Our target markets are brands and businesses of tomorrow, we are here to take those businesses to next level without burning a hole in their pockets, and give them a relaxing work journey rather than the stress of hiring, and managing the workforce

Just Built It

What we learned building our all-in-one cryptoasset trading platform If I could give one piece of advice to someone entering the market (any market) with a new product, it would be to first go and talk to your target audience (assuming you’ve narrowed in on one group). You may think that a SWOT, PESTLE or other frameworks are going to put you in the lead but you would be so wrong. To show you what I mean, I thought I’d share our , from our time for our new cryptoasset trading platform, . If you’re hearing about us for the first time, we’re a Hong Kong-based Fintech startup called FinFabrik and we’ve been building capital market software for close to two years now. Our company is made up of a mix of ex-bankers, consultants and developers. If you come from one of these ‘old’ worlds, chances are that you’ve not spent much time worrying about user experience design. It did take time for all of us to lose our biases, but when we did, WOW did our eyes open. So as a starting point I would say this, if you have to run some basic exercises before you take your product to the market, they should be surveys and interviews. Now I’ll explain why. learnings researching “CryptoFabrik” 3 critical problems we learned and are trying to solve From 2015 to 2017, the number of bitcoin wallets increased from roughly 3 million to around 25 million (not all of them are currently active). That’s quite a jump. Everyone and their dog was investing in this market. In spite of these staggering numbers, a challenge for us was finding individuals that were regularly investing or even tracking their portfolios. Most individuals that we met had made one-off purchases after getting a recommendation from a close friend or family member and then continued to hold that single currency (while the going was good anyway). In most cases, this currency was Bitcoin. So it was rather hard to get actionable insights from users that had made one-off purchases. We did eventually meet a lot of users that were not just bitten by the crypto bug, but smitten by it. When we started asking the right questions, we got some great information and insights. We identified certain problems that seemed like very very simple fixes, but so far nobody was trying to fix them. Lack of automated portfolio tracking On average, crypto investors said that they had accounts on at least 3 different exchanges. Additionally, whenever a coin listed on a new exchange, they had to open a new account to transact at an additional venue.. Tracking the prices, holdings, and their available balances across different exchanges is a major pain point for crypto investors. The majority of users that we spoke to were using spreadsheets to track their investments and performance. So our first solution is straightforward. We created one centralised trading screen, where investors could transact on and track all of their exchange balances. This integrated screen allows users to discover, compare and analyse price differences and order books between their existing exchange accounts in real time. By creating a centralised holdings page, we eliminate the need for users to pull in data from each exchange, run multiple spreadsheets and calculate manually. Time wasted on endless KYC to find liquidity 58% of our users said that one of their biggest pain points was account opening with multiple different exchanges. This shouldn’t come as much of a surprise. People don’t particularly enjoy undergoing KYC multiple times as it is a cumbersome activity that takes time away from doing real research and trading. To help solve this problem, we developed a feature called ‘CryptoFabrik guaranteed execution’. Rather than our users spending time on cumbersome account opening processes, the CryptoFabrik team will undergo verification process with new exchanges daily. This will allow the users to gain access to liquidity from multiple exchanges through a single trading venue. Additionally, our platform deploys a smart routing algorithm that will execute trades at the best prices amongst our existing list of exchanges, so that THE USER will no longer need to worry about price sourcing. Transfering crypto between exchanges IS hard Close to 60% of the users said that transferring cryptocurrencies between different exchange accounts impeded their trading activities. Wait times resulting from delays in transfers between exchanges often lead to missed opportunities. Although this isn’t the easiest problem to solve, our product team is working on a solution. The gist of it is to allow CryptoFabrik users to transfer balances between their different exchange wallets from a centralised location. We will uncover more of this capability later. What next? Those are just a few examples of how simply asking the right questions can give you a great start on your building the product roadmap. After we listened to all these grievances, we identified a clear market opportunity. We went ahead and did something crazy — we just built it. That’s it. No ICO’s, no whitepapers, no roadshows. Just a team of crazy people that like to get shit done. We demoed the product at the recently held and and got great responses! FIX conference HKTDC Entrepreneur day We decided to start with something ‘super simple’, if you’re wondering why, its so that we could solve the most basic pain points to start with and take it forward from there. This marks only the beginning of our journey. Next, we will get on to solving problems for businesses. With additional research, we will discover the specific problems faced by institutional investors such as portfolio managers, hedge funds and even crypto brokers. Over the next few weeks, we’ll share a bit more on our findings, as well as our recommended solutions to the problems. Stay tuned!! About me I’m a Product Manager at FinFabrik, get in touch if you have any questions about our research, want to talk about user experience, cryptocurrencies or complain about the Last Jedi.

UX in the Crypto World

StackOverflow, the most commonly used platform by software developers for programming support, has been through a rough ride lately. If you haven’t used StackOverflow before, it’s a / -like Q&A forum where you can ask programming-related questions. It’s been several years since I wrote production-quality code but back when I did, StackOverflow was incredible. Quora Reddit For example, if you ran into the most obscure of errors while compiling your code and got an error message you could not make sense of, you would stick it into Google search. More often than not, you would find a StackOverflow page where someone had asked the same question and got an answer. Less often, you would find another soul who had the exact same obscure problem as you but got no answer - in which case, good luck. More precisely, on StackOverflow are answered, which is pretty impressive. 69% of questions Recently, however, StackOverflow’s traffic has been in decline. Similarweb’s data shows that their traffic year over year (StackOverflow says it’s ). Nevertheless, the trend is downward and is explained primarily by the emergence of AI coding products like ChatGPT and . These products have meaningful code-writing capabilities and are therefore able to provide programming support, at least partly as good as StackOverflow does. Ironically, several of the large language models (LLMs) behind these AI products were trained using scraped StackOverflow data. dropped 14% closer to 5% GitHub Copilot The company has gotten pretty harsh media coverage with these developments. Business Insider, in their article , wrote: Death by LLM Welcome to the future of the internet in an AI world. Online communities like Stack Overflow and Wikipedia thrived as hubs for experts and curious browsers to come together and share information freely. Now these digital meeting places are being pillaged by big tech companies prowling for human data to train their large language models. The new products emerging from this generative-AI boom are putting the future of these online forums in doubt. The chatbots answer questions clearly, automatically, and often pleasantly — so humans don't need to deal with other humans to get information. In the midst of all this attention, StackOverflow has played a steady hand and articulated its two-pronged approach to addressing this challenge: A few weeks back, they announced that they will start charging large AI developers who use the platform’s 50M+ questions and answers for model training (we dug into this issue in the [data scraping](http://Data Scraping in the Spotlight) article earlier) Last week, they launched the product, which is a set of actually useful generative AI features that can help kick off their second innings - we will focus on this today OverflowAI In this article, we’ll dive deep into: AI code writing tools disrupting StackOverflow What OverflowAI does How OverflowAI accelerates the future of StackOverflow AI code writing tools disrupting StackOverflow There are several AI code writing and editing tools available in the market today. These are either independent products (like OpenAI Codex, ChatGPT, Google Bard) or products that are natively integrated inside existing platforms (like GitHub Copilot, Replit Ghostwriter, Amazon CodeWhisperer). They have a broad range of capabilities including code generation, code editing, autocomplete, and debugging. The products that have native distribution (like GitHub Copilot) are at a large advantage because they can operate seamlessly within environments that programmers already use today, and we will see more products attempting to get plugged into existing environments. For example, CodeGPT has a plugin that lets developers use the product from (a popular code editing tool). within Visual Studio Code Existing AI code-writing tools are good are certain tasks. For example, this captures feedback from several web developers about GitHub Copilot - the overarching theme is that the product is useful in a subset of situations where developers have to write net new code and don’t want to spend time writing from scratch. Even in those situations, it’s often hit or miss. Reddit thread The reason is not surprising. Conceptually, ? It’s essentially calculating the probability of a word following another, and generating output based on that. Despite this construct, given the amount of data that’s gone into training these models, the results for the more general ChatGPT use cases (like drafting an email or summarizing a page) have been nothing short of impressive. But it’s important to remember that language models, by design, have limited analytical / math capabilities. In other words, when you ask the model “What is 2+2?”, it may give you the right answer - not because it knows math but because it has seen that text pattern before in its training data. large language models (LLMs) take in a ton of data and generate output on the basis of this construct: in a particular context, for the question you asked, what is the most likely word/text to follow the previous word Similarly, when it comes to code generation, the but is predicting results based on its training with a ton of text data. The consequence of this is the GitHub Copilot feedback above - it is sometimes good at generating the base code you need, but . This will get better over time but it’s hard to say if it will ever get to the point of high accuracy / high reliability. model does not really “know” the underlying concepts behind programming its ability to actually understand code, debug and provide you explanations is limited StackOverflow CEO Prashanth Chandrasekar : describes it succinctly One problem with modern LLM systems is that they will provide incorrect answers with the same confidence as correct ones, and will ‘hallucinate’ facts and figures if they feel it fits the pattern of the answer a user seeks. At some point you’re going to need to know what you’re building. You may have to debug it and have no idea what was just built, and it’s hard to skip the learning journey by taking shortcuts. This is the opportunity for StackOverflow - their traffic drop may be permanent and it’s very likely that programmers come to StackOverflow less often for simpler questions (eg. they might not visit StackOverflow anymore for an off-the-shelf sorting algorithm). But is: 1) providing to more complex questions that language models might not have the capability to answer, and 2) providing that the models have not had previous data to train on. OverflowAI is designed to directly tap into this opportunity. where the product can shine high accuracy / high-reliability answers answers to questions in new technologies/problem spaces What OverflowAI does There are three key facets they are betting on - direct answers to questions, usability from within developer environments, and supercharging knowledge within enterprises. provides (similar to ChatGPT), but provides several . Besides helping create trust, this also provides users with the opportunity to go deeper where the answer provided by AI does not fully solve the user’s problem. This strikes the delicate balance of giving a direct answer when the question is simple, but also guiding the user along a more exploratory path for difficult questions. OverflowAI Search direct answers to users in a Q&A format links to actual StackOverflow posts If the user is not satisfied with the responses, they can enter a . If none of the answers are satisfactory, they can ask StackOverflow to , ready to be posted to the Q&A forum. This experience also saves users from the semi-often situation where the question they ask is already answered previously. chat-like interface to ask follow-up questions draft a question on their behalf The product also doubles down on usability by making all of this capability . This helps StackOverflow compete more effectively with natively integrated coding assistants by letting developers get answers from within their coding environments (instead of having to context switch and search from a browser). available from Visual Studio Code through an extension In addition to this, for enterprise customers, OverflowAI is creating the ability to plug in (internal Q&A, wiki pages, document repositories), to provide a cohesive Q&A experience for developers. Being able to utilize internal and StackOverflow data, and more important exposing this easily in a Q&A type interface, can be a big productivity boost for engineering organizations. They also intend to launch a Slack integration as a seamless interface to expose this capability. several different sources of information within a company What’s impressive about OverFlowAI’s product approach is that it takes the company’s core asset (answers to difficult questions), exposes answers in a highly usable interface wherever the users are (whether on or within developer environments), and in turn creates a loop where users can leverage generative AI to submit new questions. Slack How OverflowAI accelerates the future of StackOverflow StackOverflow is not exactly a public company - they are owned by Prosus, which is in turn part of a bigger holding company Naspers, which is publicly traded. Therefore, it’s hard to get clean revenue data but a report from Prosus published in paints a picture: May 2022 The company made ~$89M in revenue in 2022, split 50-50 between the enterprise product StackOverflow for Teams and Reach products (advertising and employer branding) From 2021 to 2022, StackOverflow for Teams revenue was +69% while Reach products revenue was -12% There could have been extraneous factors that impacted 2022 revenue like slower hiring. Nevertheless, the steep growth rate for StackOverflow for Teams shows where the business is headed I think this is generally great news for StackOverflow. Their advertising business is in decline - which is not surprising. The success of advertising products is directly tied to traffic, and the trend there is downwards. Even if the revenue drop is explained by macro factors, it’s undeniable that the rise of generative AI products like ChatGPT and GitHub Copilot is going to reduce traffic to StackOverflow. This means that their . advertising business becomes a less sizable revenue stream, which seems dire at first glance but isn’t necessarily bad The most likely long-term state of StackOverflow is that it , and the volume of questions and answers continue to grow with the company’s generative AI push to automatically draft/submit questions. I would go one step further and argue that if StackOverflow can keep the content engine running (i.e. have new questions submitted and answered), the , as repetitive/easy questions will no longer be the highest volume of content. continues to be a valuable source of answers for difficult questions quality of content on the platform will improve . For the company to thrive long-term, how they monetize should be tied to their core value. StackOverflow’s core value proposition is providing answers to users’ programming problems with the least friction possible + having a high volume of trustworthy data that future AI models can use. This aligns closely with their focus - the StackOverflow for Teams business (growing +60% year over year) and the launch of their data licensing program (where they charge AI companies for training on their data). What will change is how they make money The company has made the right product and business investments to weather what was a potential disruption, and I am optimistic these investments will set StackOverflow on a new trajectory where they continue to thrive. A world in which: , but a . If this happens (and I think it will), it would be a turnaround story for the books. they have fewer eyeballs (measured by page views) and therefore less advertising revenue thriving content platform that is monetized through an AI-first enterprise product and data licensing Also published here. 🚀 If you liked this piece, consider subscribing to Every week, I publish one in the form of a 10-minute read. Best, Viggy. . my weekly newsletter deep-dive analysis on a current tech topic/product strategy

Turning Disruption into Opportunity: StackOverflow's Transformative Pivot

So you want to learn React Native? Here's 12 exercises to that will help you do that. The post is inspired by a free self-paced workshop I've created. If you're interested in guided instructions, solutions, and extra exercises check it out! Exercise 1: Hello, World! The classic starter project. Render a square with a background color. Inside of the square render the text "Hello, world". Helpful Links React Native View Docs React Native Text Docs Exercise 2: Capturing Taps An app isn't much if a user can't interact with it. The most basic form of interaction is capturing a user's tap (or press) on a screen. Using the component, capture a tap event and "hello". Button alert Helpful Links React Native Button Docs React Native TouchableOpacity Docs Exercise 3: Custom Component Previously we've been exclusively using component from React Native. But what about our creating our own? Build your own button component that accepts an and prop. Use the TouchableOpacity and Text components to accomplish this. onPress text Exercise 4: State & Props We've taken a look at props previously - props allow us to pass data down to a component to interact with it. These props can be static values or dynamic (changing) ones. What we haven't used is state. State allows us to track changing values and update our UI automatically when they change. Using state track how many times a user presses the button and display that number. Helpful Links Using the State Hook Introducing Hooks Exercise 5: Styling If you're familiar with CSS-in-JS you should be able to fall right into styling in React Native as it follows the same principles. Create 3 squares that are vertically and horizontally centered. Each square should have a unique background color but all shared styles should only be defined once. Helpful Links StyleSheet API Documentation x Layout with Flexbo Exercise 6: Scrollable Content Rarely does all of the content you want to display completely fit into the user's screen - that's where scrolling comes in. Render 15 squares in a scrollable list. Helpful Links ScrollView Docs Explained: Each child in a list should have a unique "key" prop. Exercise 7: Building a Form Much like you would have an input in html, React Native has the TextInput component. The big difference between the two is that, with a TextInput, you get no styling with it - you're responsible for it all. Using the component capture a user's name and, upon a button press, their name back to them. Add some style to the while you're there! TextInput alert TextInput Helpful Links TextInput Docs Exercise 8: Long Lists When you have longs lists of data the ScrollView component won't cut it - it's just not built for it. Render the array of people using the component. FlatList Helpful Links FlatList SectionList Exercise 9: Working with an API Just like a web browser, React Native implements the and APIs. This means that the same logic and packages we would use to fetch data in a web browser will work in React Native. fetch XMLHttpRequest Fetch data from the url (it returns JSON) and render it in the . https://randomuser.me/api/?results=100&inc=name FlatList Helpful Links Fetch API Docs Promise Docs useEffect Hook Docs useState Hook Docs Exercise 10: Multiple Files In React Native we leverage JavaScript Modules. These allow us to export code from one file and import it into another file. Take the component and put it in its own file then import the component from that file. Button Helpful Links JavaScript modules in depth Exercise 11: Class Components Throughout this workshop we've been using functional components. In reality, this is what you'll most likely be writing. BUT they're a somewhat new concept. That means that some older tutorials/documentation will use the previous API - class components. Using the render a screen that has a button. Count each time the button has been pressed and store it in state. React.Component Helpful Links React.Component Docs JavaScript Class Docs Exercise 12: Component Lifecycle A component has a lifecycle, much like we do. They're born (mount). Things change (update). And then they die (unmount). Sometimes we need to take actions based on where the component is in the lifecycle. Upon component mount, start listening to the listener. Store in state how many times it has been called. When leaving the exercise screen and going back to the exercise list screen the listener should stop. Helpful Links Effect Hook Using the Effect Hook Good luck learning React Native! You can watch the full workshop here: Also published on React Native School.

12 Fun Challenges to Learn React Native

Learn what systems really make the money. I've had a lot of people reach out to me with a very simple question. How do you build software that makes money? The answer to that is actually very simple. You don't. . It simply doesn't do that on its own. Software doesn't make money Sales make money. Of course, that doesn't mean that you can't build that make money. systems Keep on reading ahead, if you want to understand the difference. Now, let's admit it, for a lot of software engineers, is a dirty word. sales They don't want to think about it. Instead of focusing on sales, they adopt a mindset that "if you build it, they will come." That, unfortunately, is not true. , it was the mindset that I had for my first company, PlusMore. Trust me The failure of PlusMore taught me that they won't come just because you've built it. You need to attract the right people there. PlusMore's product was a service that let people staying at hotels book different things like dinner reservations from the comfort of their rooms. Kinda like a concierge app on a tablet in each room, or on their own phones. In order to get our product in the hands of customers, my buddy Krish would call up the hotels and tell them about what we were doing. He'd go through all this manual work for each client to just get us an opportunity to go to them, show they what we could do, and maybe get a sale. It was a lot of work for what turned out to be not a whole lot of sales. I didn't think anything of it at the time though, because that's how I thought sales worked. I thought they were a very personal, one-on-one kind of process. After all, my father is a salesman, selling paint to contractors for large projects. He goes to their offices, he goes to their job sites, and he chats people up, and lands the contracts. It doesn't have to be that way, though. There are other ways to put people through your sales process. Potentially easier ways to put money in your wallet. After my company PlusMore failed, I'd been reading by Robert Kiyosaki, and something he had learned was if you want to get rich, don't work for money, work to learn skills and gain experience that can be put towards your ventures. Rich Dad, Poor Dad I decided to try to work for some experience rather than just money. I needed to build up knowledge and understand where I had gone wrong. I took a pay cut and took up a job with a pair of digital marketers selling informational products. I was blown away by the work that these two guys were doing. At PlusMore we had been 90% product and maybe 10% marketing. It was a damn good product. These guys were the exact opposite, pouring most of their energy into marketing with a seemingly relatively simple product. The difference was in the results. PlusMore had failed, and these guys were comfortably in the two comma club. When I got the opportunity, I asked one of them: how did this get started? His response? "I really like building systems." As a software engineer, I was floored. I'd been learning to build the wrong types of systems! Sure, I had learned to build the kind of systems that support business and help them thrive, systems for scaling distributed systems, systems for implementing pretty UIs, but I hadn't learned the kind of systems that make money! There are actually systems whose whole purpose is to just make money. There are systems that help you make sales. Systems that move a customer up a value ladder, moving them into gradually more expensive products and earning you more money. Systems for attracting leads. Systems for converting leads into customers. Complex systems, simple systems, effective, and ineffective systems. Instead of focusing entirely on the product, the guys I worked with were building systems that would take customers through each step of the the sales process. It was as if they were taking someone like my father and multiplying him. Instead of having to make the sales pitch with each client, they made it once, and then used marketing funnels and newsletters to bring that pitch to each customer. And they were absolutely making it rain with this system. Instead of putting your effort into individual sales pitches, you need to be putting your effort into creating systems that give that pitch to as many interested people as possible. The guys I worked for had their sales guy creating videos which were sent out to their potential customers each week. That way, he made the effort of his pitch once, but using the internet, multiplied the effect of that effort by hundreds and thousands. It's kinda like having the ability to clone someone like my dad and to send him out to multiple job sites all at once. You're going to get a lot more sales. (Please no one clone my dad.) That's the real secret of this, though. You don't build software that makes money. You build systems. Sales systems. Marketing Systems. Operation Systems. You turn a product into a full business. In fact, there are five main types of systems – five pillars of business if you will, and building systems for each is what helps you actually become a business. Each with many subsystems of their own – but that's a topic for another time. So, in conclusion: If it's money you seek, you'll need systems for making money, not just scaling services. Want to learn more about me and my story? Click here to read about me and how my agency, Unbounded, can help you turn your product into a money-making system. Image licensed by Adobe Stock Photos, by ALDECAstudio .

How Do You Build Software That Makes Money?

Qwest predicted Netflix, Hulu, and HBO GO in 1999. By Micah Rosenbloom , Partner urging people to temper the hype surrounding the Amazon/Whole Foods acquisition, essentially arguing that Amazon will be a disruptive player in the food industry, but don’t expect them to scarf down market share immediately. It’s an important point. Food represents 13% of US GDP and many of the deal’s boosters are underplaying the sheer volume of innovation in the food category at the moment. Erin Griffith wrote a great post QSRs like SweetGreen and Chipotle have redefined fast food while supermarkets like Whole Foods and Wegmans are increasingly doubling as cafeterias/restaurants with huge takeaway selections. Recent entrants in food delivery — Seamless Grubhub, Postmates, Uber Eats, etc. dramatically changed the ease and frequency of food delivery. We invested in Plated, Gobble, and SkipTheDishes (Acquired by JustEat), and with the exception of Uber, I’ve never seen companies grow revenues so quickly. In retrospect it makes sense, most of us eat every day, multiple times, but still, it’s astonishing that a group of companies that are less than ten years old now account for $10B in consumer spending. As Erin notes, many armchair analysts are beginning to write off the packaged meal category, as well as grocery stores more generally, based on the current Blue Apron share price, but let’s remember that this stuff takes awhile. The force of change in food has only just started. Change in Tech Takes Time. Lots of it. As a reminder, this commercial which premiered in 1999 at the height of the first dot-com boom. At the time, the promise of “every movie, ever made, in any language, day or night” was as fantastical as a flying car to most people. This was a time when Blockbuster, a store-sized version of Redbox popular in the last century, still issued fines to people who returned videotapes that weren’t rewound. Techies and VCs of course saw the potential and started building it, but nearly 20 years later, we’re still a ways away from enjoying that enticing value proposition. Sure, I can watch on any device that has an internet connection, but if I want to stream I need to sign into my Prime account. adds a third app, and yet another subscription fee to the mix. There is a cottage industry of websites , or , a given streaming service at any given time. A great many of our cultural treasures — and to name but two — remain in inaccessible IP vaults. Orange is the New Black The Americans Game of Thrones explaining which shows are coming to leaving ER The Fresh Prince of Bel-Air From a technologist’s point of view, video on demand is one of the simpler products to be built. It’s just packets over the internet, after all. But even in today’s world of widespread broadband, DRM, and omnipresent smart TVs and set top boxes, we’re still managing multiple services and remotes. How to think about self-driving cars, AI, and ecommerce Now imagine how long it may take one of today’s hot technologies, like self-driving cars, to take hold. If you read some pundits, they believe truck drivers will soon be as obsolete as fur trappers. Let’s examine this hypothesis critically. If it’s taken us 20 years to stream reruns, do you really think we’ll have technical solutions to route through Boston or SF’s backward traffic patterns, unexpected detours caused by exploding water mains, falling cranes, drunk drivers, and distracted pedestrians in short order? How quickly will the business case for fleet ownership be hammered out? Will 50 states be able to coalesce on laws governing the design of autonomous vehicles and licensure for the operators or coders in a timely fashion? Do we need new infrastructure, like sensors embedded in streets, to make fully autonomous driving work? Will progress be slowed the first time an AV clips a kindly grandmother crossing the road? Will the drivers of these trucks, who often pull double duty stocking shelves at their destinations, be replaced with robots? These are not problems that are going to be solved in the next year. It’ll probably take a decade or more. That’s why smart companies, who are able to play the long game, are investing in companies like Cruise today. Self-driving will only be a reality if companies are willing to do the hard work to make them a reality. Three’s Company For those panicking that AI will put most workers out of jobs, remember that Amazon’s vaunted machine learning recommendation engine is regularly a punchline in tweets because it advertises products that we’ve already purchased. Amazon has been grinding for over 20 years and ecommerce still only accounts for 10% of retail sales. Ebooks should have replaced dead trees by now, but paper has proven to be as resilient as Katniss Everdeen and their electronic counterparts have plateaued at 50% market share. Customers are weird and unpredictable. Regulation often routes technology in suboptimal directions. Established players can be surprisingly durable and loathe to adopt new technology. Food is not videos and is not AI and each will follow its own path, probably slower than you think. It’s fun to make bold proclamations about the future, but the perspective of history tells us we should be more circumspect. As Bill Gates said “We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten.”

Tech Takes a While

This is the first article, in a new series where I’ll be sharing my journey of building an app for learning a language with Jetpack Compose. I’ll show you my daily progress, how the app evolved over time, why I made certain decisions, and so on. I’ll also share the new things I learn along the way that might be useful to you. When I started programming, one of my first projects was a website that I used to review words I was learning in other languages. Fast-forward a few years, and I need something similar. This time, however, I’ll be building it using Jetpack Compose and aiming to publish it to the Play Store. If you’re a junior engineer, this series will show you my process - how I went from nothing to having an app that’s ready to be published to the Play Store. If you’re a senior engineer, this series will hopefully teach you some new things about Jetpack Compose and might even cause you to rethink the way you solve some problems. Background I’m calling this app Lingua, it’s “language” in Latin. It’s supposed to be a combination of Duolingo and Anki plus some other things I think will be useful. One of my goals for 2023 is to learn Italian, and I also need to spend some of my time actively creating something instead of just passively learning new things. I’ve read a lot about Compose, now it’s time for me to put that knowledge into practice (later I recognized that I actually knew very little about Compose). If I arrive at a version that looks good and I think it makes sense, I’ll publish it to the Play Store. This is the first draft I created in Figma; it is far from the final version but guides me in the right direction. I’m building this app iteratively, so as you will see the first versions of the app will look pretty ugly but that’s okay, the initial goal is to get something that works. Later, I can come back to it and improve the design. I’ll split my updates into days, so you can have a reasonable idea of how long it took to build this, I’m spending about 1hr on this daily. Tools and Libraries To keep track of my tasks, I’m using Notion. For the design, I’m using Figma. I’m using Kotlin and Jetpack Compose. I’ll mention the other libraries as they come into play. Day 1 Below, you can see my first draft of the home screen. I’ll have some kind of progress section that’ll show how much the user learned in a given day, in a given week, and in a given month. I also want to add a strike feature to motivate people to keep learning. Data is pretty valuable to me, so I’ll also add detailed charts about things that might be of value to the user. Instead of just having the usual “Practice” most apps have I also want to build a custom practice that lets you choose what you want to practice. And finally, I want to show the list of courses/decks the user is subscribed to. I’ll call these cards “decks” from now on but the name may change in the future. My screens are divided into route and screen. In the example above, I created and . The route just is what gets added to the navigation graph, the screen, and the actual visual content. HomeRoute HomeScreen Here we have the main for the application. NavHost NavHost(navController = navController, startDestination = Routes.Home.route) {
    composable(Routes.Home) { HomeRoute(navController) }
} And here the that defines the home screen. HomeRoute @Composable
fun HomeRoute(
    navController: NavController,
) {
    HomeScreen(...)
} I’m just using for now but I’ll probably abstract that later. For routes I just defined a sealed class with my routes, for now, this suits my needs but I’ll likely change this too in the future. NavController sealed class Routes(val route: String) {
    object Home : Routes("/")
} Keep in mind that this is my first “big” application with Compose so I’ll certainly make mistakes and learn along the way. I then created the other components but for now, I’m just assigning hard-coded values to them. One thing worth mentioning is how I achieved cropping for flag images. I’m using Coil. I was not able to add cropping directly to the image, so I wrapped it with a and defined a rounded corner for only the bottom end corner because the top start corner is already cropped by the card. Surface I also set the aspect ratio to 16/9 but I’m not 100% happy with the result. Surface(
    shape = RoundedCornerShape(bottomEnd = 8.dp),
) {
    AsyncImage(
        model = model.imageUrl,
        contentDescription = "Icon",
        // TODO: add placeholder
        contentScale = ContentScale.FillBounds,
        modifier = Modifier
            .width(36.dp)
            .aspectRatio(16 / 9f)
    )
} By the end of day 1, I had created the screen you can see below, it’s not functional yet but it’s a good step in the direction I want to go to. Day 2 On day 2, I just built the library screen, it’s pretty simple right now. It lists the available decks. In the future, I’ll add more sorting options to make it easier for the user to find what they want. It also has a button for the user to create a new deck. This is what I had achieved by the end of day 2, nothing special. Just a and a . I actually also created the for this screen so I just need to hook this up to a data source later and the correct data will be listed. LazyColumn FloatingActionButton ViewModel Day 3 On day 3, I started working on the deck creation screen. Basically, I need at least one deck with cards to be able to develop the rest of the application so that’s why I’m going this route first. As you can see below this screen is not as simple as the other ones. First, you can name your Deck, I’ll build this initially. After that you can change it to public/private, I’ll leave this for later. Everything will be public for now. On the top right, you can see a flag, that’s a cool functionality but I won’t be implementing it right now too. In the middle, we have the most important part of this screen, the cards that compose a Deck. For now, there will only be one type of card and it’s just . Input <-> Output For example, the English word Bee is Ape in Italian so it’d be something like Bee <-> Ape. This allows me to do a few things: Show the word “Bee” and the user has to type “Ape”. Show the word “Ape” and the user has to type “Bee”. If there are more words I can show something like “Ape” and the user has to choose a word from a list like (“Apple”, “Bee, “Pie”). This is basically a simple version of Duolingo. Later I’ll be adding support to sound and image but for now, that’s just additional complexity. Finally, we have a button to add new cards and a save button to save everything. I ran into a few problems like how do I pass the deck id to this screen? I’m using the same screen for adding and editing decks so the id had to be an optional parameter. I solved this by creating a new object in and adding the and methods. Routes createRoute parse object EditDeck : Routes("/deck/{id}/edit") {
    fun createRoute(deckId: String?) = "/deck/$deckId/edit"
    fun parse(bundle: Bundle?): String? = bundle?.getString("id")?.takeIf { it != "null" }
} And in my routes, I added the new route. composable(Routes.EditDeck) {
    val deckId = Routes.EditDeck.parse(it.arguments)
    EditDeckRoute(navController = navController, deckId = deckId)
} Then when my screen is created I use a to load the deck in the . LaunchedEffect ViewModel @Composable
fun EditDeckRoute(
    navController: NavController,
    deckId: String?,
    viewModel: EditDeckViewModel = hiltViewModel()
) {
    LaunchedEffect(deckId ?: "none") {
        viewModel.loadDeck(deckId)
    }

    EditDeckScreen(...)
} The code inside will get executed again if the key changes, in this case, if the deck id changes and that’s exactly what I want. LaunchedEffect By the end of day 3, I had built only a small portion of this screen. Day 4 By day 4 I stopped working on the screen to create decks and moved on to the screen to create cards. Decks and cards are the foundation of this app, everything else in the app will revolve around them. For now, I’ll only develop the “Text” type or as I called it earlier Input <-> Output. It’ll contain one field for the input and one field for the output (I need to come up with better names for these things lol). Later I also want to add an “Info” type that works a little bit differently but that’s for later. This screen is pretty similar to the previous one, I have my and EditDeckCardRoute . We also have 2 text inputs and a FAB to save the card. EditDeckCardScreen One difference here is that to create a card I need the deck id and to modify the card I need the card id. To solve this I created 2 routes: and . EditCard AddCard composable(Routes.AddCard) {
    val deckId = Routes.AddCard.parse(it.arguments)
    EditCardRoute(navController, deckId = deckId)
}

composable(Routes.EditCard) {
    val cardId = Routes.EditCard.parse(it.arguments)
    EditCardRoute(navController, cardId = cardId)
} The caller decides which route to call based on what it wants to do but on the implementation side, I chose to use the same route. So when the is created, it either calls a function to create a new card or one to load an existing card. ViewModel fun load(deckId: String?, cardId: String?) {
    when {
        deckId == null && cardId == null -> {
            Logger.e("deckId and cardId are null")
            // navigate up
        }
        deckId != null -> createNewCard(deckId)
        cardId != null -> loadCard(cardId)
    }
} Here’s what this screen looks like by end of day 4. Pretty simple screen but now I can at least create simple cards and add them to a given deck. In this update, we went over the first 4 days of this app. I shared what I’ll be building and why. I also described what I built each day and a little bit about how it was built. I’m enjoying this new series a lot and I’m hoping you’re too. If you liked this update please share it with your friends, and coworkers, …. Stay tuned for the next updates. Cover photo by on . Kelly Sikkema Unsplash Also published . here

How to Build a Language Learning App with Compose – Part 1

The coronavirus has caused havoc across the globe. Millions of people are infected and thousands are dead. The pandemic is still on the rise as all the countries in the world have imposed a lock down. Everybody has  the rights to know how this virus has affected our lives as we continue to wage a war against it. So, I write this article about my project with a voice that addresses the current situation of the world. The Voice Assistant Project In this project, you’ll learn how to build a COVID19 voice assistant using Python. The project mainly covers the fundamentals of , and in python. web scraping voice assistance multi threading Web Scraping The project makes use of a tool called . ParseHub is free web scraper that is very powerful and is easy to use. This tool allows you to scrape the web merely just by clicking on the elements you would like to get out from that website. So, the first step is to download . ParseHub ParseHub We will be scraping from the very famous website for statistics which is that is regularly kept updated about the coronavirus information. Open up ParseHub -> Start New Project -> Paste the link to worldometers. Worldometers Click on the elements you would like to scrape from the page and assign unique tags to them. Use the relative tool to link let’s say, ‘Coronavirus Cases’ on the page with its number. The ParseHub is clearly very efficient and uses an AI to understand which elements you would want. To try this, link ‘USA’ with the ‘total cases’ and it does rest of the countries in the list for you. Amazing! What this is doing behind the scenes is constructing a JSON format of what you are linking above. After you are done with the linking, hit ‘Get Data’, which may take a while. Then make sure, on the next screen you note down the API key, Project token and Run token. Voice Assistant Copy paste the code given in and I shall walk you through what is really going on. Part1.py These points explain the Part1.py file: The requirements for this part of the project are the following python packages: , , , , . So, pip install all these packages. requests json pyttsx3 SpeechRecognition re Use your API key, project token and run token in their respective variables. The class has the following methods: Data : fetch the data from parsehub using your most recent run for the project on their servers and return it. get_data() : get the total cases of COVID-19 around the world. get_total_cases() : get the total number of COVID-19 deaths in the world. get_country_deaths() : get the number COVID-19 cases and deaths in any particular country. get_country_data() : get the list of all the countries. get_list_of_countries() The method initializes the pyttsx3 engine to say a text passed into it as a parameter. speak() The method listens to the speech input by the user through the microphone using the method by Google and returns the input text. get_audio() recognize_google() Finally the main() function is responsible for recognizing the input speech using regular expression patterns and classify them as patterns for cases and deaths in world or cases and deaths in any country. The matched pattern calls that particular function and speaks the number out loud. Multi threading As of now, your assistant speaks only the data you fetched initially. It is important to keep the data updated, as this global pandemic is still on, it would be soothing to hear it when your assistant says that the number is declining. Delete Part1.py and copy the into your application. Part2.py The Part2.py has just an extra function and a handler for it in addition to Part1.py. This function is the function. The function, doesn’t fetch from the last run on parsehub, but instead initializes a new run on the parsehub servers for your project. This will take a while, so please give it some time when you are updating. update() The most interesting part is to reduce this time. So, we make use of multi threading using the library in python. By doing this, we ensure that the voice assistant is running on one thread and the updating of data takes place parallely on another. multithreading Result The most interesting part of this project is its result. Run the Part2.py file and when the console says “Listening…”, say “number of total cases” and the assistant would reply with the number of COVID19 cases across the world. Now say, “total deaths in India” and the assistant would reply with this number. You can say, “update” and the assistant will update the values for you which may take some time. Please wait patiently. You will have to say “stop” to exit from the running application. 😲Yay!!! You did it! You just built a voice assistant that updates you about the coronavirus by only using your voice. You can further develop this project into amazing things. To get complete insight checkout my . GitHub If you got any errors, feel free to reach out to me on my . LinkedIn Hopefully, you enjoyed as well as understood this project. Thank You. Previously published at https://medium.com/@KGPrajwal/covid19-voice-assistant-63c37b1f02f9

Building Covid-19 Voice Assistant

Every product or company goes through a similar set of steps on the path to success. In previous posts, we’ve looked at different aspects of going to market — from launching to building and measuring a funnel to pricing — with some ideas about how to measure things and know when you’re off track. In this post, we’ll take a step back and walk through all the stages of GTM that you’ll typically experience. Programming notes: this post is n in a series of indeterminate length on GTM topics mainly for startup people, mainly leadership, mainly coming from non-GTM backgrounds. There’s a list at the end. Going To Market What did you make? Who is it for? How does it make their life better? How do you find them? How do they find you? What will get them to try or buy? How much do they have to spend on you? How will you sell it? How many deals do you need to close or users do you need to acquire in a month, a quarter, a year? What is your (real and desired) sales motion? < — All your answers to these questions are only put to the test and can ultimately (for a while) only be arrived at by going to market. Going to market is the sense mechanism of your business. It’s how you know whether you even have a business and provides the only relevant* input to developing that business. Everyone thinks it’s great, but no one wants to pay for it? That’s not a business. A niche of users in the 100s loves it and will pay a small amount for it, but no one else cares? That’s not a business. A niche of users in the 100s loves it and will pay a very big number for it. That might be a business. Millions use the free version and 10s of 1000s pay for premium features such that revenues are growing faster than costs? That might be a business. * Raising money from VCs is not a substitute for going to market. Someone’s, even an expert’s, opinion of whether you might have a business is not a substitute for the market reacting to your proposed business by giving you money (or growth) or rejecting you. Stages of Going To Market What follows is from a B2B angle. B2C is analogous, but the mechanics can be wildly different depending on whether you’re growth and indirect revenue (like from advertising) predicated or direct revenue (user = customer) predicated. There are many models, but this one is mine. :) Discover Problem There’s something out there that’s broken, someone who’s needs are poorly or not served, or just something that could be so much better if someone just built ___. This is where it all starts. And it’s usually . You have the issue yourself or someone you know does. personal Validate Market Ask some people if it’s true. Talk to experts and specialists and the people who would be the users, influencers, and buyers. Maybe do some research, see if anyone else is already solving the problem and how. Find out how big the market might be, how many users on earth or just in your country might exist. Or not. Just solve a problem that you have on the presumption that you represent a set of users or customers big enough to sustain a business through revenue or data or users to sell to advertisers, eventually. Product Build the thing.* *Once you start, you never get to stop. Product Market Fit [Price Discovery] See if the people you built it for like it! If it does the thing you expected. If they want it. If they’ll — signing up, changing something they do, paying some money — to get it and keep it. overcome some friction At base, this is Product Market Fit*. A core function of PMF in products that are supposed to make money is discovering what price your market will bear, or is competitive, for the value you provide. *Depending on who you talk to, PMF requires GTM Fit [see below]. Network Sales Sell it to people you know, people your staff knows, people your investors know. 1–2 degrees away with warm intros. Some companies only ever do this and build solid businesses this way. Growth in the hockey stick sense happen this way.* does not *A lot of sales and biz dev hiring in ye olden days — and more often today than anyone would admit — is simply buying access to a person’s network and the positive disposition of that network towards them. This is not materially different from an advertiser buying eyeballs. You’re paying for at-bats. Novel Customer Acquisition Novel, because it’s not based on your existing rolodex. Now you’re doing lead gen work that exposes new people to your product and costs real money, either in hard currency or in paying for staff/contractors. This includes: sponsoring events, speaking at events, running ads on podcasts, SEM, buying contact lists and emailing them, hiring SDRs to contact people, running meetups, investing in content marketing for inbound search driven traffic, etc. Natural Market Limit Discovery There is a point early in your company at which you will fall off a cliff of diminishing returns on making sales or acquiring users. This will manifest as lead gen effectiveness failing or increased effort to convince prospects who fit a given profile to give your product a shot or poor conversions rates at the middle and bottom of the funnel regardless of what’s happening at the top or any general stalling of growth that seems impervious to the marketing and sales you’ve tried to date. You are discovering the bounds of your Natural Market*. Natural Markets consist of: Have the problem and know they have it, or so close that if you point at it they know they have it Want to solve it Will pay to solve it Have the authority and responsibility needed to spend their own or their team’s or their company’s money to solve it or the influence over such a party to cause it to happen Who you can reach If this happen, you are on a rocket ship — right product right market right team. In which case, your biggest (only?) risk is success. does not *Natural Market is an idea that I got from Tae Hea Nam of Storm Ventures in a board meeting. It is the single most useful thing I’ve learned from a VC. CAC Ramp Usually you don’t know you’ve hit your Natural Market bounds. It will just seem like you’re experiencing the normal diminishing returns of a particular marketing tactic(s) or marketing strategy or playbook or GTM strategy. Most teams will start with the assumption of a fading marketing tactic and try others. Whether you layer on more tactics, like adding display advertising on top of search ads, or out more resources on an existing tactic, like going from one big event sponsorship every six months to one big one every quarter — this inevitably results in more $$$ per customer acquisition.* *This is separate from the fact that CAC is rising for everyone in general due to the sheer number of products and services vying for people’s attention that all ultimately arrive at a small set of channels to do it. This is a secular trend. Tactical Experimentation A kind of flowering of marketing and sales tactics, including potentially larger GTM bets like sales partnerships, takes place in response to the challenge of rising CACs. In the B2C world, if you’re experiencing outstanding user growth, there’s an infinite appetite on the part of investors to fund that growth. Even with rising CAC. Strategic Instability Usually in parallel, you will start questioning your marketing and sales strategies. Maybe the entire GTM model, including core things like who your customer is and is this even the right product. Example: If you thought you had an inbound, inside sales business, you might start playing around with field marketing and outside sales. GTM Fit Exploration Do you have a GTM model that can be executed and scaled sufficiently to get you from where you are right now to your next stage of growth? Then you have GTM Fit. The hallmark of GTM Fit is having predictability in terms of resources in, like marketing dollars or AE headcount — and results out, like users or logos or dollars. *GTM Fit is an idea from Tae He Nam and Bob Tinker. It’s best summed up in this video . I agree with every last thing Tinker says in this talk and have found it to be true both at startups and when shipping new products at places like IBM. Playbook Scaling When you’ve locked into a GTM that works, you’ll pour resources into it to achieve maximal user or rev growth. Diminishing Returns [Again] Just like running into the limits of your Natural Market, you’ll eventually find the a peak in yield from a playbook. This will tend to look more like a slow decline than falling off a cliff.* The market will saturate, at least for you, as you hit max possible penetration and fight harder for smaller gains. Sales and marketing tactics will become less effective. Etc. *If you’re pouring millions+ of dollars and hundreds+ of people on the fire of a working playbook.. it’s more likely that you’re going to fall off a cliff. Operational Discipline If you have great GTM Fit, you’re scaling incredibly fast and just trying to stay one step ahead of your own success. In this situation, everything you’re doing in marketing and sales is massively inefficient and your funnel is leaking everywhere. It just doesn’t matter because growth is so rapid. When you start to hit post GTM Fit speed bumps, all of a sudden people will start caring about operations, analytics, process, and discipline. This is severely normal. Bloomberg didn’t have a formal marketing function until it was 1B+ in rev. New Relic didn’t have an enterprise sales team until many years after it was a tech household name. Most CFOs, CROs, and CMOs of startups in spitting distance of an IPO could not tell you what their sales efficiency is or what the conversion rates are for their funnel stages with even the slightest bit of confidence or accuracy (ask them!)*. *If you don’t have great GTM Fit, if you’re way early with a product the market is not ready for yet, or if it’s just a grind in the way that enterprise businesses are — then you’ll often end up with a great grasp of what to measure and optimize for and the resultant operational discipline much earlier, because you had to in order to survive. Market Expansion You have gotten all you can get out of a particular GTM, so it’s time to expand who you sell to, what you sell, or both.* New channels, such as distribution deals with a partner or being bundled with another product or being sold by consulting companies into accounts/geographies you don’t have access to or leverage in. New segments, such as going after enterprise where you had only been selling to SMB. New verticals, such as selling into Financial Services where you and only been selling to CPG. New products you build or buy that can be sold into your existing markets and/or give you access to new markets. Companies that you buy, get bought by, or merge with to give you access to new markets, or new products that you can sell to your existing markets or new markets, or that you effectively buy revenue streams or users or eyeballs that you can add to your own. *All attempts at Market Expansion basically start a new cycle of GTM with its own Product cycle. I’ve left that out because depicting interacting loops in a serial text format is damn near impossible. Exit or GTFO Eventually, most people want to Exit.* Your early investors want liquidity. Your employees want liquidity. You want liquidity. You might just be tired and ready to dissociate from the thing you’ve spent years and years building. Exiting is a GTM stage because, regardless of how you exit, you will have to hit specific GTM measures around costs, revenues, and growth to do it. *If you don’t want an Exit, you might have a lifestyle business. Or something bootstrapped you want to do forever. Or you’re going to buy out your investors. Or you’re going to take on some private equity (not necessarily PE) investors who get dividends. Or etc. More power to you. Final Note The road is winding and often non-linear. Posts in this series (and templates) Marketing 101 for Engineers: A Functional Introduction Marketing 102 for Engineers: Roughing Out a Funnel Marketing 201 for Engineers: Messaging & Positioning Marketing 202 for Engineers: Launching Marketing 203 for Engineers: Sales Enablement Marketing 204 for Engineers: Generating Demand Marketing 301 for Engineers: Strategy & Planning Marketing 302 for Engineers: Hiring Marketers Marketing 303 for Engineers: Pricing Frame Marketing 401 for Engineers: Stages of GTM Marketing 402 for Engineers: Diagnostics & Troubleshooting Sales 101 for Engineers: A Functional Introduction PR 101 for Engineers Analyst Relations 101 for Engineers Basic Messaging Template [Google Doc] Basic Funnel Metrics Template [Google Sheet] Basic Launch Timeline Template [Google Doc] Basic Battlecard Template [Google Doc] Detailed Battlecard Template [Google Doc] Basic Marketing Calendar Template [Google Sheet] Basic Marketing Ladders Template [Google Sheet] Reading List and Resources Bob Tinker’s GTM Fit and Talk Slides Tomasz Tunguz Tren Griffin Startup Marketing Playbook First Round Capital: PR & Marketing Magazine First Round Capital: Sales Magazine Max Olson: Advantage Flywheels Go-To-Market Fit: The Missing Link for Enterprise Growth HubSpot: Top 10 Sales Methodologies Image credits [in order] Photo by on 2Photo Pots Unsplash Expensify Value Proposition Diagram from Tom Tunguz Bob Tinker on GTM Fit and Talk Slides Using Price & Demand Curves from Tom Tunguz

JUST BUILT IT

justbuilt.it

-- employees

Since n.d.

COMPANY RANKING

JUST BUILT IT

EVERGREEN INDEX #5848

Representing this company? Click here to claim and customize this page!

Just Built It WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!

JUST BUILT IT

justbuilt.it

-- employees

Since n.d.

COMPANY RANKING

JUST BUILT IT

EVERGREEN INDEX #5848

RELATED COMPANIESRANK

STORIES ABOUT JUST BUILT IT

Representing this company? Click here to claim and customize this page!

Just Built It WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!