it took 1 application for us to get into yc

AI as we know it might not exist if @nvidia had not gone all-in more than a decade ago. CUDA was foundational to the development of AI and remains important for training. Inference, however, is different. CUDA has no role in inference. AI is now useful. For coding. For doing your taxes. For finance, legal and HR departments. People are coding their own apps, eliminating SaaS applications. And none of them know anything about CUDA. And of course it’s this way. It always is. When a new technology emerges the practitioners use low level software. It looks like math. Its written by the select few. Graduate students. Supercompute engineers. But to get popular, the technology needs to leave this environment. There simply aren’t enough technologists to make a market. There are no low level programming languages that are popular. They are just too hard. In order to get popular, computer languages move from low level to high level. Compare how many people write Assembly versus Python. It’s: 1 vs. 10,000. When a technology is used by those who know nothing about its creation, and instead care only about its usefulness, then it has broken through. It takes 12 keystrokes to link a world class AI model to your application, and start getting blazing fast tokens. You don’t need to know anything about AI. There is no CUDA. Just point to the OpenAI API. Or the Cerebras API. Or to Fireworks. Or to DataBricks. Here is what you type to move from Nvidia GPUs to Cerebras. API key= os.environ.get(“CEREBRAS_API_KEY”) That’s it.

39 startups are building alongside Vercel for six weeks. Backed by $8M in credits from AWS, Anthropic, OpenAI, ElevenLabs, and more, they have the resources to scale their infrastructure. Meet the 2026 Vercel AI Accelerator cohort. vercel.com/blog/2026-vercel-…

software generation is no longer the bottleneck. it's operations trillion dollar opportunity for whoever solves it

I realized something else AI has changed about coding: you don't get stuck anymore. Programming used to be punctuated by episodes of extreme frustration, when a tricky bug ground things to a halt. That doesn't happen anymore.

we are doubling deployed helix db instances daily

Hi friends! Big news today: Helicone is joining Mintlify 🚀 3 years ago, we started Helicone out of YC W23 with a simple bet: every hype-cycle has a massive observability market, and we felt like AI was going to be even more prevalent. We knew someone needed to build the platform to ingest and make sense of it all. So we did. Years of grinding and a handful of pivots later, we built something real that customers use every single day. I spent most of that time worried about tomorrow and never really stopped to look back. But looking back now: - 14T tokens processed - 30k sign ups - 36M users tracked - Trending on Github last year (5.2k stars) - $1M ARR - Product of the week on Product Hunt That still blows my mind. So why Mintlify? In a post-AGI world, a knowledge infrastructure layer is the thing that makes the most sense for companies to invest in. Waymo didn't reinvent the roads, but still needed to learn about them. Mintlify is the information layer for agents building the systems of the future. The overlap with what we were building at Helicone was obvious. Same world, different angle. More importantly, I actually like the team and the product. That sounds simple but after talking to a dozen companies, it's rarer than you'd think. @handotdev and @hahnbeelee are both incredible and formidable founders that Cole and I really admire. Things I learned building Helicone that I think are underrated - A 5-person team can do an absurd amount if everyone is locked in. Headcount is not a moat. - If you have data at scale and need analytics, just pick ClickHouse. Don't overthink it. - The grind is real, and it's fun. But REALLY try not to be stressed when you aren't grinding. That part took me way too long to learn. - It's never too late to pivot and try new things. - Charge more for your product. What's next I'm going to be leading an engineering team at Mintlify and I'm genuinely excited about what we're building. Keep an eye on my X, because I can't wait to be tweeting about it more. Thanks to @coleywoleyyy for being an incredible co-founder. Thanks to our team, our investors, our customers, and everyone who bet on us early. And thanks to my fiancee for putting up with the chaos of startup life. That part is deeply underappreciated. And lastly, thank you @mintlify for taking a bet on the Helicone team. We are so excited for the next phase. More to come. DMs are open.

OpenAI🤝Cerebras openai.com/index/cerebras-pa…

you will regret not using a monorepo.

Open source isn't dying because of AI. It's dying because we all fucking suck. Users. Maintainers. The whole ecosystem. All of us. We fucking suck as users. We use OSS to do our jobs. To build products. To make money. Our employers depend on it. Our startups run on it. But the second a maintainer mentions compensation? We lose our fucking minds. "It should be free." "That's not the open source way." Meanwhile we're extracting value from someone's unpaid labor every single day. We file issues like we paid for a support contract. We demand timelines. We get snippy when our bug isn't the priority. We're not customers. We're guests. Start acting like it. We fucking suck as maintainers. We wanted the GitHub stars. The conference talks. The Twitter followers. We wanted to be the person who "owns" that thing everyone depends on. But now people actually depend on it and we're annoyed. "Talk is cheap, show me the code." So someone does. They spend hours on a PR. We ignore it for months. Close it with no explanation. Or we screenshot their issue and mock them publicly for not reading our minds about what we actually wanted. We invited contributions then punished people for contributing. We don't want to maintain projects. We want to be admired for maintaining them. Not the same thing. We suck at this together. Both sides want the benefits without the responsibilities. As users we want free, maintained, high quality software but won't contribute a damn thing. Not money. Not code. Not even basic respect. As maintainers we want the status of running critical infrastructure but won't communicate, won't collaborate, won't treat people like humans. We all know this shit doesn't work. And yet here we are. Okay Josh, what do you suggest then? If we use OSS and profit from it, we contribute something. Anything. Money, docs, triage, kindness. We stop expecting infinite free labor. If we maintain OSS and we're burnt out, we say so. Archive it. Hand it off. Ghosting is worse than walking away. And if we can't treat each other with basic respect? We don't get to participate. Full stop. Open source runs on people. On us. None of us owe each other a god damn thing. But we could choose to stop being assholes and do better anyway.

lol: Determinate Nix's is cutting Nixpkgs tarball cache import time from 40s to 4s in the next release.

My data shows that basically *all* new user growth with Nix is with users choosing flakes. That flakes alone are responsible for making Nix much more approachable and understandable. [...]

Programmers of which language exhibit the most signs of being in a cult?

Had @theo as my top Channel.

Axiom sets out to build an AI mathematician. We are the underdog. 4 months old, 2 years late to the game, under 10 FTEs (recently grew to 17), and had 1:5 in funding and in valuation to our competitor. Today, AxiomProver solved Erdos Problems #124 and #481 in Lean, a 100% verifiable language. Onwards!

I'm reinstalling Arch, so I clicked on a mirror to download it, and... well I'll be.