😸 Trust fall

gm legends, happy Thursday.

Today: a no-code editor that catches what vibe-coded apps break, docs that update themselves so you stop lying in your changelogs, and an AI CEO named Astra who's ready to run your whole operation. Also: whether any of this should be funded at all.

WeWeb 3.0 adds AI generation to a visual no-code editor — meaning you get the speed of vibe-coding without the part where you can't find the bug three days later. Raphaël Goldsztejn and Marc Fabre have been building WeWeb since 2021, before "vibe-coding" was a term, and the product reflects that: the visual layer is what you maintain, and AI generation is just faster input.

🔥 Our Take: Most AI app builders assume the code is the product. WeWeb assumes the opposite: the no-code structure is the interface you live with, and AI is just how you fill it in faster. That's a real distinction. You can vibe-code anything in Cursor — the question is what happens when you need to change it six months from now and the person who wrote the prompt is gone. WeWeb's answer is structured by design. Whether that trade-off is worth it depends on how serious your app actually needs to be.

Mintlify Workflows turns documentation into a self-updating knowledge base — detecting when your actual product diverges from what the docs say and correcting itself. Han Wang left school to build a cloud monitoring system for dairy cows, eventually sold a community platform he co-founded with Hahnbee Lee, and landed on documentation as the problem he kept running into in every project he shipped.

🔥 Our Take: Stale docs are a maintenance problem pretending to be a writing problem. Mintlify's bet is that if you make them self-correcting, the writing problem disappears too. The real risk is AI-generated corrections being confidently wrong in ways manual updates aren't — especially at the edge cases where docs matter most. Mintlify already has serious traction (74 reviews at 5.0, $67M raised, a16z-backed), so the question isn't whether they can build this. It's whether accuracy holds under pressure.

Inference should feel simple. Once AI hits production traffic, you're really managing routing, retries, observability, GPU bills running while nobody's home.

DigitalOcean Serverless Inference handles all of that in one API — OpenAI- and Anthropic-compatible, 55+ models including Claude, DeepSeek, Qwen, and Llama. Scale to zero when you're not using it. Pay per token when you are. VPC and zero data retention on by default.

230 tokens/sec on DeepSeek V3.2 — 3.9× faster than AWS Bedrock (Artificial Analysis). Runs on the same bill as your databases and Kubernetes. One less thing to reconcile.

Tycoon AI runs your company with an AI CEO named Astra plus 10+ specialized agents for social, coding, growth, and ops — you set the KPIs, Astra plans and delegates. Xiaoyin Qu dropped out of Stanford's MBA to build Run the World during COVID, ran a small company with real staff for three years, sold it in 2023, and is now building the infrastructure that would replace that team entirely.

🔥 Our Take: Every "AI company in a box" product lives or dies on the same question: are the agents doing real work, or generating output a human still has to babysit? Tycoon's pitch is that Astra "asks for approval when needed" — that clause is doing all the work. The case for taking this seriously: Xiaoyin has actually run a small company, felt the bottlenecks, and sold it. She's not guessing at what breaks. Whether Astra has learned what she has is another matter.

Aleksandar Blazhev (@byalexai) opened with a stat: 81% of $297B in global VC this quarter went to AI infrastructure and frontier labs. His argument is that software is now a commodity and distribution is the only real moat, which makes bootstrapping more viable than it's ever been.

The thread split between founders who've taken money and the ones who haven't. The VC skeptics weren't anti-funding so much as anti-accountability-to-someone-else. Blake at Vassant put it plainly: taking investor money shifts your loyalty from your customers to your investors. The VC defenders said the usual — speed, connections, hiring — but nobody landed a strong rebuttal to the accountability point.