Artifica builds Artifical Intelligence SaaS solutions for businesses

Artifica

artifica

Artificial intelligence is the imitation of human intelligence processes by machines specifically computer systems. Artificial intelligence came into the picture in mid-1955 by John McCarthy at the Dartmouth conference. John McCarthy and his team members proposed their work on artificial intelligence and a year later in 1956, and the name was attributed to McCarthy. Artificial Intelligence is mounted over the years and today if we look around we are surrounded by the same. From Alexa to google assistant, artificial intelligence made our life much easier than it used to be earlier. A Little Background About Artificial Intelligence Around 1950 Alan Turing published ‘Computing Machinery and Intelligence’ which gave the proposal of simulation Game – an idea which says, And this idea is named as . In 1959 a person Samuel brought out the word which clarifies that a person who develops the program may lose a game to the program itself. The program is capable to win against the creator of their existence. Artificial intelligence focuses on learning, Reasoning, decision making, and the Turing test checks the consciousness and decision-making ability in machines. can machines think? Turing Test Machine learning Types of Artificial Intelligence AI basically has two categories, and the other is . Let’s know a little more about these two. one is narrow or weak AI strong or general AI Narrow or Weak Artificial Intelligence AI has a purpose to act very smartly on a narrow task. AI is used for narrow purposes like feature or . For instance, Siri, Google Assistant, Alexa. face recognition finger-print sensor Strong or General  Artificial Intelligence Strong AI is no less than a human mind and can work as the real human mind does. General  AI is capable of doing every task by thinking like a human. , – the human-robot was discovered by Hongkong, thinks like a human being For instance Sophia Super Artificial intelligence Super strong AI means which is beyond the thinking of human which means a machine can do something which a human can not even think of. For instance, VIKI is a supercomputer developed by USR – a robotic company. VIKI means Virtual Interactive Kinetic Intelligence and Viki is an extremely efficient supercomputer. The latest development is that a company named Hanson Robotics created a humanoid that looks exactly like a human being and the name of that robot is . Sophia is able to recognize facial expressions and voice and can even communicate through artificial super intelligence. Sophia Scope of Artificial Intelligence Natural Language Generation Natural language Generation(NLG) is basically a process that converts data into written language. Its main purpose is to generate content from the data without human guidance. There are many instances where NLG can be used, it can generate business reports based on a dataset. As we know, ECommerce is at its peak and everybody wants different and attractive things. With the help of NLG, it can create landing page content and a description of the product as well. Writing reports manually can be a laborious task for the analysts. Natural language generation, the analysts are able to know clearly without wastage of time. NLG can automate the process with proper insights on the dashboard and speaking to every person on the dashboard. In other words, Natural language generation is an efficient way that would help the analysts by eliminating the workload. In Self-driving Cars Artificial intelligence can work more efficiently than humans do. Elon Musk owner of the tesla group invented self-driving cars that are dependent on artificial intelligence. According to the latest report, he said that in the near future, there will be cars which are level 5 autonomy with no geofence which means that level 5 cars can drive anywhere in this world with no limitations Chatbots Chatbots are recently used by every online platform to take an overview and resolve the queries of the client and give them the best services. Sectors of Artificial Intelligence Manufacturing Manufacturing sectors spent on intelligence ideally because the machines work so efficiently with minimum wastage of resources that keep the companies up in the game. Be it is virtual reality or nanotechnology they always try to keep up the game by investing in this technology. $2 billion Education By the Year 2030, Education is going to be in a different place than it is now, earlier people have to go to get their education at different places like educational institutions, but today there is a lot of quality education available on the online platforms. From personalized classrooms to satellite coaching its all because of intelligent solutions. Education is evolved throughout the years and AI has a great role in providing the best of quality to the potential students and those who are studying can have access to all the content that is widely available in the market. Health The health sector also spends about every year on intelligence solutions.  There are so many machines that can diagnose any disease in a span of minutes which helps in saving the lives of so many people. $1.7 billion Tourism The tourism sector gets a boom by introducing artificial intelligence and the use of chatbots and travel assistance is quite common in use. Tourism had always been an important part of the economy but lately, it has been the most talked topic. Like there are so many travel bloggers who are promoting tourism and artificial intelligence played an important part.  According to a report by , 80%  of travel assistants will be robots by 2020 Travelzoo Technology The tech giants like Apple, Google, IBM Watson are the main leads in artificial intelligence. Elon musk said the self- driving cars are likely to lessen the accidents and a revolution for this era. These Marketers have launched their virtual assistants like Google Assistant, Siri and the latest one which is loved by everyone Alexa. It is so easy for people to work by voice recognition feature by using neural networks to analyze the language and follow the command. Cybersecurity The crime rates on the online platforms are increasing rigorously be it a transaction of money or online abusing or bullying. Trolling these days are quite common. And in monetary terms, the datasheet hacked and the personal documents of the firms are leaked so artificial intelligence plays the part by giving security to the firms. IT solutions spent a ton of money just for the security of the datasheets and important documents that are confidential. This is also predicted that Artificial intelligence will help in employment generation and it will lead to 50% more employment opportunities by 2030. is here to rule the world by giving immense comfort to the people by their intelligence solutions. Artificial intelligence

Getting To Know About Artifical Intelligence

100 Stories To Learn About Remote

88 Stories To Learn About Usa

86 Stories To Learn About Pandemic

278 Stories To Learn About Covid 19

Let's learn about via these 48 free stories. They are ordered by most time reading created on HackerNoon. Visit the to find the most read stories about any technology. Redis /Learn Repo 1. How to Implement Caching Efficiently in NestJS Using Redis In this post, we’ll look at using Redis to give cache capabilities to a NestJS project. We’ll discuss Redis, what caching is, and the implementation procedure. 2. How to use Redis Pub/Sub in Messengers How to use Redis PubSub in messenger apps 3. 3 Simple Steps to Run Redis Docker in CentOS or Linux Machine Let's discuss the easiest way of setting up the Redis Docker image using the Docker container. 4. Redis and Node.JS: Learn the Basics Redis is a powerful piece of technology that lets us store data in memory. Let's look at how to use it with Node.JS 5. Resolving the CROSSSLOT Keys Error with Redis Cluster-Mode Enabled Why "CROSSSLOT Keys in request don't hash to the same slot" error happens in a Redis Cluster, and how to solve it. 6. Migrating from Heroku to Render Migrate from Heroku to Render 7. What is Idempotency and Why We Need It Idempotency is a strategy that makes sure an event that executed multiple times will give the same result. For example, GET method on the HTTP request. If you calling the endpoint with GET method will give you the same result. But POST method is not idempotent because every time you call an endpoint with POST method will create a new record on the database (the database state changes). 8. DBngin: Instant Local Server for MySQL, PostgreSQL, and Redis Docker might be an easy solution as an all-in-one tool for setting up a Local Development Environment. But on the other hand, it eats up a huge portion of your system resources to get up and running, and we hate it when our computer becomes lagging and clumsy. 9. How to Build a Digital Auction SaaS Platform using JavaScript, AWS Lambda, and Redis Want to create your very own digital auction? We'll show you how to build this application by guiding you through each step. Click here to get started now! 10. Live Updates in Ruby on Rails with ActionCable Feature There are times when one as a Ruby on Rails developer wants to implement a real-time feature, like a chat application, and after digging a bit you find a framework's feature named ActionCable, sounds new and scary, right? Fear not, I'll try to explain it as simple as possible so that at the end of this article you'll feel comfortable with the subject. 11. Introduction to ADN: API Delivery Network A CDN designed for APIs with pre-heating and real-time purges, that sits in front of all our API servers, always serving a hit and never a miss. 12. Exploring Redis II: How Redis Stream Works What's Redis stream and how to use it. 13. A different approach to User Sessions in Microservices the most important pieces of data for your whole application is kept in-memory and shared throughout all modules of your application: the user session. 14. Building a Secure Web Chat With Redis, mTLS and GCP Securing Redis installation with mTLS and accessing it from a compute instance on GCP for free. 15. Distributed Denial Of Service (DDOS) Mitigation In the current age of big data and internet businesses it is a constant struggle for an organization to prevent data breaches and protect the data of their users. The current landscape of cyber security is still evolving and attempts are being made to use Artifical Intelligence to protect systems from attacks and data breaches. 16. Event Driven Systems: Building A Facebook Like Counter & Notification System Most of the applications have some sort of badging functionality which is used to display certain counts to the users for CTA (Call To Action). 17. Caches in Python Let's learn about Caches, Caching Operations, Cache Eviction Policies,
Implementation of Cache Eviction Policies, Distributed Caching
and Caching In Python 18. SSL Certificates and Some Other Prerequisites for Installing Orchestrator System Center Orchestrator can automate reoccurring tasks related to support and management of the Exchange organization. In particular, it can be used to create a custom interface where support organization can request precise actions like message tracking logs and mailbox moves. This makes executing basic tasks easy as one does not need Exchange permissions or be forced to use tools like Exchange Management Console or Powershell. 19. Creating a Distributed Task Queue in Python With Celery + RabbitMQ + Redis In this article, we are going to use Celery, RabbitMQ, and Redis to build a distributed Task queue. But what is a distributed task queue, and why would you buil 20. A Successful Selenium Parallel Testing With Zalenium Do you run a lot of automation frameworks using Selenium but finding it difficult to scale with a Dockerized infra? Maybe our case study can be of help. Read on 21. What Is Redis and How Can It Make Your Website 30-40% Faster? Redis is a type of database that can be used to significantly improve your website's loading speed thanks to its design and its versatile selection of modules. 22. We Built The World's Largest Real time High-Frequency, Low Latency Streaming System on Golang The evolution, failures and design decisions behind one of the world’s largest real-time, high-frequency and low-latency streaming systems. 23. How to Choose the Right Real-Time Communication Approach: Long Polling vs Redis Pub/Sub Real-time communication is essential for modern web applications that require fast data exchange between clients and servers. 24. Scale Your Microservices with an Easy Message Queue on Redis If you’re a microservices developer considering communication protocols, choosing an event-driven architecture might just help you rest a little easier at night. With the right design, event-driven architecture can help you to create apps that are decoupled and asynchronous, giving you the major benefits of your app being both performant and easily scalable. 25. Using Redis Streams with NestJS: Part 1 - Setup This is part 1 of a 3-part series, where we will explore how to use Redis streams with NestJS. 26. Exploring Redis: How the Publish/Subscribe Pattern Works Pub/Sub pattern in Redis 27. What Are Redis Hashes Redis hashes are a type of record stored in a Redis database. They are little like JSON objects and store data as key-value pairs. 28. How To Configure Redis + Redis Commander + Docker I will get redis and redis commander up and running using docker. 29. In-Memory Caching in Golang How to implement in-memory cache in Golang App 30. Introduction to Redis: The In-memory Database Redis is a type of database and it can be added to your production level application to make it more performant. I will cover the basics of Redis and show a real world example of Redis. 31. Deploying Complex Microservice Apps Using PMKFT [A How To Guide] In this tutorial we are going to expand our examples with deploying a more complex microservice. The idea is to make you more comfortable with the platform and to show you how you can leverage it for more advanced scenarios. 32. Using Redis Streams with NestJS: Part 2 - Reading from Stream This is part 2 of a 3-part series, where we will explore how to use Redis streams with NestJS. 33. Using Redis Streams with NestJS: Part 3 - Consumer groups This is part 3 of a 3-part series, where we will explore how to use Redis streams with NestJS. 34. Most Common Redis Use Cases by Core Data Structures Redis, short for Remote Dictionary Server, is a BSD-licensed, open-source in-memory key-value data structure store written in C language by Salvatore Sanfillipo and was first released on May 10, 2009. Depending on how it is configured, Redis can act like a database, a cache or a message broker. It’s important to note that Redis is a NoSQL database system. This implies that unlike SQL (Structured Query Language) driven database systems like MySQL, PostgreSQL, and Oracle, Redis does not store data in well-defined database schemas which constitute tables, rows, and columns. Instead, Redis stores data in data structures which makes it very flexible to use. In this blog, we outline the top Redis use cases by the different core data structure types. 35. The Pros and Cons of NoSQL Learn how NoSQL differs from SQL and how it's built for high performance and availability. And no, it doesn't stand for "No SQL". 36. 2019 Database Trends – SQL vs. NoSQL, Top Databases, Single vs. Multiple Database Use Wondering which databases are trending in 2019? We asked hundreds of developers, engineers, software architects, dev teams, and IT leaders at DeveloperWeek to discover the current NoSQL vs. SQL usage, most popular databases, important metrics to track, and their most time-consuming database management tasks. Get the latest insights on MySQL, MongoDB, PostgreSQL, Redis, and many others to see which database management systems are most favored this year. 37. ROLE BASED ACCESS CONTROL (RBAC) DESIGN FOR MICROSERVICES This article extends the principles and design of API gateway discussed in the article here API Gateway. 38. Digging into Postgres's Lesser Known Features Postgres Handles More than You Think 39. How to use Redis HyperLogLog How to use Redis HyperLogLog data structure to store millions of unique items. 40. GitHub Actions CI config for MySQL, Redis, Elasticsearch in Ruby on Rails project with RSpec tests How to run parallel tests with Github Actions jobs for Rails project with MySQL, Redis, Elasticsearch. 41. Django, Celery, Redis and Flower Implementation Here I’m assuming you already have your basic Django project setup. And, already know what Celery is? if not, I’ll suggest getting a basic understanding of it here. So let’s just directly jump into the steps. 42. How To Setup Caching in Node.js using Redis In this article, we are going to implement caching in a node js
application using Redis, but, before we delve into the implementation
details, let’s explore what caching is and how it can help the
performance of our application. 43. Redis Can Help You Deal With Tough Calculations for Geospatial Data Easily Learn how to use Redis and its geospatial indexes and commands for complex geo calculations. 44. Redis Stream vs. Amazon SQS Do you like boxing fights? This evening I bring you a head-to-head battle between Redis Streams and AWS SQS. If you are interested which technology is better and which will end up on the ground, check it up! 45. Let's Build It: Real-time Session Invalidation How would we build an experience like the one above? Demo Repo Some applications need to limit users to a single client or browser instance. This post covers how to build, improve, and scale this feature. We begin with a simple web app with two API endpoints: 46. How to Build a Notification System with Sidekiq, Redis and Devise in Rails 6 In this post, we are going to talk a lot about asynchronous functions, something very fashionable nowadays, but not so much just a few years ago. 47. Should You Be Using NoSQL? NoSQL got quite some hype a few years back. It was going to solve your scaling, uptime, and speed problems. There were trade-offs, of course, but, for a brief moment, seemingly everything we knew about storing and querying data was up for grabs. 48. How to Build a Concurrent Chat App with Go and Websockets Go emerged from Google out of a need to build highly performant applications using an easy-to-understand syntax. It's a statically typed, compiled language developed by some of the innovators of C, without the programming burden of manual memory management. Primarily, it was designed to take advantage of modern multicore CPUs and networked machines. Thank you for checking out the 48 most read stories about Redis on HackerNoon. to find the most read stories about any technology. Visit the /Learn Repo

48 Stories To Learn About Redis

277 Stories To Learn About Microservices

The Lex Fridman Podcast Apart from the many episodes on physics, history and art, Lex Fridman frequently discusses Artifical Intelligence on his podcast. There have been many episodes in 2022 with top AI researchers and experts from other fields discussing AI. Here are my favorite ones: Demis Hassabis: DeepMind - AI, Superintelligence & the Future of Humanity | Lex Fridman Podcast #299 On this episode with Demis Hassabis, CEO and co-founder of DeepMind, Lex and Demis discuss AlphaFold, Solving intelligence and Conscious AI. https://www.youtube.com/watch?v=Gfr50f6ZBvo&t=3364s&embedable=true Andrej Karpathy: Tesla AI, Self-Driving, Optimus, Aliens, and AGI | Lex Fridman Podcast #333 This is one of my favorite episodes of the podcast ever. Andrej Karpathy is the former director of AI at Tesla and a founding member at OpenAI. They discuss Tesla's Data Engine, Google's LaMDA and Autonomous driving. https://www.youtube.com/watch?v=cdiD-9MMpb0&embedable=true Oriol Vinyals: Deep Learning and Artificial General Intelligence | Lex Fridman Podcast #306 On this episode with Oriol Vinyals, the Research Director and Deep Learning Lead at DeepMind, Lex and Oriol discuss AGI, Neural networks and Meta learning. https://www.youtube.com/watch?v=aGBLRlLe7X8&embedable=true Sam Harris: Consciousness, Free Will, Psychedelics, AI, UFOs, and Meaning | Lex Fridman Podcast #185 Together with philospher, neuroscientist and author Sam Harris, Lex discusses the current state and development of AI, aswell as the overall impact of AI on humanity. https://www.youtube.com/watch?v=4dC_nRYIDZU&embedable=true Short version . here The TWIML AI Podcast with Sam Charrington https://twimlai.com/podcast/twimlai/?embedable=true On the ‘This Week in Machine Learning Podcast’ host Sam Charrington interviews AI and ML experts on a large variety of topics and discusses the most interesting current developements in a simple and understandable way. A new episode is released every week. The AI Podcast https://blogs.nvidia.com/ai-podcast/?embedable=true NVIDIA’s Podcast about AI features a large variety of guests ranging from CEOs to scientists talking about how they use or could use AI in their industries and fields of work. I very much enjoy the large variety of topics discussed in the podcast. New episodes are released every other week. Talking Machines Started back in 2015 by Katherine Gorman and Ryan Adams, the Talking Machines Podcast features conversations with experts about AI and ML, as well as their applications in different industries. New episodes haven’t been released since 2021, there are plenty of older ones to listen to though. Listen . here Me, Myself and AI https://sloanreview.mit.edu/audio/?embedable=true Hosted by MIT Sloan Management Review and Boston Consulting Group, the ‘Me, Myself an AI’ Podcast features guests from a large variety of companies, like Boeing, IKEA or Spotify about  applications and possibilities of Artificial Intelligence. New episodes are released every other week. I hope you will check out some of these podcasts / episodes and make sure to comment what your favorites are too! Also published . here

The 5 Podcasts to Check If You Want to Get Up To Speed on AI

What is traditional eCommerce search? eCommerce search is an essential part of the shopping experience, particularly when it comes to text search. Users type in their search queries and expect to receive relevant search results. In a traditional eCommerce site search experience (without any AI technology), a user’s keyword search is matched against a manually configured glossary of keywords. Unfortunately, due to ever-changing fashion vocabulary and the complexity of search queries, traditional eCommerce search still falls short in reaching its full potential, leading to poor matching results and frustrated shoppers. The Current State of Search Technology in 2023 In 2023, the fashion eCommerce industry has evolved to use algorithms like NLP (Natural Language Processing) to further interpret shoppers’ search language, attempting to define intent and meaning behind search queries more effectively. But, the hypothetical value is often not the same as the reality, because of: 1-high prices 2-large data sets needed to train the algorithm 3- the sheer complexity of search queries The two types of algorithm-powered text search models are: Contextual: : Matches user search intent using context clues relevant to the individual user. Search results are specific to a user. In theory Difficult to provide relevant search results, and expensive to run Reality: Semantic: Uses NLP to identify meaning and logic behind search. For example, a “work dress” search is synonymous with “wear-to-work” or “office dress” In theory: Depends on quality of product tags, which may be mismatched.
Too slow to capture up-to-date, modern fashion keywords.
Requires manually updating glossaries. Reality: In 2023, the majority of search technology still requires manual configuration, is expensive, and is still prone to errors. And, it’s not that these search models are bad–but there’s still untapped potential in making text search even better, more optimized, and more relevant. But, how? By using fashion tagging at its core, we’ve created text search that combines different types of AI, unlocking a new, revolutionary approach to traditional text search. How YesPlz Is Solving Common Text Problems in eCommerce Search The most common problems in “traditional” eCommerce search are: irrelevant search results; poor understanding of search intent; and 3) low-quality (or missing) product tags. At YesPlz, we’ve built a comprehensive text search system to solve these problems. Our formula to resolve the problems includes: Laser-accurate AI image tagging, to better create matching, accurate search results, without the need for product descriptions Machine learning that analyzes past user inputs and interactions to predict best search intentions Fresh, updated keywords to keep search keywords relevant and modern Thanks to machine learning, we can customize search results for each of our clients’ specific needs. Here’s the YesPlz antidote to inaccurate, one-dimensional eCommerce text search: Outdated Solution YesPlz AI Solution aka updated keywords with most popular trends and suggestions Fresh, popular keywords Lacks up-to-date, popular keywords. Product discovery isn’t quick or effective. Automatically update with latest keywords based on past user input Autocomplete suggestions No keywords to nudge or guide shoppers. Mismatched search results. Suggested autocomplete keywords, based on popularity or similarity. Instant search results with product preview Search results only show after typing in complete search keywords. High error rates for typos, and unmatched keywords. Updated search results displayed with each keystroke. Additional Filter Paired with Text Search Hundreds of low-accurate matching search results, leading to higher bounce rate. Fully integrated filter alongside search results. Shoppers filter down further, increasing possibility for add-to-cart. Enhanced image tagging Search results limited to matching text descriptions.Huge missed potential for matching results. Key product attributes extracted from image, not product descriptions. Increases percentage of higher matching results. Brand recognition No brand recognition from catalog. Missed opportunity to help shopper discover favorite products from brands. Brand recognition from search, leading to search results that match brands. YesPlz Enhanced eCommerce Site Search Solutions vs. Top Competitors YesPlz’s eCommerce site search solutions are designed to stand out to both retailers and shoppers, from its underlying technology to its customizable features. But, what exactly is the difference between YesPlz eCommerce search and top competitors? Here’s a quick overview of all the features your eCommerce could be unlocking: The impact of added features for text search: Accurate product tags when using enhanced image tagging, building the foundation for personalization Higher conversion rates– Customers who land on an autocomplete page suggestion are than those who don’t. six times more likely to convert Increased average cart size and volume with additional integrated product filter Lower bounce rate with intuitive UX design Fresh updated keywords keep eCommerce relevant and on top of the latest trends More relevant, precise search results when powered by both visual and text AI The power of the above features, when combined, means outstanding text search that always provides relevant, updated results. Before/After: Creating Enhanced eCommerce Search Solutions for W Concept W Concept’s story: A trendsetting online retailer for luxury fashion, W Concept has captivated the attention of shoppers worldwide with a vast selection of clothing, shoes, handbags, and beauty products. But, W Concept had a problem. Its current text search technology didn’t have instant search, lacked autocomplete to make search intuitive and accurate, and had low matching search results. Here’s the before/after snapshot of how YesPlz transformed W Concept’s text search: 1. No automated product tagging, leading to inaccurate search results Before: Product tagging for entire catalog, completely automated by fashion-trained AI, that isn’t reliant on vendor information, increasing the accuracy of matching search results After: W Concept’s robust product catalog also required robust product tags to set the foundation for enhanced text search. No product tags? No problem. With thousands of vendors, W Concept was missing product information, and had some inaccurate information from third-party vendors. YesPlz easily generated these product tags using visual AI. 2. Plain text match searching Before: Rich search suggestions based on both keyword and similar search After: By training the data to understand what other styles resemble a search term, we made W Concept’s search more accurate and robust. For example, if a shopper searched for “tank top,” behind the scenes, search results are based on data that understands the number of similar items to a tank top (including products that don’t have the term “tank top” in their description). A sleeveless shirt, ruched tank top, or square neck tank top still fall under the search query–but wouldn’t be picked up by traditional search unless the terminology was an exact match. In addition, autocomplete offers the best of both visual and text based search, with updated keywords on the most popular, modern trends. No product preview images, not sparking the visual nature of shopping Before: Product previews combined with instant search After: All it takes is a single keyword to show matching results with a product preview, resulting in a much more intuitive and better quality discovery experience. Shoppers needed to click on a search term to filter further Before: Integrated side filter helps shoppers quickly filter down as they search After: W Concept shoppers needed to click on a search term, load a new page, and then were able to filter based on their preferred product attributes. This not only wasted time, but wasn’t providing the most optimized product filtering experience. With an added integrated side filter, shoppers can filter based on their text search (for example, a white tank top text search would lead to further filter options related to brand, silhouette, fit)--with white tank top as a preselected first filter. YesPlz filters are designed to be visual, and UX-friendly, unlike text-heavy filters. By creating a custom, enhanced eCommerce search that solved all of the above problems, YesPlz transformed W Concept’s text search–resulting in always relevant, effective eCommerce site search that converts. Can your eCommerce text search do all of this? When powered by visual and text AI, YesPlz text search is revolutionizing the way shoppers search. If your text search provider isn't doing enough, contact us for a free demo. is next-gen, all-in-one product discovered, powered by artifical intelligence. YesPlz AI Also published . here

Traditional vs. Enhanced eCommerce Search: Unlocking New Text Features

How We Got Here… Before we go deep diving into the new Spatial Web Protocols, HSTP and HSML, it is important to understand what brought us to this point. to enhance our daily lives and productivity . As humans, we are naturally curious beings, always pushing boundaries to new heights of discovery. This basic factor, coupled with As technology evolves, the way in which we interact with it and utilize it naturally demands more expansion and sophistication over time Ray Kurweil’s Law of Accelerating Returns, sets forth the evolution of information technologies on a predictable and exponential trajectory of growth, creating an increasingly more complex interrelationship and interdependency with our way of life. What has this trend looked like over time? The Evolution of Computing Operating systems and computer programs on personal computers. comprised of 3 layers: Computing model Logic — programs based on sets of rules — if, then, etc… Data — user information as a result of interacting with program. Interface — provides usable access point for user to interact with program. Computers After the Web 3 new layers for shared, : networked computing Logic — Website Data Storage — Web Server Interface — Web Browser Smart Phone & Mobile Computing Enters the Scene The 3 layers evolve for : mobile computing Logic — Apps Data Storage — Cloud storage for instant mobile access Interface — mobile screens (phone, tablet, smart TV etc..) Web 3.0 — The Spatial Web : Computing in 3D spaces Logic — AI — Artifical Intelligence Data Storage — Blockchains Interface — AR/VR, IOT, Metaverse, and Digital Twin The progress and expansion of technology increases the capacity of our performance and potential, so we spend about 80% of our time in these new computing spaces. With each evolution of added layers, you gain new abilities to accommodate the growth in technological development. These new technologies will not force you to use AI, but you can, and it will save you time and effort, increasing your capacity for search, creativity, safety, and so much more. You don’t have to use blockchains in the Spatial Web. You can still use the cloud. However, blockchains solve for many of our Web 2.0 problems with security and centralized authority verses decentralized access. Same with AR/VR. These new interfaces are not mandatory. They will merely provide for better, more immersive experiences, with IOT allowing real-time updates and adjustments. Yet, with each progression, we do not lose access to any of the old systems, and the same will be true of the Spatial Web. One fascinating element contrasting this computing era as different from any previous evolution, . Artificial Intelligence, as a virtual assistant, becomes a companion user to every person participating in Web 3.0, whether they realize it or not. the Spatial Web introduces two users at the interface level Trillions of IOT sensors are coming online in these next ten years with data points that AI will interact with to update digital twin environments in real time. , parsing the gargantuan amount of data and information for us to access, use, and navigate through our daily lives. Over time, AI “assistants” will become the dominant user of the Spatial Web Human Connection + Computing Progress = Exponential Network Increase As the computing capabilities have matured and evolved over time, we have also become more connected on every level. From sending mere messages from one computer to another, to sharing and accessing scores of information on websites, to mobile computing, social media, and gaming, each new stage of computing has required updated methods for processing the new types of data and information exchange. , establishing cohesion through common rules and guidelines. Developing new protocols provides the necessary framework for increasingly complex network growth The Evolution of Computing ) 1st Internet (Late 1970’s — Only needed one protocol: TCP/IP Email Only one function — to send a message from one computer to another. ) 2nd Internet (1990 — Websites — New protocol + formatting language — HTTP + HTML, and then Javascript (1995) World Wide Web HTTP gave web browsers the ability to identify web pages and their locations through a unique ID / IP address. HTML provided a common language and set of rules to format the web pages so that web browsers can read them and display the content. Javascript is used for interaction. ) 3rd Internet (Now — Requires a new protocol and formatting language — HSTP and HSML Internet of Everything HSTP gives browsers the ability to link spaces with an ID for every person, place or thing — both digital and physical (virtual or real). Much more than a location identifier — HSTP is a query language and gate-keeper, allowing various parties the ability to agree on who, what, and where anything is in space, and what can be done with it. HSTP allows for query over multiple dimensions, identifying, localizing and updating the state of objects in space and over time. HSTP supports credentialed search of objects throughout all spatial ranges, allowing for restricted access based on any access parameters desired. HSML provides an open standard common language to format all spaces through computable context awareness defining all physical and social dimensions, by recognizing and deciphering policy, rights, meaning and culture, at all levels of embedded systems, incorporating all data and relationships throughout and between them. All context elements get structured into a universal governance graph representing and revealing correlation, interconnection, and interdependencies within. Using measurable context, HSML becomes the foundation for a new generation of artificial intelligence and data sharing. Web 3.0 Requires a New Massive Scope of Global Collaboration Initially the web was constructed as static documents, structured for information display and public access. In 2004, YouTube arrived on scene, introducing a new concept of user generated content, ushering in a new era, bringing social networks into existence, as well as websites like, Wikipedia, granting users the ability to create and update website information. Along with the increased availability and affordability of high-speed connections, the web was suddenly transformed into an interactive platform embodying a new media-rich model of information exchange, Web 2.0. The technologies of Web 3.0, including AR, VR, AI, and IOT, necessitate interaction with each other, along with the user, within multiple environments, whether real world or Metaverse/Digital Twin, allowing for the jumping in and out of each, updating in real time. Web 3.0 moves beyond users interacting with websites. This requires thinking very differently about data connection within and between all objects, circumstances, and people. A new protocol and formatting language was inevitable. To learn more about the Spatial Web Protocol and the evolution of Web 3.0, visit SpatialWebFoundation.org Also published . here

Why the Spatial Web Demands a New Protocol

A primer to understand how technology is poised to disrupt law Note: Below content is informed by online research, reading, and interviewing law practitioners. Resources can be found at end of the article. Background Emerging technologies such as mobile, data analytics, and machine learning had started affecting industries and customers beyond urban consumers and tech-savvy enterprises over the last decade. Retail and healthcare are examples of domains that have benefited significantly from this trend. Yet, there remain industries where technology has only started to a breakthrough in a meaningful way. Law is one such industry. I wanted to identify opportunities for technology to create a meaningful impact and started researching. Being an outsider, I’d to: develop a basic understanding of the field and state of technology adoption understand the ecosystem and market, especially in India’s context This article deals primarily with the first point. A follow-up article may detail out insights from the market. Need for legaltech If you’re inclined to read about the importance of law in society, a Google search is a good enough starting point. Hence, I’ll skip talking about that. What is legaltech then? as the use of technology and software to provide legal services and support the legal industry. It’s similar in spirit to how Fintech and Healthtech have augmented their respective industries and continue to do so. Unlike them, the law has been significantly slow on the uptake. This is due to resistance from a traditionally conservative industry mixed with a lack of understanding and interest from technologists. Wikipedia defines it Conversations about technology implementation in law often drift towards replacing lawyers. That line of thought misses the point entirely. Instead, legaltech’s biggest promises are: augmenting work of law practitioners and subsequently improving their quality of professional life expanding access to legal services and justice for ordinary citizens and small businesses Technology can help lower costs and increase efficiency. This can lead to a positive-sum outcome for all participants. Breaking down law Don’t worry, we aren’t going to do anything illegal. What we’ll do is categorize participants and work in the legal industry. Harry Surden, in a at Stanford Law, identified three types of legaltech users: talk about AI and Law Administrators of law Practitioners of law Users of law In terms of work, a distinction that’s useful is one between the practice of law and the provision of legal services. The defines the practice of law as “the application of legal principles and judgment with regard to the circumstances or objectives of a person that require the knowledge and skill of a person trained in the law.” That’s a mouthful, ain’t it? Simply put, that’s work that requires legal education and expertise. Legal advisory, lawsuits, and negotiations are examples of such work. American Bar Association In contrast to the practice of law, legal services require a low level of legal knowledge. Non-lawyers can perform such tasks too. These include company incorporation, patent filing, or rental agreements. Even amongst practitioners, litigation and corporate law is a high-level distinction. Prologue — technology in law Before moving on to legaltech, it makes sense to look at how non-specialized software is having an effect on the field. The below examples are specific to India and may not necessarily apply elsewhere. The use of video conferencing has exploded over the last few months due to the COVID-19 crisis. Many law firms and corporate teams had started embracing it prior to that. Cisco Webex came up multiple times during interviews. For litigators, it has been an abrupt transition that’s taking some time to get used to. Judiciary is struggling with network infrastructure issues amidst the pandemic. Bandwidth issues plague e-courts and online hearings. The office suite is used extensively for tasks such as legal document drafting. Microsoft teams may be a path of least resistance for those looking to move to the cloud for collaboration. The last point above around the adoption of cloud collaboration software is interesting. It could be an example of a makeshift solution taking hold of the market until more specialized software comes in. Then again, only time will tell. Legaltech in action Legaltech, the way we define it, is concerned with applications that directly affect the practice of law and the provision of legal services. , a database of legaltech companies, comprises 9 categories. Those categories along with some examples are listed below. Stanford CodeX Techindex — Connect various market participants (lawyers, students, users) with each other Examples: , Marketplace Upcounsel Vakilsearch — Help customization and automation of various aspects related to legal documents Example: Document automation Legito — Help streamline processes, optimize workflows, and foster collaboration on an individual and team level Example: Practice management Clio — Organize legal information and make it accessible Examples: , , Legal Research Casetext Manupatra Indiankanoon — Provide legal classes or training to students or practitioners looking for continuous education opportunities Legal Education — Helps parties resolve small cases through online mediation, arbitration, or even without third-party intervention and hence, increase access to justice Examples: , Online Dispute Resolution Sama Presolv360 — Helps a legal party procure information in an electronic format from other relevant parties for purpose of legal proceedings E-discovery — Generate legal insights and predict outcomes from documents and data Example: Analytics Lex Machina — Monitor, track and ensure compliance with relevant regulations Example: Compliance Arachnys An about legaltech business models that I’d come across offers a more comprehensive categorization of legaltech companies. Below is an insightful chart of the same. independent report Not a conclusion There’s a need for stakeholder collaboration on multiple levels for legaltech to realize its promise. Judiciary should be provided with adequate infrastructure. Law practitioners and technologists should work together to build legaltech solutions. Legal education needs to prepare the upcoming workforce for industry developments. If venture capital infusion is any measure, legaltech has in the west. India, on the other hand, is only getting started. Startups in the space are few and far between. The state of tech and is still not as robust. Promising initiatives such as those run by not-for-profit and , India’s legaltech incubator, point to a positive future though. started to take off in recent years data in the judiciary Agami Prarambh The next decade should see technology permeate law in a meaningful way. The COVID pandemic will accelerate adoption. Here’s hoping that this development increases access to justice and improves the life of stakeholders. References and resources - a great resource for keeping up with the global legaltech ecosystem Artifical Lawyer blog - a great resource for those interested in the Indian legaltech ecosystem Agami Youtube channel - a great resource for keeping up with the Indian legaltech ecosystem Indian LegalTech blog Stanford Codex website Artificial Intelligence and Law – An Overview and History by Harry Surden at Stanford Law Business models in LegalTech companies - e-book The State of data in the judicial sector by Civic Data Lab AI in Law and Legal Practice – A Comprehensive View of 35 Current Applications If this topic interests you or if you have an interesting project that we can collaborate on, please feel free to reach out on Twitter , LinkedIn , or mail (akashtndn [at] gmail [dot] com). Previously published at https://www.akashtandon.in/technology/2020-06-16-legaltech-a-10000-feet-overview/

LegalTech - An Overview and Prospective Future

In the current age of big data and internet businesses it is a constant struggle for an organization to prevent data breaches and protect the data of their users. The current landscape of cyber security is still evolving and attempts are being made to use Artifical Intelligence to protect systems from attacks and data breaches. User data is much valued in the shadowy underworld of darknet. There are amorphous hacker guilds that constantly attack internet sites to compromise the security and data. The attacks come in different forms: SQL injection, social engineering and denial of service attacks. DDOS ATTACK One of the most common and extensive form of attack is the DDOS(distributed denial of service attack). This attack is a brute force attack against a site to obtain credentials for a user on the system. The attacker would use the compromised account to launch social engineering attacks on other users or he can sell the details on the darknet. Fig 1: DDOS Attack CHARACTERISTICS OF DDOS ATTACKThese attacks are large scale and are known to leverage cloud capabilities to scale requests and widen the attack area.DDOS attacks simulates user request and spoof the origin IP address and other request information like the headers. It may not be easy to isolate requests that are part of a DDOS attack to a certain IP address(or range) or to a country/region.They cause the application/site to use more than normal computing resource due to increased load. The side effect of this could lead to resource starvation for requests associated with actual user and lead to deteriorated performance of the site or potentially make the system inaccessible to legitimate users.Login and registration modules of a site are the first targets for the attackers seeking to gain unauthorized access to the system. DDOS MITIGATION Its not always possible to completely block DDOS attacks but the impact of it sure can be reduced and its effect mitigated. Fig 2: DDOS Mitigation General Design MITIGATION STRATEGIES 1. Leverage 3rd party WAF (Web Application Firewall) tools. These tools are specifically built to monitor, alert and block certain attack patterns. They act as a reverse proxy to your application/site with added security on top of it. Some well-known WAF are from Cloudflare, Akamai , Incapsula and F5. In addition to DDOS attacks they also help block other attack type like XSS, cookie poisoning, buffer overflow, backdoor entry, stealth commanding etc. 2. Make sure that your services can scale well enough to take the blunt force of the DDOS attack till they can be blocked. Using a public cloud provider will allow you to scale up and down as needed. 3. A true DDOS attack will be difficult for a WAF to block. Mitigating such attack needs more intelligence and data. One can start by monitoring and pulling WAF request logs into their logging/analysis tools like Kibana/Logstash to figure out the URI that is being attacked, the IP address & country from which the requests are originating and the user agent of the request. Looking at these data may help come up with a certain IP address range or country or user-agent patterns which can be added to the WAF rules to block those requests. 4. It is also advisable to add a rate-limiting service in front of your key user facing services(like login/registration modules) to rate limit API requests based on IP address or user agents. 5. Usage of CAPTCHA is also effective in some extent in discouraging the scripts behind the DDOS attack from successfully submitting forms on the site.

Distributed Denial Of Service (DDOS) Mitigation

AI : A Promising Progression for Renewable Energy Sector While it is almost impossible to tell by when, it is for sure that fossil fuels will stop fulfilling our energy needs. The most reliable alternative to this is “renewable energy”, however, even with huge progress, we still have to face some daunting challenges. Amongst all the renewable sources, solar and wind energy are the leading ones. But climate change is making them unpredictable. With major industry players continually trying to cater to the need for cheap, clean and reliable energy, technologies like Artifical Intelligence and Machine Learning are providing solutions for Industries difficulties. In this article, I am going to talk about how AI is contributing its part in RE: Weather Forecast The main challenge connected to RE sources like solar and wind energy is Weather. Be it slow wind speed because of high pressure or no sunlight due to a cloudy day, these conditions considerably affect energy generation. One way to overcome this is storage technology but it’s yet far from where it should be. Energy providers are also getting help from AI to develop software and models for weather change predictions. Google and DeepMind are already doing great progress in this field. They were able to , using their neural network with the weather forecast and turbine data. In the words of Google’s Carbon Free Energy program lead Will Fadrhonc: “This approach also helps bring greater data rigor to wind farm operations, as machine learning and AI can help wind farm operators make smarter, faster and more data-driven assessments of how their power output can meet electricity demand.” improve efficiency by 20 percent Intelligent Energy Storage Without any doubts, storage technology is very helpful to solve problems related to the seasonal nature of Wind or Solar energy sources. But, due to the chaotic and periodic dynamics of these sources on both demand and supply side, battery hardware alone can’t bring true energy values. The situation is so complex that a simple solution can’t be the savior. We need proper control over production-consumption and energy storage and that is only possible through sophisticated software. This is where Artificial Intelligence is required again. By pairing RE with AI-driven storage we can make a change in paradigm. Integration of microgrids: The way of handling distributed energy goes through a broad acceptance of microgrids. Putting the community-level RE generation (in bigger numbers) is creating a challenge to balance energy flow within that particular grid. However, AI-powered control optimization can solve power quality issues and congestion. AI technologies will be ultimately an integralpart of the transition in the future revolutions. Intelligent Multi agent Systems and Adaptive Dynamic Programming (AI algorithms) can provide real-time control over the dynamic nature of the grid below the substation autonomous controls. These AI algorithms can also help to improve system optimization as new devices and new generation sources are integrated. The Other Benefits of AI: AI, Machine learning and Big data are collectively seen as a revolution in the automation of advanced analytics, IT and business processes. In the same way, AI can help the RE industry to gain future goals. According to a report : “48% of energy and utility sector agree that AI is fundamental to their organizations’ success and 46% say their organizations are building AI into the company ethos.” Here are some other benefits that AI is fulfilling for RE and utility industry : published by Infosys Dependability — Ensuring efficient usage & storage of other RE resources, Smart grids and improving operations management Supporting the advancement of smart applications that can autonomously make right decisions based on data and learning. Enabling software applications to examine large data structures, detecting anomalies, identifying patterns and making precise predictions The interface between Artificial Intelligence and RE is an excellent combo. Technological advancements are radically changing our thinking about these two. Though their integration is still in starting days, we can hope for a resilient and Smart Grid. It will definitely open new opportunities while developing its sustainability… towards a better future. <a href="https://medium.com/media/3c851dac986ab6dbb2d1aaa91205a8eb/href">https://medium.com/media/3c851dac986ab6dbb2d1aaa91205a8eb/href</a>

AI: A Promising Progression for Renewable Energy Sector

AGI doesn’t exist yet, but fictional depictions of intelligent machines eradicating or enslaving humanity abound. One startup is hoping it can use blockchain to develop AGI in a more utopian direction_._ What is Artificial General Intelligence? Artificial intelligence or “AI” refers to machines that can learn and perform tasks like a human. “Artificial General Intelligence” (AGI) is something different. It would be like combining all AI devices into a network. This network would be far greater than the sum of its parts because it would be able to generate a artificial general intelligence through the combined experiences of each machine, as well as the interactions of machines with each other. collective Thus, AGI would combine the best of human-like thinking and reasoning with the advantages that computers have over humans — perfect recall and the ability to perform near instantaneous calculations. If AGI could also control robots as dextrous and mobile as a person, it would result in a new breed of machines that could take over any role performed by humans. While AGI doesn’t exist yet, experts agree its development is inevitable. The question is what direction will this development lead? Dystopia or Utopia? We’re all familiar with fictional dystopian depictions of intelligent machines eradicating or enslaving humanity, for example, , and series_._ Not everyone thinks such scenarios are far-fetched. SpaceX and Tesla founder Elon Musk AGI the “biggest existential threat” facing humanity and famous physicist and Cambridge University Professor Stephen Hawking , “the development of full artificial intelligence could spell the end of the human race”. The Matrix 2001: A Space Odyssey The Terminator calls said At the same time, such dire predictions may go too far. An AGI equal to or better than humans could also help develop solutions to problems that seem intractable today, such as climate change and war. Nonetheless, at a minimum, we must worry whether AGI will merely empower the small group of powerful individuals that rule the world today with better means of mass surveillance and population control. Given that most of human history seems to reflect the principles of winner-take-all and might-makes-right, we (humanity) might only have one chance to get AGI right. Blockchain and Decentralized Apps: One Startup’s AGI Solution As overwhelming as the challenge of developing beneficial AGI seems, some innovators are aiming to do just that. is a new startup with the goal of creating a community-developed open-source AGI. Kassy plans to do this by creating a “Knowledge as a Service” (KaaS) Blockchain to function as an infrastructure for a global AI computing system. The Kaasy network will offer a set of kits that developers can use to expand their software applications, better optimize processing times and access load balance computing. The open-source KaaS blockchain code can also be forked to create private knowledge blockchains, for tasks which may involve privacy concerns and other data restrictions. Kassy.ai On top of the KaaS blockchain, Kassy will build an AI algorithm marketplace. The marketplace will be distributed so anyone can use and add new algorithms. The AI algorithms can be optimized to run distributed calculations and power machine learning, thus producing AI skills usable by anyone using the platform. Applications include video processing, machine learning, molecular simulation, AI conversation partners, employee-as-a-service solutions, augmented and virtual reality solutions, and real-time robotic fleet management. The marketplace will also have a built-in compensation mechanism where contributors earn rewards when someone uses their algorithm. Skill rewards for complex skills are issued as a tree, where each complex skill payment is split between itself and the integrated simpler skills. Building Blocks for a Beneficial AGI In the short term, Kaasy’s plan consists of linking a global community of AI specialists within an ecosystem of projects that allows each to contribute to the direction they want AI development to move in. However, the technical features described above support Kassy’s long-term goal of using blockchain technology to make AI accessible in the simplest of applications and devices, which will in turn act as building blocks for a future global Artifical General Intelligence. Subscribe to my Medium and Twitter channels if you want to learn about blockchain and cryptocurrency projects. If you have any questions about this article, please comment in the section below. frauswif. Thank you!

Can Blockchain Ensure Artificial General Intelligence Develops to Benefit Humanity?

Everything You Want to Know About Job Applications

GUI is . We strive to find better ways to program GUIs and we do make progress. The fact that countless GUI toolkits with different ways to develop GUIs exist indicates that the road ahead is still long. If only there was a benchmark to compare GUI programs with one another. Then identifying good ideas would become much easier. programming challenging In this article I present . I reflect on it and look into the future in the second half by pondering how a better GUI programming benchmark could manifest itself. 7GUIs: A GUI Programming Benchmark Retrospective on 7GUIs is my attempt at creating a GUI programming benchmark. In a traditional benchmark competing implementations are compared in terms of their resource consumption. In 7GUIs implementations are compared in terms of their . 7GUIs notation To that end, 7GUIs defines that represent typical challenges in GUI programming ( ). Plenty already exist and could be one of them. In addition, 7GUIs provides a recommended set of . seven tasks live version implementations yours evaluation dimensions The ‘Flight Booker‘ task from 7GUIs The project was conceived as a spin-off of my from 2014. Since then the GUI programming sphere has anything but stopped evolving. Four years later I take a look back. master’s thesis Recent Updates For a longer time I was not happy with the GitHub project’s structure. I was not good at responding to pull requests either. Nonetheless, people continued to bookmark the repository and add implementations. Thanks a lot to all contributers! Then my sent me an email asking if he could help with the pull requests. Altogether this led to my decision to improve the project. : Make its presentation and ownership clearer, and reduce the maintenance effort on my side. thesis’ advisor The goal Instead of using the GitHub Wiki for documentation there is a dedicated website now. The repository (mostly) contains links to implementations but not the implementations themselves anymore. Another addition is a of the 7GUIs tasks in React/MobX so that you can conveniently play with the resulting GUIs in your browser. live version I am happy with how everything turned out but it is still too early to tell if the goals have been reached. In any case, I am open to feedback! Call for Contributions I would love to see more contributed implementations. For example, my Elm implementation is dated and incomplete. Don’t you want to show how awesome Elm handles the 7GUIs tasks? What about implementations in popular SPA frameworks such as React, Ember, Angular or Vue? I think at creating a better implementation than my version in React/MobX. you will fail Prove me wrong! Impact As I was looking back on the project I wondered: Did 7GUIs achieve its goal to improve comparability of GUI programming approaches for framework designers and GUI implementers? To some extent sure. It is hard to give a definite answer though. The amount of contributed implementations, forum discussions and some emails I received suggest that 7GUIs is hitting a need. However, to my knowledge no comparisons have been created. From experience I know how hard it is to write a comparison. Alternatively, there is simply no demand for analyses. Maybe this is all there is to it. On the other hand, I believe that certain aspects of the benchmark could have been done differently to invite more comparisons (see the rest of this article). written down All in all, 7GUIs appears to be useful despite the lack of manifested comparisons but having them would be even better and there are potential ways to make them more likely. Shortcomings During the four years after 7GUIs inception I was gaining much more experience in building GUIs. I noticed that some important GUI programming challenges were not covered by 7GUIs. 7GUIs put its focus more on traditional desktop applications. Like it or not, now more than ever, the most important platform for delivering GUIs to users is the browser and the interconnected web. In addition, I looked at similar projects such as and and tried to identify which aspects make them successful. TodoMVC HNPWA Just to be clear: I strongly believe that . 7GUIs remains useful Still, I also believe that there is potential for a better benchmark. The results of my reflection are presented in the next section. Towards a Better Benchmark If I created a GUI programming benchmark today, what would I do differently from 7GUIs? Photo by on Matt Sclarandis Unsplash Challenges As hinted in the introduction 7GUIs does not cover all relevant challenges that I observed in my professional experience as a software engineer. I will simply list the challenges without deeper explanations that I believe are worth tackling in a more complete benchmark. Please note that this neither is a complete nor a required set but a good starting point for a hypothetical new benchmark: Handling a flaky and slow remote API. Not just read but also mutation operations (e.g. POST requests). This implies handling loading and error scenarios, caching and handling race conditions. Handling loading in a smart way. For example, only showing a loading indicator after a certain amount of time. Showing it for a minimal amount of time to prevent flickering. Showing partial data during loading if it was previously fetched. Optimistic updates Prefetching Change propagation Rendering a large filterable or searchable list of items where results are fetched remotely Undo/redo. Potentially together with remote syncing. Navigation/routing between screens. Preventing navigation (e.g. unsaved changes) Tabbed content Modals. Potentially over several steps like a wizard: Dialog control Pagination or infinite loading Adaptive/responsive layout Constraints (forms) Notifications/toasts. Potentially triggered by remote events or scheduled tasks. State restoration Offline support Two alternative views for the same data Potentially something were Rx is typically suited for: Operations on past events, coordinating multiple interdependent requests. Photo by on NASA Unsplash You might have noticed that many challenges involve communication with a remote API. The world is becoming increasingly connected and this is reflected in the applications that are built and used nowadays. A good benchmark should take this into account. Artificiality For 7GUIs I made the conscious decision to have a set of several isolated tasks. A side-effect of this approach is increased artificiality. In itself this is not a bad thing—but it’s a trade-off. By focusing on a few challenges per task it is easier to see how they are tackled. Also, it should be easier to compare solutions. As pondered above the question is whether this is important. A downside of artificiality is that possible interactions between challenges or approaches to them might be missed. But the biggest downside is that writing a “real” application is simply . If you create something, alongside solving the rather abstract challenges, that you can actually play with and use in practice you are more into it. Ultimately, the more engaging the task the more people would put the more effort into it which leads to a greater exchange of ideas—at least that is my impression. more fun To be concrete, a movie browsing/reviewing application would nicely fit the bills. It is general enough to map all relevant challenges and real enough to be fun. Required Effort If a benchmark is too small it cannot cover enough interesting challenges. It may also encourage very specialized solutions. If a benchmark is too big the required effort for its completion may disincentivize someone from writing an implementation. Now that I have reimplemented 7GUIs recently in React/MobX I noticed that it feels slightly too big. Especially the tasks “Circle Drawer” and “Cells” could have been their own benchmarks. TodoMVC in its original form feels too small. HNPWA, on the other hand, feels significantly too big. However, if the benchmark involves creating a non-artifical application then I believe the increased fun during development can outbalanace a higher required effort. For me the ideal benchmark would therefore lie somewhere between TodoMVC and HNPWA but closer to HNPWA. Specificity of Criteria How specific should the acceptance criteria be? Be too specific and you might introduce bias towards certain solutions. It would decrease fun, too, as you want to be able to “put your mark” on your implementation. Be too lenient and you make it hard to compare implementations missing the point of a benchmark. In general, I feel that you should start by being quite specific and once people start complaining about certain criteria preventing good approaches, relax them. Over time, you should arrive at a good set of acceptance criteria. Measuring Implementations When it comes to having a set of criteria for comparing implementations against each other I am still a fan of the framework which is “an approach to analysing the usability of information artefacts”. Cognitive Dimensions of Notations (CDs) The CDs logo However, I concede that its presentation is dry. I tried to make it for 7GUIs but I feel the dimensions are still far from being easy to apply in practice. more digestible In my mind, having concrete small snippets for each dimension would help. A good and a not so good JavaScript version for a small problem. Then a short explanation why according to the discussed dimension the good version is better than the not so good one. JavaScript On the whole, measuring source code of implementations in a non-subjective, easy and useful manner is a hard problem so I’d like to see further ideas here. Inviting Contributions I noticed that some contributions to 7GUIs did not put much value into their presentation. Which is totally fine as this was never a requirement. Still, having higher demands contributions could actually lead to more contributions. The idea is to make it somewhat prestigious to be listed as a contribution. This in turn should make having your contribution listed more attractive. Plus, comparing implementations against each other should become easier. Some concrete ideas for requirements: Have clear instructions how to run your project, have screenshots, have a complete implementation, have your source code be easily browsable (e.g. host it on GitHub), have an article/document describing what’s good/special about your implementation, have someone else vouch for your contribution, have a jury judge your contribution (“peer review”). It’s hard to find a good balance between being lenient and strict but being stricter might workout better in the end. It should be clear that you must have a high-quality reference implementation: Lead by example! Threadit.Js During my research I recently stumbled upon . I took a cursory glance at it and this benchmark seems to go into the right direction. Threadit.js It is a “real” application. It provides a remote API that is flaky thus forcing implementations to handle error and loading states. It provides utility code so that you can concentrate on the core challenges. It is web-based. Threadit.js has many good ideas and could at the very least serve as a starting point. Unfortunately, the live website is down. http://threaditjs.com/ Conclusion Projects such as TodoMVC, HNPWA and 7GUIs show that there is demand for GUI programming benchmarks. 7GUIs is useful but has its shortcomings. ThreadIt.js is generally close to what I believe is the right direction for a better benchmark. I will not be the one to drive such a project though. In this article I merely envisioned a better benchmark. Nevertheless, I am willing to support an endeavour to create a benchmark as I am convinced that the result will be impactful. What are your ideas?

Towards a Better GUI Programming Benchmark

, connect with me on and are two articles on my Learning Path to Self Driving Cars You can find me on Twitter @bhutanisanyam1 Linkedin here Here Here If you want to read more Tutorials/Notes, please check this post out You can find the Markdown File Here These are the notes for the Course (2018), Taught by Lecture 1 MIT 6.S094: Deep Learning for Self-Driving Cars Lex Fridman Lecture 2 Notes can be found here Lecture 3 Notes can be found here Lecture 4 Notes can be found here Lecture 5 Notes can be found here All images are from the Lecture slides. Deep learning: Set of techniques that have worked well for AI techniques in recent years due to advancement in research and GPU capabilities. SDC are Systems that can utilize these. Instructors are working on cars that understand environment inside and outside the car. devoloping Competitions: DeepTraffic: Deep Reinfocement competition, the code runs in browser. v2.0 now allows Multi-Agent training. Learning SegFuse: Dynamic Driving Scene Segmentation competition. Given the Raw video, the kinematics (movements) of automobiles in the video. Training set gives us ground truth labels, pixel level labels, scene segmentation and Optical flow. Goal: To perform better than the State of the art in Image Based Segmentation. Need: Robots need to interpret, understand and track details of scene. Deep Crash: Goal: Use Deep RL to avoid High Speed Crash Avoidance. Training: 1000 runs to train a car model running at 30mph+ speeds with input from a Monocular camera. DeepTesla: Using large scale networks to train end to end steering, using a monocular video as input for training. Why Self Driving Cars? Goal: Applying Data Driven learning methods to autonomous vehicle. It’s a biggest integration of personal robots. Wide Reaching: Huge number of vehicles on road. Profound: An intimate connection between the car and human. Trusting the Robot with your life, a ‘transfer of control’ to the car. The life critical nature of the system if profound and will truly test the limits of the system. Autonomous vehicle: It’s a Personal robot rather than a Perception-Control. The systems will need assistance from humans via a Transfer of control during situations. A truly Perceptional System having a dynamic nature with human equivalence maybe a few decades away. Cognitive load: A fully connected CNN takes in the raw 3D input to analyse the cognitive load,body pose estimation, drowsiness of the driver. Argument: Full Autonomy requires achieving human level Intelligence for some fields. Human-Centred Artifical Intelligence Approach Proposal: Consider human presence in design of every algorithm. Perception control can handle 90% of the cases. Human Control: Dominant in 10% of the cases. Why Deep Learning? Deep learning Perform really well with huge data. Since human lives are directly dependent on the machines, techniques that learn from real world data are needed. Perception/Control Side Human based collobaration and interaction. What is Deep Learning? AI: Ability to accomplish complex goals. Understanding/Reasoning: Ability to turn complex information into simple and useful information. Deep learning (Representation Learning or Feature Learning) is able to take raw information without any meaning and is able to construct hierarchical representation to allow insights to be generated. The most capable branch of AI that is capable of deriving structure from the data to allow to derivation of insights. Representation Learning Representation is important. Ex: Earth Centered Vs Sun Centred. Cartesian Vs Polar Cordinates to seperate circles and triangles Ex: To Seperate Blue Vs Red Curve using a 1 layer hidden Neural Net. The learning (Using raw input to generate outputs) of the function is achieve using Deep Learning Deep learning improves with more data. Generalisation over Edge cases is the main challenge for Deep Learning. Neural Networks. Inspired loosely by biological human neurons. Human NN: 100 billion Neurons, 1,000 Trillion Synapses State of the Art ResNet-52: 60 million Synapses. The difference is a 7 order of magnitude difference. Differences: Human NN have no stacking, ANN are stacked. No Order Vs Ordered Synchronous Learning Vs Asynchronous Learning Unknown Learning Vs Backprop Slower Processing Vs Faster Processing Less Power consumption Vs Inefficient Similiarity: Both are distributed computation on a large scale. A basic Neuron is simple, connected units allow much complex use cases. Neuron: Neuron consists of inputs of a set of edges with weights The weights are multiplied A bias is added A non-linear function determines if the NN is activated. Combination of NN: Feed Forward NN: Successful in Computer Vision. Recurrent NN: Feed back into itself, have memory. Successful in Time Series related data, much similar to Human (hence are harder to train). Universality: Mutiple Neural nets can learn to approximate any function with just 1 hidden network layer* *Given good algorithms. Limitation: Not in power of the Networks, but in the methods. Categories of DL Supervised Learning: Human annotation of data is needed. Augmented Supervised Learning: Human + Machine approach. Semi-Supervised Unsupervised Learning: Machine Input. Reinforcement Learning: Machine Input. Currently being used: 1,2 Future and better categories: 3,4,5. DL Impact Spaces: Defining and Solving a Particular Problem. ex: Boston Housing Price Estimation. General Purpose Intelligence (Or Almost): Reinforcement and Unsupervised Learning. Supervised Learning Training phases: 1. Input Data 2. Labels 3. Training of Data Testing Stage: 1. New data 2. Inputted to learning system 3. Produce Output Learning Forward pass: Input data is fed into the NN and prediction is generated. Backpropagated: Measures deviation from the expected output and calculates error, adjusts the parameters (hyperparameters) used to predict the values according to the magnitude of error. What can we do DL? One to One Mapping. One to Many Many to Many. Asynchronous Many to Many. Terminologies: DL = NN (Deep Learning = Neural Nets). DL is a subset of ML (Machine Learning). MLP: Multi layer Perceptron. DNN: Deep Neural Networks. RNN: Recurrent Neural Networks. LSTM: Long Short Term Memory. CNN: Convolutinal Neural Networks. DBN: Deep Belief Networks. Neural Network Operations: Convolution Pooling Activation Function Backpropogation Activation Functions Sigmoid. Cons: Vanishing Gradient, Not Zero Centred Tanh. Cons: Vanishing Gradients. ReLu. Cons: Not Zero Centred Vanishing Gradients: When the output or gradient of the NN is very low and results in slow learning. Backpropagation Learning Process of the NN. Goal: Update the weights and biases to decrease loss function. Subtasks: Forward pass to compute network output and error. Backward pass to compute gradients. A fraction of the weight’s gradient is subtracted from the weight. Since the process is modular, it’s parallelizable. Learning Learning is an optimization process. Goal: To minimize the Loss Function by updating weights and biases. Techniques used: Mini-batch Gradient Descent and Stochastic Gradient Descent. Learning Challenge Loss function is highly non linear. Vanishing Gradient. Dying ReLU: Derivate = 0 for 0 inputs. Saddle Points. Overfitting: NN Learns training data but fails to generalise well to real world data. Detected by: Low Training error but high Testing Error. Regularization: Techniques that help in generalising. Create Validation set: Subset of Training Data Early Stoppage: To save a checkpoint and evaluate how is the NN performing on Test Data. Dropout: Randomly Remove some of the nodes (along with the incoming and outgoing nodes) Denoted by Probability of keeping a node (p) Input Node p should be much higher. Goal: To help generalise better. Regularization Weight Penalty: L2 Penalty: Penalised squared weights: Keeps smaller weights unless error derivative is high. Prevents from fitting sampling error. Smoother Model. For 2 similar inputs, the weight gets distributed. L1 Penalty: Penalise absolute weight: Allows weights to remain large. Neural Network Playground: To play around with techniques and practise Deep Learning Breakthroughs What Changed? Compute Power increased. Large Organised Datasets available. Algorithms and Research in GPU Utilization. Software and Infrastructure. Financial Backing. DL is hard Human Comparision: Human Vision: Devoloped 540,000,000 years of data. Bipedal Movement: 230,000,000 years of data. Abstract Thoughts: 100,000 years of data. Neural Net: Add Distortion to pixel data, causes incorrect predictions. Vision Problems: Illuminabilty, Poses, Occlusions, Intra class variations. Object Recognition/Classifcation: Goal: Input an image and predict an output ImageNet: 14M+ categories with 21.8k+ Categories Competition: ILSVRC: AlexNet (2012) made a significant jump in accuracy. Resnet (2015): Human Level Performance was defeated. Subtle example: DL is still distant from ‘Human Generalisation Ability’ Same Architecture, Many Applications: We can change the predict layer to make predictions to as many number of classes as per requirements. Image Classification. Image Captioning. Object Localisation. Image Segmentation. FCNN: Every pixel is assigned a class and it inputs an image and produces another image as output. Goal: Image to Image Mapping. Use-Cases: Pixel-Level Full Scene Segmentation. Colorisation Mapping. Object Detection. Background Removal. Pix2PixHD: Generate High Res Photo Realistic images from semantic label maps. RNN: Work with Sequences Use-Cases: Handwriting Generation Image Caption Generation. Video Description Generation. Modeling Attention Steering. Drawing with Selective Attention. Major Break Throughs Pong to Pong (2012): A step towards AGI. AlphaGo (2016): Learned from Human Expert Games AlphaGo Zero (2017): Beat AlphaGo and Co and it was trained without any Data! (It learned by practising against itself). DeepStack (2017): First to beat Professional Poker Players (in Heads up Poker) Current Drawbacks Defining good Reward Function is difficult. (Coast Runner Examples), The outcomes can be surprising. Lack of Robustness: Adding noise to Pixels cause wrong predictions. Current Challenges: Transfer Learning: Works well with domains closely related. Challenge: Cross-Domain Transfer learning is missing. Reason: Understanding of reasonsing or ability to derive understanding Huge Amount of data needed. Annotated Data is required. Not fully automated: Hyperparameter tuning Reward: Defining a good reward function is difficult. Transparency: NN are mostly blackboxes (even after we visualise the under the hood processes). Edge cases: DL is not good at dealing with edge cases. (Specially when it comes to Autonomous driving). Re: Why DL? It’s an oppurtunity to apply techniques to real world problems effectively. (And DL is the most effective at these). , connect with me on and are two articles on my Learning Path to Self Driving Cars You can find me on Twitter @bhutanisanyam1 Linkedin here Here Here Subscribe to my Newsletter for a Weekly curated list of Deep learning, Computer Vision Articles

MIT 6.S094:Deep Learning for Self-Driving Cars 2018 Lecture 1 Notes

What’s the story? Human culture is ruled by narrative. Everywhere from the board room to the playground, from casual conversation to complex political discourse the story is key. And yet, it seems, we are increasingly at a loss to understand this web of fiction that surrounds us. A few hundred years ago a good proportion of mankind relied on manual dexterity and strength to generate income. Now the ability to turn out a well-structured sentence is of much more economic value than the ability to plough a field, manipulate a loom or blow glass. In the last twenty years this shift toward a knowledge-based society has accelerated further with the internet driving people to spend more and more of their day manipulating and refining the basic building blocks of story. From the CEO who needs to create a compelling narrative to avoid bankruptcy to the twitter user crafting a 280 character tweet; the creation of a good story is at the core of so many tasks today. Yet with all these man-hours spent creating narratives do we truly comprehend the impact this has on our understanding of ourselves? Is this merely a shift in market demand or does it mark some difference in our understanding of the world? Why might this growth in narrative be important and hat importance should we attach to the skill to create and interpret a story? Storytelling as a guide to intelligence One key reason that narrative dominates our culture is that it is a primary method by which we, as humans, gauge the intelligence of others and formulate our responses to complex situations. The leading artificial intelligence theorist, Roger Schanck, discusses the fact that human society has long considered two aspects of narrative as key indicators of intelligence in his book “ ” (Schank, 1999). Here he outlines the proposal that it is not only our comprehension of a particular story that shapes our understanding but also the role that story plays in building conceptual models of others around us. Tell Me a Story: Narrative and Intelligence “We assess the intelligence of others on the basis of the stories they tell and on the basis of their receptivity to our own stories.” Roger Schanck As Wikipedia defines it “A narrative or story is a report of connected events, real or imaginary, presented in a sequence of written or spoken words, or still or moving images, or both’’ — a pretty broad scope. And it is not only human intelligence that we gauge based on storytelling. Going back to the Turing test, one of the earliest attempts to measure machine intelligence, the idea of narrative and discussion is essential. In his seminal paper ‘ ’ (Turing, 1950), Turing discusses not a quantitative measurement of intelligence but an imitation game. Computing Machinery and Intelligence The interrogation he proposes is designed to see whether a machine can fool a human into ascribing it attributes that they would otherwise attribute to a person. Turing makes no mention of measuring intelligence, no empirical certainty, only a dialogue between human and machine. The human judge attempts to decide if the respondee is human or machine based on the sum of its answers. The narrative the machine tells through its answers, combined with the judge’s understanding of that narrative, decides whether the machine is good enough at imitation. If it is only then the human can say that they believe the respondee is intelligent. “Narrative is not a single entity, or a single tightly related set of concepts… Narrative can mean many things.” (Mateas M, 2003) As Turing states, ‘It might be urged that when playing the “imitation game” the best strategy for the machine may possibly be something other than imitation of the behaviour of a man. This may be, but I think it is unlikely that there is any great effect of this kind. In any case there is no intention to investigate here the theory of the game, and it will be assumed that the best strategy is to try to provide answers that would naturally be given by a man.’ This imitation game, of course, is no different from the process (or processes) we use to judge the intelligence of other humans around us. ( ) We are constantly engaged in dialogue and constantly struggle to create a consistent narrative that gives us clues as to the intelligence of others around us. K. Dautenhahn, 1998 However, even if we agree that storytelling is the sole way we measure intelligence there still remains the question of what exactly we are assessing when we say somebody, or something, possesses intelligence. If we read and understand a text are we making an assessment of the intelligence of the author? Or of something else? More on this later. Life in the realm of fiction When presented with a new narrative does everybody approach it in essentially the same way? Is the process of understanding fixed — in the same way that vision is largely fixed by the mechanical processes of the eye? Is it something we can objectively measure? Or is narrative merely a word that covers a , a process that evolves and changes, such that narrative understanding depends only on context and can never be said to be intrinsic. disparate set of responses When we talk about vision there is a certain physical basis that defines what sight is; while people may have hallucinations or other aberrations in their vision, there is a neuro-physical basis that we generally agree underlays what we ‘see’. The mechanical processes that control the cornea, the retina and the optic nerve are fairly well understood and provide a basis or standard framework when talking about sight. Are there analogous processes in the brain when we process narrative? The complexity of human strategies in relation to story-telling suggest there is little hope of any such ‘narrative pathway’. If we accept that narrative, in its broadest sense, is any account that connects a series of events then the scope becomes impossibly vague. Modern neuroscience of course does not have time to worry about such assertions and instead tells us that we can safely ignore most of the stories we spend time telling ourselves since the self — the central actor in most of our stories — does not exist. Research increasingly suggests that the brain will go to fantastic length to create fictional explanations for our actions that have little to do with the real causes at a chemical level. From an empirical perspective the self simply does not exist. It may seem that our sense of self is deeply ingrained but science is pretty unequivocal — the evidence shows that in many situations the self is a fiction that we are each continuously rewriting. From a neuroscience perspective, stories or narrative in its broadest sense, seem nothing more than a curious byproduct, something to be discounted from empirical consideration. There is no “I” or “we” except for in the made-up stories we tell ourselves. But this of course is not how anybody actually lives.While the neuroscientist’s perspective may be empirically consistent it ultimately fails to capture the real experience of self and so can make it that much harder to understand what we talk about when we use the term intelligence in everyday speech. If when we use the word ‘intelligence’ we are referring to a property of a fictitious entity then it is not certain we will ever find it on a brain scan. It seems clear there are certain empirical criteria that are not applicable to fictional entities. Try measuring the average body mass of a unicorn? Or determining the chemical composition of Aladdin’s lamp? If the self is to be classed as a fiction — if “I” is simply a story that we retrospectively create— then it must follow that “we” exist in a world of fiction. If that is the way that the real and unreal are demarcated so be it but when we talk about the realm in which “we” exist we must accept that it is not wholly the mechanistic realm. The role of technology in shaping narrative For most of human history our understanding of narrative was reliant solely on body language and aural interpretation but with the introduction of the first tools that allowed people to inscribe letters onto papyrus we have steadily changed to a culture that is closely entwined with technology. In the last two decades this shift has moved into overdrive as we now scarcely imagine consume any narrative that is not in some way mediated by technology. From the earliest uses technology has provided a crutch for our (human) understanding, and consequently expectations, of intelligence. Four hundred years ago knowledge could be equated with intelligence fairly simply. In the 1,600s, the fact that someone had gained specific knowledge on a topic (say the operation of a steam engine) implied they had invested significant time in understanding the problem space. Now any problem space can be accessed very easily via Google such that knowledge on its own is generally insufficient to claim intelligence. A five year old can easily find and regurgitate a compelling explanation of Einstein’s relativity from the internet but we would be right to hesitate in calling the child intelligent on this evidence alone until we had probed a little deeper. We have seen circumstantial evidence that technology affects and we understand stories in much compressed forms. Nowadays we tend to furnish our understanding by searching on Google rather than spending time puzzling over meanings. A side effect of this reduced attention span is the result that we cross reference our stimuli many times more frequently than previous generations. our attention span If we are reading an article, and a thought occurs to us, it is unlikely we will wait long before jumping to a new tab and searching for information on that point. If we are reading a novel length work it is highly unlikely that we will get through it in one sitting without referencing other material that will influence our ‘reading’ of the text. One can say that this interplay is simply a diversion that comes about because of the easy availability of resources, but it has a much more profound effect on the way we understand. We assume that the way we understand narrative may be affected by our knowledge but that the way we consume it has relatively little impact. We may read The Odyssey via an ipad rather than a papyrus, or hear a human voice relayed through a cinema sound system rather than in person, but we assume that the story itself is essentially the same as it has always been. After all is the difference in consumption important? Surely it is the message that is important in any given text? If we read Homer’s Odyessy on an ipad — while sporadically checking Wikipedia, googling words and browsing beach holidays in Ithaca — then the process is radically different from anything that was possible even fifty years ago. Certain key points may remain with us, certain universal messages, but it is by no means certain that we understand these in the same way as we always did. We are used to an almost constant dialogue with search engines, inserting questions and queries into every narrative to better understand it and test how much we trust it. Homer In reality our approach to understanding story is, and always has been, evolving. The way we read today is drastically different to the way Oscar Wilde read and that in turn is drastically different to the way that Aristotle read. When we are told a story in spoken language, most people directly realise the importance of the delivery. The way somebody tells a story is certainly very important to our understanding of both the narrative and the storyteller’s intelligence. If Shakespeare’s ‘Hamlet’ is performed by idiots not only do we judge the intelligence of the performers but the words themselves lose their power to capture our attention. A poorly delivered version of Hamlet will appear little better than a poorly performed play by a third rate playwright. Will we recognise other intelligence? When we now turn to consider the case of artifical intelligence or indeed the idea of an artificially intelligent singularity the question of narrative becomes much muddier. The idea that computers and machines will reach a level of complexity that exceeds human intelligence and beyond this point will be able to design ever superior versions of themselves, rapidly accelerating the rate of intelligence far beyond what we can imagine raises a number of questions. If we accept that narrative is a key indicator of intelligence then it is reasonable to ask what kind of dialogue we might have with such a machine that could convince one way or the other. How might we notice that it has equalled or surpassed our own understanding. Is this something we, or any of our descendants, are equipped to recognise, even in principle? “One thinks that one is tracing nature over and over again, and one is merely tracing round the frame through which we look at it… When philosophers use a word — “knowledge”, “being”, “object”, “I”, “proposition/sentence”, “name” — and try to grasp the of the thing, one must always ask oneself: is the word ever actually used in this way in the language in which it is at home?” essence Ludwig Wittgenstein We face a tricky problem when it comes to recognising any potential singularity event in that our understanding of narrative is increasingly entwined with the technology we use to consume stories. What are the implications for recognition of a super-intelligent machine when to interrogate it we rely on the same technology that we use to determine if its stories are trustworthy? When the subject is a human this is a fairly straightforward task, and something that we do on a daily basis. However we have seen that the neuro-physcial description rarely gives us insight into this. Whenever we meet a new acquaintance or colleague we sub-consciously create a representation of their intelligence based on the stories that they tell (and the stories others tell about them). We may check their Facebook page, read a blog post they have written, listen to them recount the joys of a recent trip to Paris and all of this we seamlessly join together to form a mental representation of this person, a sense of their intelligence. We do not generally get useful information about their intelligence from a brain scan. When it comes to a machine the situation is more complex. If we seek to understand the narrative that it creates, we can ask it questions, listen to the stories it spins, look at the code it is based on, but the modern machines we interact with are in one sense a single machine, co-joined by the internet. If intelligence is a conversation whereby we create a mental picture of the other then the process of asking if a machine is intelligent is inherently dependant on our knowledge, or lack thereof, of its goals or intentions. If in turn our level of understanding is in part dependent on information generated by the machine, can we truly have an objective representation of its intelligence? We can interrogate the programming language or try to understand its physical processes but, as with the human brain, a scan might explain mechanisms but it does not feed our sense of intelligence. Imagine asking a human outright if they are intelligent — whatever their answer it tells us nothing about the real situation. A person may reply ‘Yes, I am intelligent’ or ‘No, I am not intelligent’ but they may be joking or obfuscating or self-deprecating. We build our mental picture based on experience. To put this another way, we cannot expect to see a super-AI emerge because any intelligent machines we build will be explicitly designed for, and intricately linked to, the systems we use to consume narratives. “No man is an Iland, intire of itselfe; every manis a peece of the Continent, a part of the maine” John Donne They will only become more deeply embedded with us not more distinct. It may not be possible to recognise an intelligent machine because it is not something anybody would, or could, design to exist on its own. To understand narrative in the future we will be increasingly reliant on machines to spread and parse the stories we share and as a result the very concept of intelligence will become less and less meaningful outside the context of this machine infrastructure. Thank you for reading. If you enjoyed this, please consider sharing, or follow on Twitter . For reprint or licensing enquires please get in touch via: lochlanbloom.com .

Stories, News & Mentions about ARTIFICA

Artifica’s Stories on HackerNoon