Lee Nelson

Here's a step-by-step tutorial on how to build a @figma plugin that uses @OpenAI to generate color schemes. designerscancode.com/tutorials/crea… Even if you're not a Figma user, It's a great intro on how to integrate LLMs with web apps and familiarize yourself with the chat completions API. If you are a Figma user, It provides great insight into what will soon be possible using natural language to update our canvases and files. I hope you enjoy it. -- #figma #ai #buildinpublic #codetutorial #devtutorial #aitutorial #figmatutorial #figmaplugins @FigmaPlugins @zoink @rogie

@felixleezd @ycombinator Join

Designers, YC is calling for you. @ycombinator just put out a "request for startups" for more designer founders. starting a group chat for designers who want to build their idea/startup... we’ll share: - product ideas - best frameworks and tools - design engineering talks - design founder chats reply if you wanna join

@awilkinson If the new site is anything like the current one, I wouldn’t even use a CMS. When building the site, include a md file with instructions for the agent so less technical folks can easily edit & upload content in the future.

Working on a new site for Tiny. We typically use Webflow, but I'm wondering if I should/could just have Claude code the front-end. My worry is it becomes difficult to manage. What is the best way to vibe code front-end but have an easy to manage backend/CMS that is secure?

@Tesla deciding to do an update 10 minutes before my family and I need to board a ferry at 7pm is the worst UX I’ve experienced in my entire life. Microsoft updates don’t even come close. @elonmusk

@felixleezd @ycombinator +1

@awilkinson @Replit @lovable @cursor_ai I think we’re still in the honeymoon phase. I too frequently find myself having to correct Cursor. Most value I see is for up-leveling eng. Use cases that involve comparison. Architecture, etc. Vibe coders I know are still running into walls and don’t know how they got there

Is your software business a sandcastle on the beach? In 2025, the tide is coming in, and it wont' be pretty. Software that once required millions in R&D can now be vibe-coded by teenagers over a weekend using @Replit, @lovable, and @cursor_ai. It's a commodity. Kind of like salt: Once rare, valuable, and controlled by the privileged few, now it's ubiquitous, dirt-cheap, and available to all. If you want to succeed in software, your only hope is to have one (or ideally multiple) of the following moats: Network Effect: The more users that use your software, the better the experience becomes for everyone, eventually establishing it as the standard (Stripe, Shopify, Figma, Notion). Proprietary Dataset: Your unique data and insights can't be easily recreated by others (Bloomberg, Zillow, Palantir). Complex Technology: Something not easily AI-coded (this won't hold up long-term, but works for a few years). Hardware Moat: Physical devices that lock in your software (Dexcom medical devices, Apple/Google phone ecosystems, Square POS hardware). High Switching Cost: It's too annoying or difficult to replace incumbent software because organizations are deeply integrated with it or people are trained on it (Microsoft Office/Teams). Brand and Reputation: Established trust that matters most in security-sensitive areas (Coinbase, 1password, Crowdstrike). What are the implications of this? RIP Pricing Power: Margins directly correlate to competition levels. Charge $10/mo for a calorie tracking app? If a college kid launches a competitor for $1, your price will have to drop and your acquisition cost will increase. This is Bad for Legacy Software Investors: Those who paid 8x ARR for SaaS companies without the qualities listed above are in trouble. If your product is merely an interface on top of a database, beware. This is Amazing for Consumers: Software is about to get 100x better. An incredible long-tail of specialized apps will emerge, unleashing unprecedented human creativity. Value Will Accrue To The Search Box: Just as Google disintermediated TripAdvisor and other travel sites, ChatGPT will disintermediate many software companies. It will eventually connect to all of your APIs and data, then use them to complete tasks you'd otherwise need software for. For most queries requiring software, it will simply code custom software on the fly. (I might be wrong about on-the-fly coding, but I'm extremely confident about API disintermediation). How we are approaching this at Tiny If you'd sat me down and told me that Google Search dominance was under threat 3 years ago, I'd tell you to stop huffing gas. AI is profoundly disruptive to all digital businesses (and has major implications for many real-world businesses as well). Anyone who tells you otherwise is a fool. As a result, we have passed on dozens of deals that we otherwise would have jumped at a few years ago because they lack one of these now critical moats. We've learned this lesson the hard way. A few years ago, we bought a software company that was hugely dominant in a niche. It was a gem of a business: juicy profit margins, fast organic growth, and a passionate user base. But all that slowed the moment competitors flooded in. It's capitalism 101: if you're the only place in town selling apples, you can charge $5/apple and probably get away with it. But the moment other farms start charging $2/apple, your prices have to drop almost immediately. That is, unless you have some sort of special, insanely tasty or healthier apple. But for most people, apples are apples, and price matters. The same thing is true in software: Unless people can't live without your specific software and it's hard to replicate the functionality it provides, your pricing ultimately has to be competitive with the market. In that particular software business, we found that users would simply jump to cheaper options once they got good enough (it happened quickly). Before we knew it, the company we acquired (an early pioneer and innovator) came under pricing pressure and growth slowed once competitors flooded in. A lot of people are about to learn this same difficult lesson. Godspeed. Stay frosty out there, folks 🫡 What did I miss? Curious if you think I'm off base. What companies are the walking dead? Who's untouchable?

@thesamparr @willroman Parsable alumni ✊ Way to go, Will. These boots look incredible.

This is my buddy Will (@willroman). How can he turn his cowboy boot company into a $100m/year company? You see, a few years ago Will quit his job and moved to Mexico to apprentice under one of the best cowboy boot makers in the world. Spent years there, mastering boot construction. Then he started a cowboy boot company in Austin, Texas. It’s called @ChisosBoots. They have one location, are bootstrapped, and great margins. What I like about Will is that he’s obsessed with the craft of making super, super high quality boots and unwilling to compromise on quality. This is not how most D2C companies start. Most of em are marketers who are good at ad buying. I don’t think Will spends much, if anything, on ads. If you google reviews, people on Reddit say Chisos are the most comfortable boot and that for $600 they’re a steal given the craftsmanship and durability. I’m not a financial investor in Will but I am emotionally invested because I think his boots are damn good. I also think this business can be huge (tacovas, a startup in the space, does hundreds of millions a year in sales), and I think it’s fun to show off small businesses before they get huge. So, my question to yall: What would you do if you owned Chisos and how would you make it huge in the next 10 years?

@negativespaceca @rogie @figma Do either of you know if it's possible for devs to access enterprise features if they don't belong to an enterprise org? Some sort of enterprise sandbox perhaps?

Yessssss! Node just shipped the ability to run TypeScript files directly!

Looking for a talented freelance designer to collaborate with the team over Stripe.com on events-related web projects for the next 5 months! A strong eye & experience in visual design, interaction, and systems are essential. Looking for someone to start in August through the end of the year! US based time-zone working hours preferred. Interested? Contact me directly!

I'm hosting a pitch event in Victoria with a few local angel investors next week. Who wants to pitch their company?

@daniel__designs Nice work. I think @MengTo is doing something similar.

Some very early experiments

I'm almost done with a new @figma plugin that generates vector fields and patterns. It's super fun to play with. It supports custom svgs, different vector field patterns, and automatically adjusts to frames. Just need to knock out a few bugs before I make it free and open source

@thesamparr @dennishegstad @BroScienceLife Nvm he’s based in LA. I thought with a face and name like that, he had to be a Yankee.

@thesamparr @dennishegstad sk8 with @BroScienceLife and do a pod at the same time.

Trying to convince @dennishegstad to skateboard in NYC with me. What are some cool spots and any other old dorks skate too?

If you're a designer who is creative and imaginative, someone who is pushing the boundaries of UI/UX, now is your time to shine. Every software product in the next 3 years is about to feel & behave the same. We're about to enter the: Bootstrap of AI era. #config2024

@marcofyi @diagram @figma Great work! 👏

a year ago @diagram joined @figma to further explore AI in design today, at #Config2024 we helped launch Figma AI alongside so many amazing people we can’t wait to see what you create with it ✨

New Tutorial: @figma plugin that generates typeface pairings. These GPT functions are a little unpredictable at the moment but nevertheless, if you want to learn more about using @OpenAI GPT functions, sign up for designerscancode.com.

Someone needs to buy hats.com and make it 1000x better.

Unpopular opinion #Figma should not release a native table builder. They should stick to primitives - tables are a component. This is a perfect use case for a widget/plugin. They should embrace that community instead.