owing to the generative AI functionality that it anticipates will be added to the iPhone 16. raise forecasts Apple appeared to have taken a more steady approach towards embracing generative AI in the months that followed the launch of ChatGPT and its widespread introduction to the vast potential of large-language models (LLMs) and their limitless industry capabilities. However, CEO that the implementation of generative AI tools for iPhone and other Apple products would soon be on the way. Earlier this year, Cook claimed that substantial investment in "a tremendous amount of time and effort" towards GenAI has been taken, and we recently saw the launch of OpenELM, or ‘Open-source Efficient Language Models’ as a tool that could ramp up Apple’s bid to embrace the emerging technology. Tim Cook highlighted It’s clear that the best of Apple’s generative AI bid will be found in the iPhone 16, which is due to be launched in the second half of the year. But will it be capable of turning the tide of Apple’s falling market share? Turning the iPhone 16 Into a Success The launch of the iPhone 16 this year is going to be a significant moment for Apple because it will provide the answer to many questions currently being asked about the company. Can Apple recover its market dominance? How will Apple use generative AI to improve iOS? Is Apple suffering from a lack of innovation? The latter question is becoming increasingly crucial for the company and its various stakeholders. Echoing Barclays analyst Tim Long’s sentiment off the back of Apple’s worrying downgrade on Wall Street, there’s a that CEO Tim Cook isn’t as innovative as his iconic predecessor, Steve Jobs. growing school of thought Now, as generative AI emerges as a key opportunity for Apple to embrace exciting new technology and utilize it within its products, Cook has the chance to outperform Apple’s rivals in delivering meaningful innovation in a smartphone industry that’s ripe for growth following the arrival of 5G connectivity. So, can we use what we already know about the iPhone 16 to anticipate whether it will be a success? Available information suggests that the screen size of the iPhone 16 will be , which is 0.2” . 6.9 inches larger than the iPhone 15 Pro Max At 48 megapixels, the iPhone 16’s camera is likely to be , but with an and ‘pixel binning’ to combine data from four pixels into one single ‘superpixel’ to improve quality overall. similar in spec to the iPhone 15 ultra-wide lens The expectation of a larger battery and an A18 Pro chipset weighs in at an , but so far, there isn’t much evidence of innovation that appears ready to threaten the dominance of Huawei and Samsung in Chinese and global markets respectively. improvement on the iPhone 15 Pro However, the expectation of generative AI features throughout many existing apps and a series of new ones will make the iPhone 16 a watershed release for Apple. This raises the question, are we waiting for the next iteration of the iPhone? Or a major iOS update? Will iOS 18 Save Apple’s Slump? Apple’s biggest launch of 2024 will be its integrated LLMs into the next major iteration of its iOS operating system. The reason that this is exciting is because of the sheer effort that Apple has been undertaking in making its generative AI capabilities truly transformative. The company’s recent AI acquisitions , which specialize in building AI models on limited hardware. include DarwinAI and Datakalab Implementing generative AI capabilities into the next generation of iOS will open the door to unprecedented intelligence for users. Because of their generative nature, apps can work alongside one another to actively update calendars the moment plans are made on WhatsApp, emojis or GIFs will be created using text-based prompts, and users will be capable of automatically responding to messages with GenAI algorithms capable of mimicking their writing style and personality. This is only a glimpse at some of the transformative capabilities offered by generative AI, and it’s not only a matter of time before they become commonplace throughout iOS. Whether the scale of Apple’s generative AI strategy means that an iPhone 16 would be essential to unlock the full capabilities of the technology remains to be seen, but this could be crucial in the company’s overall market performance in 2024 and beyond. According to a , “this year, [Apple] will not only greatly strengthen the AI computing power of M3 and A17 processors, but also significantly increase the number of AI computing cores and performance of the new generation of M4 and A18 processors.” recent report from Economic Daily News Could this mean that the iPhone 16 will be essential to fully utilize generative AI? Considering that Bloomberg has suggested generative AI will become a , the lure of GenAI within the next generation of iPhone could be strong enough to spark a market revival. $1.3 trillion market by 2032 The Long Road to Recovery Apple’s future performance will depend on other factors as well as its rollout of generative AI capabilities. The current geopolitical issues between the United States and China threaten the firm’s global dominance, but recapturing its sense of innovation can be a major stepping stone in Apple’s recovery. Although the iPhone remains a Western powerhouse, recapturing the spirit evoked by Steve Jobs in the launch of the iPhone 16 could go some way in transforming the sentiment surrounding Apple. In a market that’s full of capable global players, the next generation of iPhone and iOS has never been more important.

Weight of AI Expectation Weighs Heavy as iPhone 16 Launch Becomes Pivotal for Apple

আইফোন 16 লঞ্চ অ্যাপলের জন্য গুরুত্বপূর্ণ হয়ে উঠেছে বলে এআই প্রত্যাশার ওজন ভারী

Die Erwartungen an die KI sind groß, da die Einführung des iPhone 16 für Apple von entscheidender Bedeutung ist

El peso de las expectativas de IA pesa mucho a medida que el lanzamiento del iPhone 16 se vuelve fundamental para Apple

Le poids des attentes en matière d’IA pèse lourd alors que le lancement de l’iPhone 16 devient crucial pour Apple

एआई की उम्मीदों का वजन भारी है क्योंकि iPhone 16 का लॉन्च Apple के लिए महत्वपूर्ण बन गया है

iPhone 16の発売がAppleにとって極めて重要になる中、AIへの期待が重くのしかかる

iPhone 16 출시가 Apple의 중추적인 역할을 하게 되면서 AI에 대한 기대의 무게가 무거워졌습니다.

O peso da expectativa de IA pesa muito à medida que o lançamento do iPhone 16 se torna fundamental para a Apple

Ожидания в отношении искусственного интеллекта тяжелы, поскольку запуск iPhone 16 становится решающим для Apple

iPhone 16'nın Lansmanı Apple İçin Önemli Hale Gelirken Yapay Zeka Beklentisinin Ağırlığı Ağırlaşıyor

Kỳ vọng về AI ngày càng nặng nề khi việc ra mắt iPhone 16 trở thành vấn đề quan trọng đối với Apple

iPhone 16 发布对苹果来说至关重要，对 AI 的期望也随之增加

Authors: (1) Ruohan Zhang, Department of Computer Science, Stanford University, Institute for Human-Centered AI (HAI), Stanford University & Equally contributed; zharu@stanford.edu; (2) Sharon Lee, Department of Computer Science, Stanford University & Equally contributed; zharu@stanford.edu; (3) Minjune Hwang, Department of Computer Science, Stanford University & Equally contributed; zharu@stanford.edu; (4) Ayano Hiranaka, Department of Mechanical Engineering, Stanford University & Equally contributed; zharu@stanford.edu; (5) Chen Wang, Department of Computer Science, Stanford University; (6) Wensi Ai, Department of Computer Science, Stanford University; (7) Jin Jie Ryan Tan, Department of Computer Science, Stanford University; (8) Shreya Gupta, Department of Computer Science, Stanford University; (9) Yilun Hao, Department of Computer Science, Stanford University; (10) Ruohan Gao, Department of Computer Science, Stanford University; (11) Anthony Norcia, Department of Psychology, Stanford University (12) Li Fei-Fei, 1Department of Computer Science, Stanford University & Institute for Human-Centered AI (HAI), Stanford University; (13) Jiajun Wu, Department of Computer Science, Stanford University & Institute for Human-Centered AI (HAI), Stanford University. Table of Links Abstract & Introduction Brain-Robot Interface (BRI): Background The NOIR System Experiments Results Conclusion, Limitations, and Ethical Concerns Acknowledgments & References Appendix 1: Questions and Answers about NOIR Appendix 2: Comparison between Different Brain Recording Devices Appendix 3: System Setup Appendix 4: Task Definitions Appendix 5: Experimental Procedure Appendix 6: Decoding Algorithms Details Appendix 7: Robot Learning Algorithm Details Appendix 7: Robot Learning Algorithm Details Object and skill learning details We utilize pre-trained R3M as the feature extractor. Our training procedure aims to learn a latent representation of an input image for inferring the correct object-skill pair in the given scene. The feature embedding model is a fully-connected neural network that further encodes the outputs of the foundation model. Model parameters and training hyperparameters are summarized in Table 7. Collecting human data using a BRI system is expensive. To enable few-shot learning, the feature embedding model is trained using a triplet loss [80], which operates on three input vectors: anchor, positive (with the same label as the anchor), and negative (with a different label). Triplet loss pulls inputs with the same labels together by penalizing their distance in the latent space, as well as pushes inputs with different labels away. The loss function is defined as: where a is the anchor vector, p the positive vector, n the negative vector, f the model, and α = 1 is our separation margin. We test our algorithm in the following generalization settings: Generalization test set. For position generalization, we randomize the initial positions of all objects in the scene with fixed orientation and collect 20 different trajectories. For the pose generalization, we randomize both the initial positions and orientations of all objects. Position and pose. . The context-level generalization refers to placing the target object in different environments, defined by different backgrounds, object orientations, and the inclusion of different objects in the scene. We collect 20 different trajectories with these variations. Context . The instance generalization aims to assess the model’s capability to generalize across different types of objects present in the scene. For our target task (MakePasta), we collect 20 trajectories with 20 different kinds of pasta with different shapes, sizes, and colors. Instance . To understand the separability of latent representations generated by our object-skill learning model, we visualize the 1024-dimensional final image representations using the t-SNE data visualization technique [81]. Results for the MakePasta pose generalization test set are shown in Figure 7. The model can well separate each of the different stages of the task, allowing us to retrieve the correct object-skill pair for an unseen image. Latent representation visualization One-shot parameter learning details We empirically found that using DINOv2’s ViT-B model, alongside a 75x100 feature map and a 3x3 sliding window, with cosine similarity as the distance metric, resulted in the best performance for our image resolutions. Design choices. We test the generalization ability of our algorithm on 1008 unique training and test pairs, encompassing four types of generalizations including 8 position trials, 8 orientation trials, 32 context trials, and 960 instance trials. Generalization test set The position and orientation generalizations, shown in Fig. 8 and Fig. 9 respectively, are tested in isolation, e.g. when the position is varied, the orientation is kept the same. Position and orientation The context-level generalization, shown in Fig. 10, refers to placing the target object in different environments, e.g. the training image might show the target object in the kitchen while the test image shows the target object in a workspace. Here, we allow for position and orientation to vary as well. Context To test our algorithm’s capability of instance-level generalization, shown in Fig. 11, we collected a set of four different object categories, each containing five unique object instances. Our object categories consist of mug, pen, bottle, and medicine bottle, whereas the bottle and medicine bottle categories consist of images from both the top-down and side views. We test all permutations within each object category including train and test pairs with different camera views. Here, we allow for position and orientation to vary as well. Instance We test our method against baselines on 1080 unique training and test pairs, encompassing four types of generalizations including 8 position trials, 8 orientation trials, 32 context trials, 960 instance trials, 48 trials where we vary all four generalizations simultaneously, and 24 trials from the SetTable task. Test set for comparing our method against baselines Finally, we test our algorithm’s ability to generalize when all four variables differ between the training and test image, shown in Fig. 12. Here, the only object category we use is a mug. Position, orientation, context, and instance simultaneously. This paper is under CC 4.0 license. available on arxiv

NOIR: Neural Signal Operated Intelligent Robots for Everyday Activities: Appendix 7

Authors: (1) Ruohan Zhang, Department of Computer Science, Stanford University, Institute for Human-Centered AI (HAI), Stanford University & Equally contributed; zharu@stanford.edu; (2) Sharon Lee, Department of Computer Science, Stanford University & Equally contributed; zharu@stanford.edu; (3) Minjune Hwang, Department of Computer Science, Stanford University & Equally contributed; zharu@stanford.edu; (4) Ayano Hiranaka, Department of Mechanical Engineering, Stanford University & Equally contributed; zharu@stanford.edu; (5) Chen Wang, Department of Computer Science, Stanford University; (6) Wensi Ai, Department of Computer Science, Stanford University; (7) Jin Jie Ryan Tan, Department of Computer Science, Stanford University; (8) Shreya Gupta, Department of Computer Science, Stanford University; (9) Yilun Hao, Department of Computer Science, Stanford University; (10) Ruohan Gao, Department of Computer Science, Stanford University; (11) Anthony Norcia, Department of Psychology, Stanford University (12) Li Fei-Fei, 1Department of Computer Science, Stanford University & Institute for Human-Centered AI (HAI), Stanford University; (13) Jiajun Wu, Department of Computer Science, Stanford University & Institute for Human-Centered AI (HAI), Stanford University. Table of Links Abstract & Introduction Brain-Robot Interface (BRI): Background The NOIR System Experiments Results Conclusion, Limitations, and Ethical Concerns Acknowledgments & References Appendix 1: Questions and Answers about NOIR Appendix 2: Comparison between Different Brain Recording Devices Appendix 3: System Setup Appendix 4: Task Definitions Appendix 5: Experimental Procedure Appendix 6: Decoding Algorithms Details Appendix 7: Robot Learning Algorithm Details Appendix 3: System Setup The robot we use in our tabletop manipulation task is a standard Franka Emika robot arm with three RealSense cameras. For mobile manipulation, we use a Tiago++ model from PAL Robotics, with an omnidirectional base, two 7-degrees-of-freedom arms with parallel-yaw grippers, a 1-degree-of-freedom prismatic torso, two SICK LiDAR sensors (back and front of the base), and an ASUS Xtion RGB-D camera mounted on the robot’s head, which can be controlled in yaw and pitch. All sensors and actuators are connected through the Robot Operating System, ROS [74]. The code runs on a laptop with an Nvidia GTX 1070 that sends the commands to the onboard robot computer to be executed. Robot platform. A list of primitive skills along with their parameters can be found in Table 6, eight for Franka (16 tasks) and five for Tiago (four tasks). Human users can accomplish all 20 tasks, which are long-horizon and challenging, using these skills. Primitive skills list. This paper is under CC 4.0 license. available on arxiv

NOIR: Neural Signal Operated Intelligent Robots for Everyday Activities: Appendix 3

This paper is available on arxiv under CC 4.0 license. Authors: (1) Berrenur Saylam; (2) Ozlem Durmaz ¨ ˙Incel. Table of Links Highlights Abstract & Introduction Federated Learning: Collaborative and Privacy-Preserving Model Training State of the Art Federated Learning Software Platforms and Testbeds Sensors and Edge-Sensing Devices Federated Learning Applications on Sensing Devices Discussion Conclusions, Acknowledgment, and References 5. Federated Learning Applications on Sensing Devices Small sensing devices can capture location-based or person-based data related to specific environments or behaviours. Data collected from multiple devices can be combined according to the application’s requirements to gain a comprehensive understanding of situations or behaviours. This section explores the diverse application areas where FL techniques have been utilized (see Figure 4). These areas include Human Activity Recognition (HAR), health monitoring, affect detection, abnormality detection, and other relevant areas where sensors on edge devices are utilized. We are inspired by the study Nguyen et al. (2021) and extended it by adding different application areas. Due to the page limitations, we did not include all studies covering these application areas; instead, we provided examples of studies from these areas in the literature. In Table 1, we present an overview of the covered studies, which are explained in the following sections in detail. 5.1. Medical applications In today’s world, computational-based techniques are widely used in the healthcare domain, especially with image recognition, e.g., the detection of diseases, the decision of cell types according to sizes and forms. The benefits of using these techniques are fast and accurate diagnosis and treatment Szegedi et al. (2019); Wang et al. (2023). However, data protection regulations make it impossible to collect and analyze these data easily. In this regard, FL techniques are well suited to healthcare-related applications from the privacy perspective. In addition, intrinsically, FL is designed for a vast number of client participants. The study Ogbuabor et al. (2021) proposes an intelligent and privacy-oriented context-aware decision support system for cardiac health monitoring using physiological and activity data of the patient during rehabilitation, utilizing a federated machine learning approach for activity recognition while maintaining users’ privacy. • Disease Diagnosis: The system allows healthcare professionals to assess the health status of patients by considering multiple relevant parameters, such as heart rate, electrocardiogram (ECG) signals, and activity data. It uses the data collected from Holter monitors and smartphones equipped with accelerometers. The federated machine learning approach enables the development of a collaborative model without sharing sensitive patient data, ensuring data privacy while achieving model generalization. Affective computing Picard (2000) is an active research area for automatically monitoring a person’s mental and emotional state via physiological and physical signals. In Can and Ersoy (2021), the authors applied FL on stress data with the aim of privacy preservation. • Affect Detection: The authors applied FL on heart activity data collected with smart bands, and the targeted devices include wrist-worn wearable devices (e.g., Samsung Gear S, Empatica E4) and smartphones, with various physiological sensors such as accelerometers, electrodermal activity, heart rate, etc. The proposed federated deep learning algorithm achieves encouraging results in stress-level detection while ensuring privacy protection for health-related research conducted with widespread mobile unobtrusive devices. Two use-case scenarios are considered: creating a general model from individual models while preserving user privacy and aggregating separately collected event data to develop an improved shared model. They applied deep learning algorithm in a federated manner and obtained better results compared to traditional learning methods. HAR aims to detect users’ activities, such as sitting, standing, and walking, mostly via motion sensors. The sensors used in HAR tasks typically include triaxial accelerometers and gyroscopes found in smartphones and smartwatches. • Human Activity Recognition: These sensors measure acceleration forces and rotational movements, allowing the models to recognize various activities such as walking, running, biking, and more. In Feng et al. (2020), FL models were applied for trajectory prediction. The framework utilizes various sensing devices and sensors present in mobile devices to gather location-based data. The paper introduces a group optimization method for training on local devices to achieve a better trade-off between performance and privacy. In Sozinov et al. (2018), authors applied FL to detect basic human activities, where the goal is to recognize different human activities using sensor data from smartphones or smartwatches. They obtained good accuracies but slightly lower than a centralized one. In Ek et al. (2020), they focused on the HAR domain and applied several aggregation algorithms (FedAvg, FedPer, and FedMA) and compared their results. 5.2. Environmental applications FL has emerged as a powerful approach to tackle environmental challenges like climate change, air pollution, and extreme weather events. FL is employed to predict air pollution using sensor networks in Nguyen and Zettsu (2021). The utilized sensing devices are environmental monitoring stations with various sensors, which are not specified in the paper. • Air Pollution Monitoring: Nevertheless, a combination of sensors is typically deployed to collect data on various environmental parameters such as particulate matter (PM), nitrogen dioxide (NO2), ozone (O3), carbon monoxide (CO), sulfur dioxide (SO2), meteorological data (temperature, humidity, wind speed, etc.), and more. The paper focuses on predicting air pollution levels, particularly Oxidant warning levels, using Convolutional Recurrent Neural Networks (CRNN). The CRNN models are trained locally on data from different spatial areas, such as cities and prefectures, and then the common parts of these models are aggregated at the server to create a global model. This federated learning framework allows for cooperative training among participants from different regions without exchanging raw data, ensuring data privacy and reducing data transmission latency. • FL is utilized for weather prediction using distributed weather stations Chen et al. (2023). The stated paper introduces a novel machine learning approach, MetePFL, for weather forecasting on a global scale. The proposed approach utilizes a foundation model (FM) pre-trained on extensive weather data and finetuned by regions to capture local weather patterns. Weather Forecasting: MetePFL employs prompt federated learning to train a Transformer-based FM collaboratively across participants with heterogeneous meteorological data. A spatiotemporal prompt learning mechanism is introduced to handle multivariate time series data efficiently, improving forecasting accuracy. The experiments conducted on accurate weather forecasting datasets demonstrate the effectiveness and superiority of the MetePFL approach over traditional fine-tuning and other federated learning methods. By utilizing only a fraction of the model’s parameters, MetePFL achieves excellent performance while reducing inter-device communication overhead, making it a promising solution for comprehensive weather forecasting on large-scale meteorological data. 5.3. Smart city applications FL techniques are also applied to integrated city solutions related to food, water, and energy So et al. (2021). Applications use data collected from cameras, temperature sensors, and traffic sensors to understand human activities, population density, and pollution Jiang et al. (2020). FL is employed for traffic flow prediction in smart cities Liu et al. (2020). The paper Liu et al. (2020) presents FedGRU, a novel privacy-preserving traffic flow prediction algorithm based on FL. • Traffic Pattern Analysis: The goal is to develop a collaborative model that can accurately predict traffic patterns and congestion levels using data from various sources, such as traffic cameras and sensors. The proposed algorithm uses a secure parameter aggregation mechanism. The paper introduces a joint-announcement protocol to handle large-scale scenarios, randomly selecting a subset of organizations for each training round, effectively mitigating communication challenges. Empirical case studies demonstrate that FedGRU achieves comparable prediction accuracy to centralized models while maintaining robust privacy protection. An ensemble clustering-based approach is proposed to enhance model performance by leveraging spatiotemporal correlations among organizations’ data. Transportation is composed of two parts: vehicular traffic planning and resource management Nguyen et al. (2021). Traffic planning is essential for traffic prediction and congestion minimization. In Elbir et al. (2022), authors replaced centralized ML models with FL models for traffic prediction cases and ran the model on edge devices. As data, they utilized traffic flow, weather, and road geometry. In Liang et al. (2022), a traffic simulator is designed with FTL to guide Reinforcement learning agents’ (FTRL) driving. The proposed framework provides collision avoidance as well. • Smart Transportation: 5.4. Industrial applications Processing industrial data with AI techniques makes possible high-precision location detection, the creation of early warning systems, and the detection of abnormal behaviors. Thus, with smart security, pre-warnings, and postevent analysis can be performed. The manufacturing process comprises modeling, monitoring, prediction, and control stages Ge et al. (2017). The smart industry is a concept about integrating ML techniques into this process. FL is applied to develop models for predictive maintenance in industrial settings in Pruckovskaja et al. (2023). The paper evaluates FL aggregation methods for predictive maintenance and quality inspection in Industry 4.0, ensuring data privacy. Four strategies, FedAvg, FedProx, qFedAvg, and FedYogi, are explored. Data distributions significantly impact FL performance. • Predictive Maintenance: The authors introduce a real-world quality inspection dataset (FLADI) with diverse product variants as clients. FL relies on decentralized training, where clients train local models and share updates with a central server. Targeted sensing devices include machines, production units, sensors, and intelligent factories. FL’s suitability varies based on data distribution and feature heterogeneity among clients. In Han et al. (2019); Liu et al. (2020), authors use FL for monitoring the environment collaboratively using various sensors to detect defects. In Zhai et al. (2021), authors developed an anomaly detection system using FL for IoT devices equipped with an ammeter, water meter, and camera sensors. In Mowla et al. (2019), it is used for attack detection for Unmanned Aerial Vehicles, which can be considered anomaly detection. • Monitoring and Anomaly Detection: FL is used for allocating devices for realtime data processing in robotics. For instance, in Zhou et al. (2018), authors used FL to decrease the communication time, facilitating learning on the device. • Robotics and Industry 4.0: In Liu et al. (2020, 2019), FL is applied for learning an imitation scheme. Each device runs the NN model independently and shares the parameters with the server, which increases the global knowledge knowledge leading to more accurate results. 5.5. Aerospace applications The aerospace industry is equipped with cameras (for visual data), LiDAR (Light Detection and Ranging) sensors (for 3D mapping and obstacle detection), GPS (Global Positioning System) receivers (for location information), and inertial sensors (gyroscopes and accelerometers for attitude and motion tracking), is continually seeking innovative ways to enhance the safety, efficiency, and reliability of aircraft operations. Key to achieving these objectives is the development of advanced technologies that enable real-time monitoring of aircraft health and optimizing critical resources Fu et al. (2023). FL is employed for aircraft monitoring using distributed sensor networks in Li et al. (2022). The study addresses the challenge of intelligent fault diagnosis in aerospace equipment, focusing on bearings as critical components. Data-driven approaches have shown promise, but the scarcity of labelled data and data distribution discrepancies among different equipment hinder model training. • Aircraft Health Monitoring: The authors propose a novel method using Clustering Federated Learning (CFL) with a self-attention mechanism to overcome this. Sensors collect vibration signals from various devices, and the CFL approach clusters clients with similar data distributions. The proposed method achieved superior fault diagnosis results compared to other methods, demonstrating its effectiveness in leveraging distributed data for improved safety and maintenance of aerospace equipment. FL is applied to optimize power and schedule in the aviation industry. The goal is to develop a shared model to improve fuel efficiency, reduce emissions, and enhance flight safety. Airlines or aviation companies train local models using their flight data, and the models’ updates are aggregated to create a global model. • Power Allocation and Scheduling: By collaboratively learning from diverse flight patterns, FL enables optimized flight planning without sharing sensitive flight information. The paper Zeng et al. (2020) presents a novel framework for distributed FL in a swarm of unmanned aerial vehicles (UAVs). Each UAV trains a local FL model based on its data, and a leading UAV aggregates these models to generate a global one, optimizing convergence using joint power allocation and scheduling. The study analyzes how wireless factors impact FL performance, highlighting the effectiveness of the proposed approach. 5.6. Agricultural applications The agricultural applications exemplify how integrating smart sensors and FL revolutionizes the state of the art. By empowering farmers with automated disease detection, crop monitoring, and efficient resource management, these technologies play a pivotal role in maximizing agricultural productivity and sustaining livelihoods in the agricultural sector. FL is utilized for collaborative crop disease detection in Ebenezer et al. (2022). The method involves training machine learning algorithms across a decentralized network of edge devices or servers, each retaining its local data samples. • Crop Disease Detection: By utilizing image datasets of crop diseases and employing convolutional neural networks (CNNs), FL enables the creation of a powerful deep-learning model for disease identification. This approach eliminates the need for costly and continuous professional inspection, making disease detection more accessible and cost-effective for farmers, even in remote areas. FL’s privacy-preserving nature ensures farmers can contribute their crop images without sharing raw data. Implementing FL for crop disease detection promises to enhance agricultural productivity, ensuring food security and economic growth for farming communities and the entire country. Another application area of FL is agricultural production Abu-Khadrah et al. (2023). In this study, authors utilize various smart sensors for agricultural monitoring and control. The specific sensors used include optical sensors, mechanical soil sensors, location sensors, electrochemical sensors, dielectric soil moisture sensors, airflow sensors, and electronic sensors. • Agricultural Production: These sensors are deployed in smart agricultural plots to sense crucial parameters such as temperature, pressure, humidity, soil moisture, crop health (infections and growth level), and climate conditions. The target device for this application is the smart sensor system, which integrates with IoT technologies and cloud computing for decentralized data processing and global actuation. The sensor system, using FL, collaborates with the cloud server and other sensor devices to optimize sensor controls based on past accumulated data. By modifying the sensor operations according to the validated and optimal data, the smart sensors enhance agricultural production and maximize productivity in various agricultural plots. This paper is under a CC 4.0 license. available on Arxiv

Federated Learning Applications on Sensing Devices: Everything From Medical to Environmental

This paper is under CC 4.0 license. available on arxiv Authors: (1) Prajwal Niraula, Department of Earth, Atmospheric and Planetary Sciences; (2) Julien de Wit, Department of Earth, Atmospheric and Planetary Sciences; (3) Benjamin V. Rackham, Department of Earth, Atmospheric and Planetary Sciences; (4) Elsa Ducrot, Astrobiology Research Unit, University of Li`ege; (5) Artem Burdanov, Department of Earth, Atmospheric and Planetary Sciences; (6) Ian J. M. Crossfield, Kansas University Department of Physics and Astronomy; (7) Valerie Van Grootel´, Space Sciences, Technologies and Astrophysics Research (STAR) Institute, University of Li`ege; (8) Catriona Murray, 5Cavendish Laboratory; (9) Lionel J. Garcia, Astrobiology Research Unit, University of Li`ege; (10) Roi Alonso, Instituto de Astrof´ısica de Canarias & Dpto. de Astrof´ısica, Universidad de La Laguna; (11) Corey Beard, Department of Physics & Astronomy, The University of California; (12) Yilen Gomez Maqueo Chew, Instituto de Astronom´ıa, Universidad Nacional Aut´onoma de M´exico, Ciudad Universitaria; (13) Laetitia Delrez, Astrobiology Research Unit, University of Li`ege, Space Sciences, Technologies and Astrophysics Research (STAR) Institute, University of Li`ege & 0Observatoire de lUniversit´e de Gen`eve; (14) Brice-Olivier Demory, University of Bern, Center for Space and Habitability; (15) Benjamin J. Fulton, NASA Exoplanet Science Institute/Caltech-IPAC; (16) Michael Gillon, Astrobiology Research Unit, University of Li`ege; (17) Maximilian N. Gunther, Department of Physics, and Kavli Institute for Astrophysics and Space Research, Massachusetts Institute of Technology; (18) Andrew W. Howard, California Institute of Technology; (19) Howard Issacson, Department of Astronomy, University of California Berkeley; (20) Emmanuel Jehin, Space Sciences, Technologies and Astrophysics Research (STAR) Institute, University of Li`ege; (21) Peter P. Pedersen, Cavendish Laboratory; (22) Francisco J. Pozuelos, Astrobiology Research Unit, University of Li`ege & Space Sciences, Technologies and Astrophysics Research (STAR) Institute, University of Li`ege; (23) Didier Queloz, Cavendish Laboratory; (24) Rafael Rebolo-Lopez, Instituto de Astrof´ısica de Canarias & Dpto. de Astrof´ısica, Universidad de La Laguna'; (25) Sairam Lalitha, School of Physics & Astronomy, University of Birmingham; (26) Daniel Sebastian, Astrobiology Research Unit, University of Li`ege (27) Samantha Thompson, Cavendish Laboratory; (28) Amaury H.M.J. Triaud, School of Physics & Astronomy, University of Birmingham. Table of Links Abstract & Introduction Observation Analysis and Validation Future Prospects & References 2. OBSERVATIONS 2.1. A Candidate in Archival K2 Data EPIC 249631677 was observed by K2 in Campaign 15 from 2017-08-23 22:18:11 UTC to 2017-11-19 22:58:27 UTC continuously for about 90 days as part of program GO 15005 (PI: I. Crossfield). The pointing was maintained by using two functioning reaction wheels, while the telescope drifted slowly in the third axis due to radiation pressure from the Sun, which was corrected periodically by thruster firing (Howell et al. 2014). As a consequence of such a modus operandi, uncorrected K2 light curves can show saw-tooth structures. Many pipelines have been built to correct for such systematics. Two popular detrending algorithms for K2 light curves are K2SFF (Vanderburg & Johnson 2014) and everest (Luger et al. 2016). These pipelines have helped to achieve precision comparable to that of Kepler by correcting for systematics caused by intrapixel and inter-pixel variations. In the case of EPIC 249631677b, the standard deviation of the flattened light curve for K2SFF was observed to be 1230 ppm, compared to 685 ppm for everest. Considering this, we use the light curve from the everest pipeline, available from the Mikulski Archive for Space Telescopes, throughout this analysis. We use a biweight filter with a window of 0.75 days, as implemented in w¯otan (Hippke et al. 2019), to generate the flattened light curve for further analysis, and use only data with quality factor of 0. This light curve can be seen in Figure 1. The simple aperture photometric light curve has a scatter of 2527 ppm, which improves to 685 ppm after everest processing. We searched the flattened data for periodic transit signals using the transit least squares algorithm (TLS) (Hippke & Heller 2019), and found a prominent peak around 3.14 days as can be seen in Figure 2. We assessed the presence of additional candidate signals after modeling out the 3.14-d signal by re-running TLS, but did not find any with a significant signal detection efficiency (i.e., SDE>10). 2.2. Candidate Vetting with SPECULOOS Telescopes We followed up on the planetary candidate by observing with SPECULOOS Southern Observatory (SSO) two transit windows on UT 25 February 2020 by Ganymede and on UT 18 March 2020 by Io, and one transit window with SPECULOOS Northern Observa-. tory on UT 18 May 2020 by Artemis. SSO is composed of four telescopes: Io, Europa, Ganymede and Callisto, which are installed at ESO Paranal Observatory (Chile) and have been operational since January 2018. SNO is currently composed of one telescope (Artemis), which is located at the Teide Observatory (Canary Islands, Spain) and operational since June 2019. All SPECULOOS telescopes are identical robotic Ritchey-Chretien (F/8) telescopes with an aperture of 1-m. They are equipped with Andor iKon-L cameras with e2v 2K × 2K deep-depletion CCDs, which provide a Field of View (FoV) of 12 0 × 12 0 and the corresponding pixel scale is 0.35 00 pixel−1 (Delrez et al. 2018; Jehin et al. 2018). To schedule those windows we used the SPeculoos Observatory sChedule maKer (SPOCK), described in Sebastian et al. (in prep.). Observations were made with an exposure time of 40 s in an I+z filter, a custom filter (transmittance >90% from 750 nm to beyond 1000 nm) designed for the observation of faint red targets usually observed by the SPECULOOS survey (Delrez et al. 2018; Murray et al. 2020). SSO data were then processed using the SSO Pipeline, which accounts for the water vapor effects known to be significant for differential photometry of redder hosts with bluer comparison stars (Murray et al. 2020). SNO data were processed using prose, a Python-based data reduction package, which generates light curves from raw images (Garcia et al. in prep.). It creates a stacked image to extract the positions of the stars in the field, and uses the positions to perform aperture annulus photometry. A differential light curve is produced using a weighted light curve derived from the field stars, while the instrumental systematics, such as pointing shift and full width half maximum, are recorded to assist later in detrending. We show detrended light curves from SPECULOOS in Figure 3, where we recovered the transit events within 1σ of the calculated ephemeris from K2 data. Since these observations were obtained two years after K2 Campaign 15, they improve the precision of the transit ephemeris by an order of magnitude.

π Earth: a 3.14-day Earth-sized Planet from K2's Kitchen Served Warm: Observation

Software Repositories and Machine Learning Research in Cyber Security: Abstract & Introduction 

Authors: (1) LI LI, Beihang University, China; (2) XIANG GAO, Beihang University, China; (3) HAILONG SUN, Beihang University, China; (4) CHUNMING HU, Beihang University, China; (5) XIAOYU SUN, The Australian National University, Australia; (6) HAOYU WANG, Huazhong University of Science and Technology, China; (7) HAIPENG CAI, Washington State University, Pullman, USA; (8) TING SU, East China Normal University, China; (9) XIAPU LUO, The Hong Kong Polytechnic University, China; (10) TEGAWENDÉ F. BISSYANDÉ, University of Luxembourg, Luxembourg; (11) JACQUES KLEIN, University of Luxembourg, Luxembourg; (12) JOHN GRUNDY, Monash University, Australia; (13) TAO XIE, Peking University, China; (14) HAIBO CHEN, Shanghai Jiao Tong University, China; (15) HUAIMIN WANG, National University of Defense Technology, China. Table of Links Introduction Background of OpenHarmony The State Of OpenHarmony Ecosystem Overview Of Mobile Software Engineering The Research Roadmap Discussion Related Work Conclusion & References 1 INTRODUCTION Mobile Software Engineering has been a hot topic for many years. It concerns all the aspects of software engineering in mobile, including the design, development, validation, execution, and evolution of mobile applications. This has been considered extremely important as nowadays our lives have been empowered by the massive increase in the use of mobile apps. Indeed, the number of mobile devices will reach 7 billion in 2023. The number of mobile apps that can be run on each mobile device (for both Android and iOS) has exceeded the 2 million mark. Furthermore, these figures are constantly increasing, thanks to app stores and marketplaces that allow users to effortlessly download and install applications. Mobile platforms are rapidly evolving as well in order to continuously integrate diverse and powerful capabilities, including various sensors, cameras, wireless communication channels, as well as on-device memory and disk capacities. As a result of ingeniously applying these technological developments, developers of mobile software are pushing the boundaries with innovative mobile services and exciting mobile applications. Consequently, due to the rapid development and evolution of mobile software, developers face new software engineering challenges. To address these challenges, researchers in the software engineering community have explored various research directions and developed lots of novel tools supported by formally grounded methods. Indeed, researchers have proposed various static program analysis approaches (i.e., by just scanning the code without actually running mobile apps) for characterizing issues (including ones related to mobile security, compatibility, energy consumption, etc.) of mobile apps [64]. For example, Arzt et al. [7] have designed and developed the famous FlowDroid approach that performs static taint analysis of Android apps for pinpointing privacy leaks. Except static analysis approaches, researchers have also invented various dynamic testing approaches (i.e., by actually running mobile apps on devices) for detecting potential defects of mobile apps at runtime [54]. For example, Amalfitano et al. [5] have proposed an GUI ripping approach for automated testing of Android apps. Su et al. [98] have proposed to achieve the same purpose through a model-based approach. The aforementioned research approaches have contributed to the huge success of the current flourishing mobile ecosystem, including both Android and iOS. Unfortunately, these approaches cannot directly benefit OpenHarmony[1] , which is a new opensourced mobile platform launched by the OpenAtom Foundation after receiving a donation of the open-source code from Huawei. These approaches, theoretically, should be generic and hence should also work for OpenHarmony. However, it still requires significant engineering efforts to achieve that due to the following reasons (more details will be given in the background section): (1) the Openharmony platform empowers a new framework supported by layered architecture, (2) Openharmony apps are written by a newly designed language called ArkTS. Unlike Android and iOS, which have been mature for many years and each has a healthy ecosystem to support their growth, the development of the Openharmony ecosystem is still at an earlier stage. We, therefore, argue that OpenHarmony requires more help from the software engineering research community.[2] We call on actions for conducting software engineering research for OpenHarmony. As our initial attempt, we decide to present to the community an initial research roadmap for guiding our mobile software engineering community to achieve that. We start by conducting a lightweight literature review of Mobile Software Engineering to understand the current achievements. We then conduct a comparative study to locate the technical gaps between mature platforms and OpenHarmony. Based on that, we summarize the technical deficiencies of OpenHarmony and propose a roadmap for our research community to complete. [1] https://www.openharmony.cn [2] This could be regarded as new opportunities for the mobile software engineering community This paper is under CC 4.0 license. available on arxiv

Software Engineering for OpenHarmony: A Research Roadmap

Hi All! Here is my weekly wrap-up email. If you love this content (please share it), but also… Check out my and connect with me on / . Podcast YouTube Twitter If you love this kind of content, please subscribe to my or newsletter. weekly daily What’s in today’s newsletter? Regret Minimization Framework Idea: Uncomfortable Truths Question: Expect Failure Quote: Podcast: Mike Mandell - Principal Attorney at Mandell Law Dr. Emily Bashah - Founder of Bashah Psychological Services Dr. Steven G. Rogelberg - Chancellor’s Professor at UNC Charlotte The Long View Article: Idea: Regret Minimization Framework Before Jeff Bezos founded Amazon, he faced a pivotal moment. He had a comfortable finance job but couldn't shake the idea of an online bookstore. He asked himself: "When I'm 80 years old, what will I regret more - trying this or not trying this?" This question led Bezos to create the Regret Minimization Framework, an invaluable tool for overcoming fear and making bold, long-term choices. Here's how it works: Imagine your 80-year-old self reflecting back on your life's trajectory. What pivotal decisions along the way make you most proud? Which ones do you wish you’d made differently? Use these insights to guide your choices right now. This simple yet profound framework helps pierce through short-term thinking. We often make choices chasing immediate gratification that leads to long-term regret. Think of the companies that failed to adapt - Kodak dismissing digital cameras or Blockbuster ignoring streaming. In retrospect, their choices seem comically short-sighted. But in the moment, steering into uncertainty feels daunting. Here's the key insight: regret often stems more from inaction than failure itself. Bezos knew trying to build Amazon carried risks, but he'd regret not pursuing his vision even more. Better to dare and fall short than die wondering, "what if?" The regret minimization framework builds on several powerful concepts: Second-order thinking - considering how today’s choices ripple into future outcomes Premortem analysis - envisioning your endeavors collapsed, then course-correcting anticipated failure points Surrogate regret - learning vicariously from others' regrets To apply it, reflect on your own past choices to extract timeless principles about yourself. What worked? What didn't? Why? Anticipate future opportunities and ask, "Will I one day regret not doing this?" Get clear on your goals and let possible regrets guide you. Channel your wiser future self - what would 80-year-old you advise doing right now? What bold vision or endeavor would they tell you to begin, even if success isn't guaranteed? Of course, minimizing all regret is likely impossible. But regularly reviewing decisions through this lens leads to bolder choices aligned with your long-term fulfillment. It helps balance risks and rewards instead of defaulting to safety. So simplify the complex by asking, "What will I regret more?" Let possible regrets guide you, not paralyze you. Make choices that your future self will high-five you for. Question: Uncomfortable Truths “What uncomfortable truths about myself am I avoiding?” This question stopped me in my tracks recently. It cuts through the layers of rationalization my ego hides behind. Forces me to confront aspects of myself I’d rather suppress. Parts that feel messy, undesirable, or incongruent with my ideal self-image. We all have inner demons. Behaviors rooted in childhood stories or past traumas. Fears that limit us. Outdated beliefs about who we are. Impulses are misaligned with our values. It’s easier to avoid facing them. But each time we look away, we grant these demons more power over our lives. By avoiding uncomfortable truths, I give these shadows control over my thoughts, emotions, and actions. I become a slave to their whims, acting in ways I know contradict my vision. Several psychological models reveal why we shy away from self-examination: Cognitive dissonance theory shows how we rationalize contradictions to reduce inner tension. Attachment theory traces avoidance to childhood bonding patterns. But the root cause is simple: it’s excruciatingly uncomfortable to admit we’re not who we want to be. Still, hiding from our inner demons only allows them to dictate our fate, as Jung put it. Great leaders know this. They skillfully excavate their own shadows. They courageously unearth limiting beliefs, unhealthy tendencies, and suppress traumatic memories from their subconscious. Only then can they nullify their demons’ control. As Holocaust survivor Viktor Frankl wrote: “Everything can be taken from a man but one thing: the last of the human freedoms - to choose one's attitude in any given set of circumstances.” So, how do we develop this level of fierce self-honesty? First, bluntly ask: What truth about myself have I been avoiding? What impulse doesn’t align with my values? Sit with these insights and unpack their roots. This kind of sincere self-inquiry rewrites neural pathways and transforms self-sabotaging behaviors. Of course, excavating our shadows is intense inner work, requiring ruthless courage and honesty. But once we confront buried aspects of ourselves, we unlock our full creative power. We stop unconsciously sabotaging ourselves. We step into radical self-responsibility. Bottom line: it’s not enough to perfectly analyze business models or market trends. Successful leaders also courageously confront their inner worlds - their self-deceptions, childhood stories, and contradictory impulses. Who are you, really? It’s time to dig deep and find out. Quote: Expect Failure "Success is the ability to go from one failure to another with no loss of enthusiasm.” – Winston Churchill Failure gets a bad rap. We're conditioned to avoid it at all costs. One F on a test, and your parents freak out. Mess up at work, and you get reprimanded. Strike out asking someone on a date, and your friends ruthlessly tease you. We're taught that failure is something to be ashamed of. The opposite of success. A dead end. But Churchill flips that notion on its head. Success isn't the absence of failure - it's the ability to fail again and again with undying zeal. Failure is fertile soil. An essential ingredient for growth. Each fumble and flop provides an invaluable chance to learn. To experiment. To expand the boundaries of what you're capable of. Failure forces you to get creative, determined, and scrappy. The most successful people fail more than anyone else. They plunge headfirst into the unknown, fall flat on their face, dust themselves off, and jump back in. Failure is the feedback that propels them forward. It's rocket fuel for reaching new heights. Next time you "fail," see it as one step closer to success. Let curiosity conquer fear. Reframe the situation into a captivating game and play to win. Study those who have gone from zero to hero and realize that failure paved their way. Wallowing in shame, doubt, and self-pity after a failure leads to decay. Continuing to act with enthusiasm leads to emerging skills, wealth, and fulfillment. Become addicted to progress, not perfection. Thrive off adversity. Let perseverance in the face of failure become an art form. The bigger the failure, the more meaningful the rebound Podcast: 1. Mike Mandell - Principal Attorney at Mandell Law | | YouTube Spotify Apple Mike Mandell is a lawyer, social media influencer, and founder of Mandell Law, a firm that transforms the process of legal advocacy. He is also known as the #1 Lawyer on TikTok, where he has amassed over 7 million followers by sharing legal advice and tips in an entertaining and informative way. He also has a strong presence on Instagram, Twitter, and YouTube, where he posts videos, podcasts, and testimonials about his work and life. He is on a mission to connect with a younger, internet-savvy audience and educate them about their legal rights and options. 2. Dr. Emily Bashah - Founder of Bashah Psychological Services | | YouTube Spotify Apple Dr. Emily Basha is a renowned psychologist, author, and podcast co-host who specializes in clinical and forensic psychology. She has a private practice in Scottsdale, Arizona, where she provides cutting-edge psychotherapy, psychological evaluations, and expert testimony for various legal cases. She has worked on high-profile cases involving domestic terrorism, capital offenses, and first-degree murder, as well as immigration, civil, and family law matters. She has testified in federal, state, and local courts and has consulted for the FBI, the Department of Homeland Security, and the Department of Justice. 3. Dr. Steven G. Rogelberg - Chancellor’s Professor at UNC Charlotte | | YouTube Spotify Apple Dr. Steven G. Rogelberg is a Chancellor’s Professor at UNC Charlotte and a leading expert in Organizational Science, Management, and Psychology. He boasts a distinguished career with over 150 publications addressing topics from team effectiveness to leadership. He is also the author of the bestselling book The Surprising Science of Meetings, which has been featured on various media outlets such as CBS This Morning, NPR, BBC, and Forbes, as well as his latest book Glad We Met: The Art and Science of 1:1 Meetings. Dr. Rogelberg is a sought-after speaker and consultant who has worked with many Fortune 100 companies and served on multiple advisory boards. Article: The Long View I just discovered one of my new favorite writers. Vicki Tan. She was a product designer at companies like Spotify, Headspace, Lyft, and Google, and she’s currently taking a break to bless us with some writing on behavioral science/psychology. She wrote an article breaking down a (in her words) metaphysical design framework to help us design better possible past/present/futures that create the human experience we want for all time. Here are some of the questions she asks (and answers). How do we navigate decision-making processes more effectively? What guides our initial choices in determining what to build? What do we invest our time and resources in? How do we decide what to improve or get rid of? Read it here. Also published . here

On Regret Minimization, Uncomfortable Truths, and Embracing Failure

Hi All! Here is my weekly wrap-up. What’s in today’s newsletter? Dunning-Kruger Effect Idea: Self Compassion Question: Second & Third Order Consequences Quote: Podcast: David Greenberg - Ex-Chair NYMEX ($12 Billion Commodities Exchange) David Greene - Host of Bigger Pockets Podcast How To Unlock Your Inner Polymath (And Think Like Leonardo da Vinci) The Netflix Effect Article: Idea: Dunning-Kruger Effect Have you ever met someone so confident in their abilities, but so clueless in reality? Maybe they thought they were a great singer, but sounded like a dying cat. Or maybe they claimed to be an expert in something but made basic mistakes. That’s the Dunning-Kruger effect. Let me share a study and a story. A Study . A study was conducted by researchers from Cornell University and published in the Journal of Personality and Social Psychology They asked college students to take a 10-question test on American history and then estimate their own scores and the average scores of their peers. They also asked them to rate their own knowledge of American history compared to their classmates. The results showed that the students who scored the lowest on the test (bottom quartile) were the most likely to overestimate their performance. They estimated that they answered 5.8 questions correctly, when in fact they only got 2.6 right. They also thought that they scored better than 68.6% of their peers, when in fact they only did better than 12.5%. They also rated their own knowledge of American history as above average, when it was actually far below. A Story A bank robber named McArthur Wheeler was arrested in 1995 after robbing two banks in Pittsburgh. He was easily identified by the security cameras because he did not wear a mask or disguise. When he was caught, he was shocked and confused. He said that he had applied lemon juice to his face, thinking that it would make him invisible to the cameras. He had tested his theory by taking a Polaroid picture of himself and not seeing his face on the film. He did not realize that he had made a mistake with the camera settings, and that his idea was completely absurd. This is a classic example of the Dunning-Kruger effect in action. Wheeler was so confident in his own logic and reasoning, but so ignorant of the reality. He did not have the skills or knowledge to evaluate his own plan or to learn from the feedback of others. He overestimated his abilities and underestimated the risks. The History The Dunning-Kruger effect is not a sign of stupidity, but of human nature. We all have blind spots and biases that can affect our perception of ourselves and others. The key is to recognize them and to overcome them. David Dunning and Justin Kruger showed us this with an experiment. They asked people to rate their own skills in logic, grammar, and humor. Then they tested them on those skills. The results were shocking. The people who scored the lowest on the tests also overestimated their abilities the most. They thought they were above average, when they were actually below average. Why? Maybe ignorance, maybe arrogance. Why? Here’s a possible explanation, competence is like a mirror. We judge ourselves by what we see. But what we see is not always accurate. It’s distorted by our biases, emotions, and expectations. We think we know more than we do. It’s like looking at a warped reflection. But asking for feedback from others? That’s where you could get a reality check. See, we fool ourselves with false confidence and miss the opportunity to improve. Think about it like bragging about a skill when you should be practicing it. But you can overcome the Dunning-Kruger effect. Try this: Seek honest feedback. Ask for it from people you trust. Listen to it with an open mind. Test your knowledge on topics. Challenge yourself with quizzes, puzzles, and problems. Learn from your mistakes. Admit your limitations. Be humble about what you don’t know. Be curious about what you can learn. One question to guide you all. Ask yourself: How do I know this? Knowing about the Dunning-Kruger effect is just step one. Acting on it, that’s the leap. Here’s a framework to help. The Four Stages of Competence The four stages of competence is a powerful mental model to help you learn anything faster and better and reduce the chance of you falling victim to the Dunning-Kruger effect. Here’s how it works. 1. Unconscious Incompetence You don’t know what you don’t know. You are blissfully ignorant of your lack of skill. You think you are better than you are. This is where the Dunning-Kruger effect kicks in. You overestimate your abilities and underestimate the difficulty of the task. How to escape this stage? Seek feedback. Ask for honest opinions from experts and peers. Don’t be afraid of criticism. Test your knowledge. Take quizzes, exams, or challenges. See how you compare to others. Recognize your limitations. Admit that you have a lot to learn. Be humble and curious. 2. Conscious Incompetence You know what you don’t know. You are painfully aware of your lack of skill. You realize you need to improve. This is where the learning process begins. You start to acquire new knowledge and skills. How to advance in this stage? Set goals. Define what you want to achieve and how you will measure your progress. Seek guidance. Find mentors, coaches, or teachers who can help you learn faster and better. Apply your learning. Practice what you learn. Apply it to real-world situations. Learn by doing. 3. Conscious Competence You know what you know. You are confident in your skill. You can perform well with effort. This is where the mastery process begins. You start to polish and perfect your skills. How to excel in this stage? Practice. Repeat what you learned until it becomes second nature. Aim for deliberate practice. Automate. Make your skills subconscious and effortless. Reduce the cognitive load. Optimize. Find ways to improve your performance. Seek feedback and fine-tune your skills. 4. Unconscious Competence You don’t know what you know. You are fluent in your skill. You can perform well with ease. This is where the innovation process begins. You start to create and expand your skills. How to innovate in this stage? Challenge yourself. Seek new and difficult problems. Push yourself beyond your comfort zone. Explore new domains. Learn new skills or combine existing ones. Cross-pollinate your ideas. Share your expertise. Teach others what you know. Help them grow and learn from them. The Four Stages of Competence is a mental model that helps you learn anything faster and better. It helps you assess your current level of skill and guide you to the next level. It also helps you avoid the pitfalls of the Dunning-Kruger effect by being humble, curious, and open-minded. Remember, it’s not about how good you think you are. It’s about how good you can be. So, the next time you feel confident about something, question it. Are you seeing yourself clearly, or are you blinded by your ego? It’s not about how good you think you are. It’s about how good you can be. Focus on the growth, not the illusion. Question: Self Compassion Here’s a question on how you treat yourself when you face difficulties: How often do you practice self-compassion in your life? Self-compassion is the ability to be kind, understanding, and supportive of yourself when you experience pain, failure, or hardship. It’s the opposite of being harsh, critical, or judgmental of yourself. Self-compassion is not a sign of weakness, selfishness, or complacency. It’s a sign of strength, resilience, and growth. Research has shown that self-compassion can have many benefits: Reducing stress, anxiety, and depression Enhancing well-being, happiness, and optimism Increasing self-confidence, self-esteem, and self-acceptance Improving relationships, empathy, and compassion for others Promoting healthy behaviors, such as exercise, diet, and sleep Facilitating learning, creativity, and productivity Some interesting studies: A meta-analysis of 79 studies found that self-compassion was associated with lower psychological distress and higher psychological well-being. A study of over 2,000 adults found that self-compassion was positively correlated with self-esteem, self-acceptance, self-confidence, and self-worth, and negatively correlated with self-criticism, self-doubt, self-blame, and self-deprecation. A study of over 200 college students found that self-compassion was positively related to academic motivation, performance, and learning goals, and negatively related to academic procrastination, anxiety, and fear of failure. A study of over 300 health care workers found that self-compassion was positively associated with physical health and negatively associated with unhealthy behaviors. A study of over 100 romantic couples found that self-compassion was positively related to relationship satisfaction, trust, intimacy, and commitment, and negatively related to relationship conflict, insecurity, and resentment. However,  even with the data and the benefits clearly researched, many people struggle with self-compassion. They  think that they don’t deserve it, that they need to be perfect, or that they need to punish themselves for their mistakes. Some people might also worry that self-compassion will make them lazy, indulgent, or irresponsible. But these are all myths. Self-compassion is not about avoiding responsibility, lowering your standards, or ignoring your problems. It’s about acknowledging your reality, accepting your imperfections, and taking care of your needs. Self-compassion is not about giving up, giving in, or giving yourself a free pass. It’s about giving yourself a break, giving yourself a chance, and giving yourself some love. This week, try to practice self-compassion in your daily life. Ask yourself: How do I talk to myself when I face a challenge, make a mistake, or feel inadequate? Do I use words of kindness, encouragement, and support, or words of criticism, blame, and shame? How do I treat myself when I feel hurt, disappointed, or frustrated? Do I offer myself comfort, care, and understanding, or do I ignore, deny, or suppress my feelings? How do I cope with stress, pressure, or uncertainty? Do I use healthy strategies, such as relaxation, meditation, or hobbies, or do I use unhealthy strategies, such as avoidance, distraction, or substance abuse? If you notice that you lack self-compassion, don’t beat yourself up. Instead, use this as an opportunity to practice it. Here are some simple ways to cultivate self-compassion: Use positive affirmations, such as “I am doing my best”, “I am worthy of love”, or “I am enough”. Write a letter to yourself, as if you were writing to a friend, expressing empathy, validation, and encouragement. Practice mindfulness, which is the awareness of your present moment experience, without judgment or resistance. Do something that makes you happy, such as listening to music, reading a book, or spending time with a loved one. Seek support from others who can listen, understand, and empathize with you. Remember, self-compassion is not a luxury, it’s a necessity. You deserve to be treated with kindness, respect, and dignity, especially by yourself. Quote: Second & Third Order Consequences “Failing to consider second- and third-order consequences is the cause of a lot of painfully bad decisions, and it is especially deadly when the first inferior option confirms your own biases. Never seize on the first available option, no matter how good it seems, before you’ve asked questions and explored.” - Ray Dalio It’s about the ripple effects of your actions. Successful people don’t just think about the immediate outcomes of their choices. They think about the long-term implications and trade-offs. First-order thinking is easy and tempting. It’s what most people do. Second-order thinking is hard and rare. It’s what sets you apart. Consider Jeff Bezos. He didn’t just see the internet as a way to sell books. He saw it as a way to transform commerce and customer experience. He didn’t just focus on short-term profits. He invested in long-term growth and innovation. He didn’t just react to the market. He anticipated and shaped the future. Second-order thinking is the key to making better decisions. It helps you avoid unintended consequences and costly mistakes. It helps you find hidden opportunities and creative solutions. It helps you outsmart the competition and achieve your goals. How do you practice second-order thinking? Ask yourself: And then what? What will happen next? What are the side effects? What are the alternatives? How will this look in 10 years? Don’t settle for the obvious answer. Dig deeper. Challenge your assumptions. Seek feedback. Learn from others. When you wake up today, think: Are you just skimming the surface? Or are you diving into the depths? Because after all, The quality of your thinking determines the quality of your life. Podcast: 1. David Greenberg - Ex-Chair NYMEX ($12 Billion Commodities Exchange) | | YouTube Spotify Apple David Douglas Greenberg has over 30 years of experience in the financial industry, serving as an executive chair and board member of the New York Mercantile Exchange (NYMEX), the world’s largest physical commodities exchange. He was part of the team that led the exchange to be the first to reopen after the 9/11 attacks. He played a key role in growing the NYMEX from a valuation of $800 million to $12 billion. 2. David Greene - Host of Bigger Pockets Podcast | | YouTube Spotify Apple David Greene is a former police officer turned real estate investor, agent, and author. He is the co-host of the BiggerPockets Real Estate Podcast, one of the most popular and influential podcasts in the industry. He is also the author of four best-selling books published by BiggerPockets including his latest book, the Pillars of Wealth. David has been featured on CNN, Forbes, HGTV, and other media outlets as a leading expert on real estate investing, wealth building, and personal development. 3. How To Unlock Your Inner Polymath (And Think Like Leonardo da Vinci) | | YouTube Spotify Apple In this Lessons episode we explore how to think like Leonardo da Vinci, a polymath who mastered the art of blending disciplines to unlock unprecedented creativity and innovation. Article: The Netflix Effect I’ve been a big fan of Scott Galloway’s podcast podcast for a while, and I think he really did Netflix justice in the written analysis he put together below. He examines how Netflix continues to defy the odds and retain its Iron Throne in streaming. There are some great business / professional / entrepreneur lessons. Some of the highlights: How Netflix's international expansion and content strategy paid off during the writers' strike Why the streamer's continued investment in originals separates it from the competition How the platform is raising prices thanks to its new utility status The threats posed by TikTok and YouTube Read it here. Also published here. Lead image by . King Lip on Unsplash If you love this kind of content, please subscribe to my or newsletter, check out my and connect with me on / . weekly daily , Podcast YouTube Twitter

Breaking the Illusion of Competence:  The Dunning-Kruger Effect 

Something that has always baffled me is the way that fashion trends go around in circles, seemingly at whim and without warning (for those of us who don't attend Fashion Week, anyway). Round glasses have been resurrected from the '20s, mom jeans have made a triumphant return from the '90s, and even shoulder pads have made a brief reappearance in recent years. And yes, I'm dreading the day that fedoras make a comeback. While some of these resurgent trends are pretty whack, a lot of them make sense – and it isn't just the fashion industry that sees the old become new again. Featured in my newsletter today is David Wachs, CEO of . He revived the ancient art of handwritten letters, but with a modern spin... he produces them en masse with the help of 115 custom robots. Handwrytten I'm mind blown! Let's talk about it. From New to Old David Wachs is an engineer and serial entrepreneur who's managed to exit his first company – Cellit – and launch an incredibly successful second company called Handwrytten, of which he's now the CEO. Let's talk about Cellit first. Interestingly, David started this company as a real estate service. "The whole purpose of Cellit was to provide information on real estate – so you would drive by a house, text a number, and get info on the house. And then the realtor would capture a lead. We were kind of the first ones doing that." It slowly turned into something else, though. As their work grew more and more promotional, Cellit branched out into the hospitality space, and then onto big name brands like Abercrombie and Fitch, Toys R Us, and Sam's Club. It would eventually become a full-service mobile marketing agency. Cellit was sold to HelloWorld in 2012 which eventually became Merkle; and with that, David knew it was time to pivot. “I felt that text messaging had kind of run its course. I always felt that way – even when I started Cellit. People are now inundated with 1000s of text messages and emails a day... an average office worker spends about a quarter of their time managing their inbox. It's all just noise at this point.” The Handwrytten Revolution I told you that today's newsletter guest was someone who resurrected an old trend – and that's exactly what David did with his next project after Cellit, which he called Handwrytten. The inspiration struck when David was leaving his old company behind. “When I left, I wanted to send my employees and my customers 'thank you' notes for sticking with me, and helping us grow this business. I didn't want to send them an email – I wanted to send them a handwritten note.” He sat down with pen and paper and started writing handwritten notes. But of course, his hand cramped (we've all been there) and he ran out of stationery. “I thought there must be a better way; a way to scale handwritten note outreach, so that it was as easy as sending an email, but would still be noticed. And it would be easy for me to do. Even better if we could automate the whole process.” Handwrytten was born – and David started the iterative process pretty much the day after leaving his previous company. How It All Works Handwrytten is a software platform where you can use handwrytten.com or their iPhone/Android app, as well as integrate with salesforce.com and HubSpot. The company uses a number of ways to get data in, including Zapier and Integromat. You order the message, stationery, handwriting style, and volume, and Handwrytten does the rest. But don't worry – it isn't run by a factory of workers with impeccable handwriting and carpal tunnel. "What we actually do is we have 115 custom robots here in Phoenix that write the notes out in pen. Currently, each robot is custom-built; I wrote the software that actually does the writing. But they all can write in realistic handwriting, which is the most realistic handwriting available on the market today.” What's more, the company now offers a premium option where you can have your own handwriting style programmed into the robots. I'm absolutely blown away by the tech this guy has developed, and according to David, his robots are top of the range. “There's just nobody that comes close. So we're doing about 5,000 to 10,000 notes a day, on average, for everything from realtors to luxury brands and car dealerships, solar panel installers, etc.” Why Does It Work? So we know how Handwrytten manages to pump out thousands of notes every day – insane robot writers. (They're seriously cool. You can see a video of them in action .) here But my question to David was... why does this work? What's the appeal? If we've moved way beyond snail mail and into the cloud, why are people using a service like this? The Uniqueness Factor “Email is just junk at this point," David explained. "Sending somebody a text doesn't really say 'I care about you.' But when you receive a handwritten note in the mail, it's unique now. What's old is new again, right?” His explanation reminded me a lot of antique furniture. People love handmade things because they're unique – and they were made with time, effort, and care. I also thought about Mother's or Father's Day gifts; it's always lovely to receive a bought gift from your kids, but there's no denying that homemade gifts feel extra special. From a customer perspective, we all want to feel that we're cared about by the brands we buy from – even on a subconscious level. (I don't think any of us truly believe that a brand cares about us as individuals, but it's still nice to feel like they do.) The Spam Factor "When you're going through your mail, it's just bill, bill, bill, bill, bill – but if you see this nice handwritten note, you'll save that. You'll open that last." I definitely second that one. I know I've been way more likely to open and save a handwritten envelope than one that looks like it contains a bill. Plus, as David pointed out, with so much spam and junk mail floating around, getting an actual piece of mail feels special in itself. It's not often you get something physical in the mailbox these days. "Not only will you read it (which is pretty much a 99 to 100 percent read rate, because everybody wants to know who took the time and effort to write a note), but often you keep it.” The Time Factor And of course, time. We're all short on it. We all feel like we don't have enough hours in the day. “Nobody has any time anymore, right? Everybody's responding to 12 emails at once, and their Slack is going off, their Teams is going off, their emails are going off... nobody can sit down to focus and concentrate. So when you receive a handwritten note, you realize time went into it. And that's the gift that you're receiving.” What a beautiful sentiment. If someone takes time out of their day to write you a handwritten note, it really does mean the world – especially knowing how busy we all are. Of course, Handwrytten is in the business of creating that illusion. And I think it's one of the most fantastic business concepts I've heard in a while. It's seriously a win-win scenario; the customer feels cared about, the company gets repeat business, and the world gets a little more handwritten love. The Take-Home: Go Analog or Go Home When the whole business world was going digital, David and his team thought outside the box – or rather, they went back to an old-school way of thinking. And it's paid off in a big way. I'm not suggesting that Apple goes back to making handsets with rotary dials (though I'm sure there are a few of us who would be into that), or that Nintendo starts selling board games again. But there's always room for old trends to be revived in new and interesting ways. Let's look at some examples: Fujifilm revived Polaroid-style cameras with its line of brightly-colored Instax camera kits. Impala Skates relaunched the classic roller skate with a modern twist, and they're now one of the most popular brands in the industry. Depop gave thrifting a major facelift by providing an online platform for people to sell their retro and vintage finds. The list goes on – and B2B and SaaS companies are jumping on board as well. Moo.com brings the classic business card into the 21st century with their stylish and customizable designs. Trello revived the Kanban board, a system that was originally developed in the 1940s using pen and paper. The bottom line? There's always room for old trends to make a comeback – and businesses that are able to tap into that nostalgia are bound to find success. Wrap-Up Hearing David talk about Handwrytten was such an incredible experience for me. It was a throwback to the movie Her – the one where Joaquin Phoenix works for a company that generates handwritten letters based on a person's relationship with the recipient. Well, turns out it's not too far from reality. If you found this whole concept as mind-blowing as I did, make sure you where we discuss Handwrytten, entrepreneurship, pivoting, and more. check out our interview Remember – if you're tired of the new, go back to the old. You might just find your next big break.

David Wachs | Resurrecting an Old Trend

The term 'entrepreneurship' is used so often in the present day. In reality, the vast majority of people aren't entrepreneurs – but it almost feels as though everyone aspires to be one. Being immersed in the business world tends to make you forget that, once upon a time, the entrepreneurs were the 'crazy' ones; the people who took unwarranted risks for a nothing income. Not so in 2022. Now, entrepreneurship is a household word. Everyone knows what it means and aspires to achieve it in some capacity, whether you literally start your own business or simply take an entrepreneurial approach to life and business. In today's recap, we'll be focusing on the former; those entrepreneurs who are starting a literal business endeavor. The first steps are always the trickiest. Like a baby deer taking its first leaps and falls, there's a lot of wobbling and uncertainty. But with the right mindset and tools, you can make those first steps less daunting and pave the way for a successful business future. Evan Carmichael, Entrepreneur and YouTube Success Story Now that entrepreneurs are touted as the courageous success stories they undoubtedly are, we've almost forgotten about the of the world – the guys and gals who forged ahead into the entrepreneurial space before it was cool to do so. Evan was a recent guest on my podcast, and in our interview, we reflected on the early days of his entrepreneurial journey. Evan Carmichaels Evan Carmichael is an entrepreneur, author, and prolific YouTuber. At just 19 years old, he built and then sold a biotech software company. At 22, he was a venture capitalist raising $500k to $15M. He now for entrepreneurs with 3.5 million subscribers, he's written four books, and he speaks globally. runs a YouTube channel As Evan consistently reminded me, however, the road was long to get here – and there was plenty of resistance to deal with in the beginning. “I didn't know anybody who was an entrepreneur; you kind of had to be a little crazy, basically unemployable to be an entrepreneur. But I had a lot of entrepreneurial tendencies. I was selling baseball cards and doing garage sales and all that kind of fun stuff.” Evan started his entrepreneurial career with a business partner in biotech software. Despite the incredibly low income that accompanied the opportunity, he chose uncertainty over a dream job offer from Merrill Lynch. “That was the toughest decision of my life, Scott. I got the chance to go work at Merrill Lynch, my dream job; get paid six figures, travel around the world doing deals, or have three hundred bucks a month and 30 percent ownership in this startup.” It's a conundrum that tends to throw many self-starters off their course – the appeal of certainty, consistency, and a regular income. But for Evan, the allure of entrepreneurship was simply too strong to ignore. “I didn't want to live with the regret of not knowing. And so I told myself, even if it didn't work out, at least I would know. When life is over, and you're finished, and you look back on your life, you're gonna regret not knowing." The road was anything but straight, as you'll hear later on. But this was the pivotal moment in which Evan decided to value experience and learning over a regular paycheck – and to take the plunge into entrepreneurship. Why Do So Many Startups Fail? I've spoken to an incredible number of thought and business entrepreneurs in my time as host of the Success Story podcast. Through hundreds of conversations with brilliant minds, I can say with confidence that one of the big differentiators between successful and unsuccessful entrepreneurs is mindset. Why? Because no matter how intelligent and experienced you are, being an entrepreneur is taking a massive leap on . We know that 90 percent of startups fail; 10 percent in their first year, and then another 80 percent in the years to follow. unbeatable odds There's a common thread that weaves its way through every successful entrepreneur I've had the pleasure of meeting. These people know their purpose; they are so incredibly strong in their vision that they can't be shaken, no matter how many times they're told "no." They also have an unwavering belief in themselves and their team, which allows them to face any obstacle head-on. Evan Believes We're Built to Serve Unsurprisingly, Evan's purpose and passion behind his work as a YouTube thought leader for entrepreneurs is an inspiring one. “I'm a big believer that your purpose comes from your pain; that whatever you struggle the most with as a human is what you want to help other people with. I struggled so much as an entrepreneur myself, so I wanted to make the path easier for other entrepreneurs. It's still what I'm doing 22 years later – trying to help the 19-year-old Evans of the world.” Beneath this specific mission statement is Evan's broader principle that he appears to live by at all costs: he lives to serve. In his work, in his words, and in his actions – serving others is at the forefront. "Every time you take a step up, reach back and help somebody else get to where you're at. Most people are still stuck at zero; most people are stuck just taking the first step. And so they can be inspired by your story. They can learn from you. And knowing that you've been through what they are going through, gives them the hope and the belief that maybe they can get out of the hole they're in." Having the Wrong Mindset We'll move away from the topic of failure and into some advice in a moment – but I think it's really important to dig deeper into this idea of having the right mindset. More specifically, what does it mean to embody the wrong mindset as an entrepreneur? The Pride-Driven Entrepreneur Many of the people who attempt entrepreneurship are hard-headed by nature; they've got a no-nonsense personality, they're Type A go-getters, and they often have a lot of pride. This isn't necessarily a bad thing, but it can be a major obstacle when they hit their first roadblock. The problem with pride is that it often clouds our judgment and causes us to make irrational decisions. We become inflexible in our thinking and unable to listen to feedback or constructive criticism – both from others and from within ourselves. When we're pride-driven, we become attached to our ideas and our ego gets in the way of our ability to grow. We stop learning and progressing, which are two essential aspects of being an entrepreneur. The Self-Serving Entrepreneur I've said this before and I'll keep repeating it; it's okay to want to make money, and people who aim to drive a profit aren't inherently selfish. However, I would caution against having profit as your reason for starting a business. A profit-first approach can quickly erase any good intentions you may have had when starting your business. Growth, figures, and gains become your sole focus, at the expense of your team, your product, and your customers. You'll quickly find yourself in a race to the bottom, as you try to undercut other businesses and sacrifice quality for quantity. The Fear-Driven Entrepreneur This is the entrepreneur who is crippled by fear. They're afraid to take risks, they're afraid to make mistakes, and they're ultimately paralyzed by their trepidation. This can lead to a lack of action, indecision, and in some cases – quitting altogether. The Fear-Driven Entrepreneur often falls into the trap of self-doubt. They doubt their ability. They doubt their idea. They doubt their team. And ultimately, they doubt themselves. Problem is, being an entrepreneur comes with heightened risk. It tends to propagate a lot of fear in people who are unaccustomed to it. The key is learning how to manage that fear, how to work with it, and how to use it as a motivator. In the beginning stages of your entrepreneurial journey, it's critical to be of right mindset – so what's the better alternative to these three wrong mindsets? Evan walked me through some of the advice he's picked up during his own journey. Taking Your First Steps as the Growth-Minded Entrepreneur Entrepreneurs are often painfully aware of their terrible odds. The entire industry is a mix of those that flop, those that thrive, and the ones that face the worst entrepreneurial fate – disgrace. Naturally, then, startup founders tend to begin their journey as being incredibly failure-averse. Not only that, but they refuse to even think about the possibility of failure. This, Evan explained, is the total wrong way to go about your first steps as an entrepreneur. You need to embrace failure like it's your inevitable future. “Expect your first interview to suck. Expect your first video to suck, your camera to break, and the sound to not record. You'll be a nervous wreck. Just expect to suck." Pretty hardcore words for a guy who's all about building young entrepreneurs up – but don't misinterpret his words. Expecting to suck actually means allowing yourself to embrace the growth mindset. "It doesn't mean that you suck as a human. It just means that you don't have the skill set yet.” Don't Set Yourself Up To Give Up This is the growth mindset; the realistic approach to entrepreneurship that acknowledges you still have a long way to go and a lot more to learn. If you assume you're already one of the greats, you set yourself up to fail. “You do your first show, and you realize you're not a B-minus here. You're a D-minus. When you go live, you freeze up; you’re a deer in the headlights. And you just realize that you're way worse than you thought you were going to be.” Unfortunately, setting the bar too high for oneself is all-too-common in the entrepreneurial world, as Evan confirmed. “That's where most people quit; they look back and say, ‘This is terrible. I guess I suck at this.’ And then they never keep going. Expect to suck. Again, it means you're amazing as a human because you tried. And if you keep trying and keep improving, you will not suck anymore.” Evan has learned this lesson from experience; he wasn't one of the greats right out of the gate, and in fact, he didn't see the full fruits of his labor until years later. “My first video, in the first year of it being up, had three comments. One was my mom, one was my sister, and one was some random guy. So I didn't hit it big out of the gate; it took five years or something to get the five figures in subscribers.” Maintaining Momentum Evidently, Evan's growth mindset worked in the long-term (his channel now sits at around 3.5 million subs). But I was curious to ask how Evan maintained his motivation and momentum in the years when he wasn't seeing results. His answer was simple: focus on who you're serving, and not on who you aren't serving. "How do you keep going when you're struggling, and failing, and not seeing the results? People ask me this all the time. What kept me going through those days is this: I always focused on who I was serving instead of who I wasn't. Most people will focus on who they're not serving. And that game never ends." Evan's right. When the people we see as 'successful' are the ones reaching hundreds, thousands, and even millions of people, it's easy to focus on all the subscribers we're yet to gain. “If you were gonna go and give a speech at a library or YMCA or something, and 50 people showed up, you'd be pretty amped. I mean, it's great, right? 50 people showed up to listen to you. But because it's online, we feel like that's nobody and it doesn't matter.” Focus On Your 50 Evan challenged entrepreneurs to think: what about the 50 subscribers you've got now? What would happen if you gave your undivided attention to serving them? Could it be that, for one or two of those people, you actually have the impact you'd hoped to have? "For one of those 50 people, your video is a life changing video. The message is having a major impact; not on everybody, but at least one person in those 50. Maybe your words and the way you said them today, with your tone and your story, actually punctured through where other people haven't. They heard your message.” If you can shift your focus to a glass-half-full perspective, serving the people who are already there and appreciative of your work, you'll be more likely to maintain your motivation and momentum. What's more, the small audience you've got now will lead you to the bigger audience you're hoping for. “That makes you feel like your work is meaningful. And at the end of the day, we all want to wake up and feel like we're going to do something meaningful.” Wrap-Up If you're an entrepreneur who's just starting out, I really want to encourage you in the direction of Evan Carmichael's work. Everything he does is for the purpose of serving entrepreneurs and helping them grow their businesses; he knows exactly what it feels like to start out on rough footing. He's been in the proverbial trenches, and has learned what works. Thanks once again for reading through my weekly wrap. If you want to hear more from the man himself, find our and Evan's YouTube channel . He's got some incredible insights and advice to share – and with a growth mindset, your goal is to soak in all of the advice and wisdom you can get! full interview here here Also Published Here

Stories, News & Mentions about AICAMERA

Aicamera’s Stories on HackerNoon