Robin Leman

@akeshwan1 @charlieholtz In your setup script you can add something like `ln -s $CONDUCTOR_ROOT_PATH/.env .env` this will symlink the .env from your repo dir.

@charlieholtz hey charlie, maybe i’m using it wrong but what’s an easier way to manage env variables? If i create new branches, I’m having to manually type env files back in on every new branch. Am I doing something wrong?

@benspringwater I feel like it’s a protection mechanism. If people convince themselves LLMs aren’t “smart”, they don’t need to fear them

This is embarrassingly wrong and at this point anyone who believes it is just not paying close enough attention.

@Tyler_A_Harper So a Ph.D. in Comparative Literature says that researchers with a Ph.D. in Machine Learning don’t understand how AI works? Bold

Taps the sign

@matt_elevenlabs 11B

We just raised $500M at an 11B valuation 🎉 To celebrate, we’re giving away 1,000 free credits so you can test our platform. For the next 6 hours, comment “11B” below and we’ll DM you the credits (must follow) 👇

Merci à @HugoDecrypte pour l'invitation à faire un tour d'horizon des risques de l'IA ainsi que des solutions nécessaires pour se diriger vers une meilleure trajectoire. Depuis notre discussion en octobre, mon travail scientifique à @LawZero me rend de plus en plus confiant qu'il est possible de créer des IAs sécuritaires et fiables.

@KeithLee2049234 @miniapeur You’re allowed to work part time I believe in the UK, but it’s hard to find part time positions

@miniapeur Inability to side gig is one of the worst parts of being an international PhD student. There were so many short term, exciting design/consulting work that I had to pass on to friends or watch slip by because I couldn’t legally work outside of my RAship.

Best ways to make money during your PhD?

@alz_zyd_ Why did the projects fail generally? Was it because you didn’t spend time on them (but they could have been successful), because the idea didn’t work, or maybe someone was faster than you? Genially curious, as a grad student it feels like the cost of a failed project is huge

A student asked me: do research projects die? What fraction of projects die? So I looked at my dead projects folder to estimate this fraction. It turns out I have 21 papers posted/published/etc., and my dead projects folder has 126 subfolders

@mathladyhazel If you have a 20cm squared board, and it takes you 10min to cut the board in half (10min to cut 20cm). This results in two 10cmx20cm boards. It makes sense that you could cut 10cm in 5min for one of the boards. Just cut in the orthogonal direction to the original cut.

I can't believe my eyes. 😵‍💫

@damekdavis @littmath @Archivara Just pre-train on @littmath tweets directly

@littmath @Archivara RL for math but the reward function is just @littmath responding to your posts

Archivara (powered by 5.2 pro and Opus 4.5) breaks the 2019 benchmark for 5×5 circular (cyclic) matrix multiplication. The best verified construction required rank-8 (8 multiplications); we now provide a practical, fully verified rank-7 algorithm. This realizes a long-known theoretical possibility as an explicit, working method.

@NathanBurtnett @yacineMTB You also can’t underestimate having a set curriculum with modules that were created and refined over time. Sure I can learn ML on my own, but being taught by folks at Gatsby and Deepmind in modules that are sort of guaranteed to be relevant, it helps

@yacineMTB Tell me if I’m wrong but it seems the three things colleges provide is: > seeing different perspectives > doing at least something challenging > networking

You do not need university to learn anything at all The mind of a slave:

@chompie1337 I often get the opposite, where I ask a question in French and it suddenly switches to English

WTF??? ChatGPT switched languages mid-sentence. Yeah, that’s my sign to log off for the day.

@Flako772_fl @Sneketoshi @Math_files Wdym? It is

@robinleman @Sneketoshi @Math_files Yes they can but this is not one of them

2 is the only even prime. But the total number of primes is infinite. Therefore the probability that a given prime number is even is 1 over infinity, or zero. Hence it’s impossible for a prime number to be even — and hence 2 does not exist.

@cloneofsimo Yeah, except the “make a game engine” part isn’t the easiest one..

Trazillion dollar idea. Make a game engine, a decent one, that supports basic stuff, but that could be heavily extended and modified. But make it unconventionally console / test driven, make it easily verifiable. Then, make a lot of toy game with it. Get lot of instructions, Train (fine-tune) an LLM on it. Sell the LLM & game engine as a paired product. Train on the user-generated games. SFT. RLVR. I would do this if I didnt have a job

@OwainEvans_UK Video game, but Detroit: Become Human is amazing.

is there a novel that has an AI as a proper character (with some interiority) and is science fiction? if not, i expect one soon.

@Sneketoshi @Math_files Events with zero probability can happen

@Math_files There’s a flaw in your logic. I’m too lazy to do the work but I can say with absolute certainty that 2 does exist.

@ChaddWeston @charliegreenman @edenchan @xai @grok MIT, and McGill is Canada’s Harvard (unbiased) youtu.be/cc5vN2XReWs?si…

@charliegreenman @edenchan @xai @grok Waterloo is Canada’s Stanford

if you graduated from waterloo and want to work at @xai I want to talk to you

@conne_psd @Waymo x := -x is probably enough

@Waymo did entire machine learning code have to be rewritten or just flipped horizontally?

Hello London! 👋 Our vehicles are now driving in London as we prepare for commercial service in 2026.

We’re announcing a research collaboration with @CFS_energy, one of the world’s leading nuclear fusion companies. Together, we’re helping speed up the development of clean, safe, limitless fusion power with AI. ⚛️

@anw_g01 @Julien5050 @gdb @grok matrix multiplications are essentially a massive number of linear equations

@Julien5050 @gdb @grok large language models are essentially massive neural networks > neural networks are essentially massive layers of matrix multiplications

wild what a few matmuls can do

@SFSEgo GameDev here, assets are HEAVY. Yes most are compressed, but there’s a limit to how much you can compress while preserving quality. Players want 4K textures, but not everyone has the hardware to run it, so you need the same texture duplicated in different resolutions

I know some technical game dev people follow me, so WHAT IS MAKING THESE GAMES SO BIG???