💖 ➞ λ: CS PhD, into compiler design, programming languages, music, simulation, ML/AI, robotics. Follow me code code reviews, stock picks and dating advice.

Joined November 2011
744 Photos and videos
More fuel for open models and local models to catch up 👌
Andrej Karpathy (not a US citizen), the top AI scientist at Anthropic, is now barred from accessing Anthropic's top AI model. This is beyond stupid. _
18
1,592
Fable just released. Meanwhile I've been trying to use Claude Code's remote control feature for the last 3 days, and it's broken. CC gets stuck as soon as I try to approve something from my phone.
3
10
1,497
I put together a minimalistic pure-Rust, CPU-only implementation of the recent LFM2.5-8B-A1B language model from @liquidai that you can directly link into your Rust projects😎 github.com/maximecb/bebelm
6
11
186
13,712
This is one of those things that's obvious in hindsight, but there are implications if you are ever asking an LLM to make "random choices" or pick among items in a list: sanand0.github.io/llmrandom/

4
1
9
2,744
Trying out Qwen 3.5 0.8B to see what it's capable of. Has anyone found viable uses for tiny LLMs with ~1B parameters or less?
15
21
4,107
The thing is: if you maximize trying to make everyone happy, you end up optimizing for the lowest common denominator. That's not good taste. Good taste is opinionated and it upsets many people when confronted with it. Makes it very difficult to build good taste as a product.
Software engineers, what's your plan when AI develops better taste & architectural/systems knowledge than you in next 3-4 years?
4
4
42
4,820
I tried to get Claude Code to recreate the Quake start map in my game engine but it really struggled. I even gave it access to a screenshot feature and the original Quake map source. Neither was helpful. I think CC can't process visual information very well.
11
25
2,953
I had Opus 4.7 build me a little castle in my custom map format. It came out pretty nice 👌 No manual edits: github.com/maximecb/tinkerfi…
3
1
29
1,966
Hard to understand why, in 2026, audio quality on phone calls is so bad. Super heavy compression, ultra low bitrates, often hard to make out what people are saying... In a world where everything works over IP, why is the default 8 kilobits per second or some shit?
115
12
605
74,163
One shot attempt at 3D pelican on a bicycle (Claude Sonnet 4.6, 17m13s)
6
1
48
5,157
I had Claude vibe code some simple 3D worlds in my SDF game engine, in a custom textual map format. Not bad! 😆
5
2
60
6,526
A little update on Tinkerfield, my toy 3D game engine / construction game based on signed distance fields :)
3
2
46
4,588
I'm finding that asking LLMs to debug 3D graphics problems can be an experience in frustration. They can't see the program running so you have to describe the issue to them. They speculate as to the cause of the issue and are often confidently wrong.
38
3
113
13,601
There's many reasons to intentionally minimize the number of external packages your software depends on. pointersgonewild.com/2026-05…
3
2
27
1,980
For anyone still doubting that humanoid robots are happening and will provide value.
That nod of understanding Figure demonstrates Helix-02 model for coordination. - Two humanoids tidy the room, fully autonomously. Acting as independent agents. - The robots coordinate only by watching each other, not through any shared planner or messaging.
2
17
3,086
Which coding agent do you use for your personal projects, and why?
35
1
22
10,969
Maxime Chevalier retweeted
Jensen is one the smartest and most far seeing folks the world. "If an AI scientist warns people that AI is going to permeate across radiology and radiologists are going to get wiped out, it might seem helpful but it's hurtful. If we convince everybody not to be radiologists and we now need radiologists, that actually is hurtful to society. "It is hurtful to convince all the young college graduates not to study software engineering because we are going to need more software engineers than ever. That's hurtful." "Scaring people with nonsensical things, which are not going to happen, that this is an existential threat, there's a 20% chance that is is existential, that's ridiculous. "That it's going to wipe out 50% of college level jobs. "That is it going to completely destroy democracy. "These kinds of comments are not helpful. They are made by...CEOS. And you become a CEO, maybe you adopt a God complex and somehow you know everything." Brutal. And right.
245
812
5,281
854,186
Maxime Chevalier retweeted
Passive dynamic stability. Intelligence in mechanism: no sensors, no feedback control.
33
279
2,995
213,137