Bertil Chapuis

829 posts

Bertil Chapuis

Bertil Chapuis

@bchapuis

Loves hacking and collaborating at the intersection of several disciplines. Building in public. 🚧 🏗️ ⚒️ https://t.co/mA9j4suhNm

🇨🇭Switzerland Katılım Ocak 2009
91 Takip Edilen385 Takipçiler
Sabitlenmiş Tweet
Bertil Chapuis
Bertil Chapuis@bchapuis·
Started exploring extensively how AI-assisted coding changes web development in 2025 and it escalated into an open source visual automation platform built on @cloudflare called Dafthunk. 2026 plan: build more, share more 🏗️
Bertil Chapuis tweet mediaBertil Chapuis tweet media
English
2
0
4
401
Bertil Chapuis
Bertil Chapuis@bchapuis·
@grok thoughts? Is one of these a long-awaited load-bearing emoji?
English
1
0
0
32
Bertil Chapuis
Bertil Chapuis@bchapuis·
Can't contain my excitement for these 8 new emoji. Worth every one of the 7150000000 bytes
Bertil Chapuis tweet media
English
1
0
1
144
Bertil Chapuis retweetledi
Leonard
Leonard@LeonardCser·
@mitchellh been building cursortab.nvim for this: github.com/cursortab/curs… sweep-next-edit-1.5B with speculative decoding would work the best on M4 max with llama.cpp for better but slower completions zeta-2 is the best local model
English
0
1
8
817
Bertil Chapuis
Bertil Chapuis@bchapuis·
@francoisfleuret Teaching students, some of them are surprisingly good with agents. To me, drive and systems thinking seem to matter more than raw experience. TBH I'm still not sure which traits matter most
English
0
0
1
433
François Fleuret
François Fleuret@francoisfleuret·
I have the impression that using coding agents for complex tasks requires a sense of feasibility and consistency of the results that can only comes from years of actual programming. I am deluding myself? A sub-par programmer can really shine with agents?
English
112
10
458
39.5K
Bertil Chapuis
Bertil Chapuis@bchapuis·
@lucatac0 Quick question: is it possible to get the API schemas of the models programatically? I'm introspecting schemas on replicate and this would be very useful on AI Gateway
English
0
0
1
10
Luis Catacora
Luis Catacora@lucatac0·
companies use an average of 3.5 AI models across multiple providers that's multiple SDKs, multiple bills, multiple points of failure AI Gateway gives you one API for all of them: blog.cloudflare.com/ai-platform/
English
1
2
21
3.4K
Bertil Chapuis
Bertil Chapuis@bchapuis·
I'm currently experimenting with cloudflare durable objects and starting to think they're a new kind of primitive for scaling multitenant applications. Every tenant gets compute and an isolated 10GB of storage, scales to zero when idle, billed per-usage. Does anyone know if an open source alternative already exists?
English
0
0
1
68
Bertil Chapuis
Bertil Chapuis@bchapuis·
Side project update: my globe built with Claude in C/WebGPU/WASM can now serve and render custom tiles that include DEM-derived meshes and vector layers from Overture Maps... YOLO mode can brings you quite far
Bertil Chapuis tweet media
English
1
0
4
128
Bertil Chapuis
Bertil Chapuis@bchapuis·
@lemire Someone should write "How Societies Actually Work" for curious and competent people (who start by acknowledging a knowledge gap). Because those traits become handicaps when everyone around you is trained to project confidence on half-understood topics
English
0
0
1
90
Daniel Lemire
Daniel Lemire@lemire·
We are now reaching stage 4 of “Damn it, this AI can do this guy’s work better than he ever could.” On paper, a human with access to AI should outperform AI alone. Not so. Let me explain why. I work for a state university. Anyone with a university degree gets “professional” status, regardless of actual competence. A technician with no degree could easily outperform many degree holders. It is what it is. These degree holders tend to be particularly uncurious. I call this the Bach syndrome, after James Marcus Bach. He never went beyond high school. When Microsoft claimed Internet Explorer could not be removed without breaking Windows, they called Bach. He showed it could be done in minutes. Why was Bach the expert? In one of his books, he explains that when he started at Apple, he studied constantly. He soon realized that many colleagues with prestigious degrees did not. They believed university had taught them everything they needed. They now held positions of authority. Done. Yesterday, some of these people (university degree + position of authority) confronted me on a topic I know well and they did not. I stayed polite—I could not say “you are full of shit”—and calmly explained why they were wrong. I won the point, but what struck me was this: the person could have asked ChatGPT, Grok, Claude, or even Copilot and gotten the correct information instantly. They didn’t, because they don’t care to know. When they encounter something unfamiliar, they don’t research it. They simply proceed with assumptions. It is all about power relationships. I am on record as predicting that AI won't cause mass unemployment, but I sometimes wish it would. Meanwhile, please, for the love of God, if don't know a topic at least as well as ChatGPT, don't speak as an expert.
English
24
40
361
41K
Bertil Chapuis
Bertil Chapuis@bchapuis·
More and more unsolicited emails from OpenClaw bots in recruiting mode, scraping GitHub for developers interested in agentic ai. Is this really the current state of automation?
English
1
0
2
94
Bertil Chapuis retweetledi
Christos Tzamos
Christos Tzamos@ChristosTzamos·
1/4 LLMs solve research grade math problems but struggle with basic calculations. We bridge this gap by turning them to computers. We built a computer INSIDE a transformer that can run programs for millions of steps in seconds solving even the hardest Sudokus with 100% accuracy
English
249
809
6.1K
1.8M
Morgan
Morgan@morganlinton·
After hearing Timothy Chalamet insult ballet and opera - two things I’ve enjoyed a lot in my life, I’m not going to watch his ping pong movie. I don’t care why he thinks ping pong is better, or more relevant than ballet or opera, he’s wrong in my book. My predictions for tonight - Sinners is going to break some serious records, and I hope Michael B. Jordan wins best actor, he deserves it. Also loved Bugonia so hoping it wins some awards, super creative film.
English
5
1
15
2.8K
Bertil Chapuis
Bertil Chapuis@bchapuis·
🎉 Just got accepted into @Cloudflare's Worker Loader API beta. Codemode lets the LLM skip the one-tool-at-a-time loop and write code instead. The integration in Dafthunk was just straightforward and my agentic node can now call tools with code. Can't wait to get accepted in the email service beta 😘
Bertil Chapuis tweet media
English
4
2
41
5.3K
dina kozlov 🐀
dina kozlov 🐀@dinasaur_404·
Code More via Dynamic Worker Loader for the win!! @thomasgauvin you’re up next with email 😛
Bertil Chapuis@bchapuis

🎉 Just got accepted into @Cloudflare's Worker Loader API beta. Codemode lets the LLM skip the one-tool-at-a-time loop and write code instead. The integration in Dafthunk was just straightforward and my agentic node can now call tools with code. Can't wait to get accepted in the email service beta 😘

English
2
2
17
2.7K
Bertil Chapuis
Bertil Chapuis@bchapuis·
Same experience. Something I want to try earlier in the curriculum: Teach the basics, give students a problem and an AI-quality solution, then tell them vaguely what's potentially achievable with the problem, domain, and hardware knowledge acquired during the course. Their job is to get there, using AI however they want.
English
1
0
0
48
Daniel Lemire
Daniel Lemire@lemire·
How do you teach programming in 2026? I have been teaching programming professionally for two decades. It doesn’t work anymore. I now consistently catch students who produced good software during a 15-week course be unable to write a simple loop at the end of the course. « Just forbid AI. » You can’t put someone in a cage for 100 hours and force them to program without AI. We need to change the whole approach. Maybe go AI first. 🤖 This means MUCH harder homeworks.
English
289
25
478
79.5K