Tweet Disematkan
How Kaggle load-balances a gRPC application across multiple GKE clusters. buff.ly/2XHiMBG
English
Philippe Modard
3.7K posts
Philippe Modard
@Philmod
Account not used anymore. Find me on https://t.co/wIOPoOgCPf
Bergabung Ocak 2011
2.3K Mengikuti5.7K Pengikut
@doodlestein @Beetcoin I haven't been able to find the reply you're referring to. That said, I’m just looking at the data—there’s no need for personal insults when we’re just discussing technical performance.
English
English
"LLMs optimize for plausibility over correctness. In this case, plausible is about 20,000 times slower than correct." x.com/KatanaLarp/sta…
Hōrōshi バガボンド@KatanaLarp
English
Is AI singularity here already? x.com/karpathy/statu…
Andrej Karpathy@karpathy
I packaged up the "autoresearch" project into a new self-contained minimal repo if people would like to play over the weekend. It's basically nanochat LLM training core stripped down to a single-GPU, one file version of ~630 lines of code, then: - the human iterates on the prompt (.md) - the AI agent iterates on the training code (.py) The goal is to engineer your agents to make the fastest research progress indefinitely and without any of your own involvement. In the image, every dot is a complete LLM training run that lasts exactly 5 minutes. The agent works in an autonomous loop on a git feature branch and accumulates git commits to the training script as it finds better settings (of lower validation loss by the end) of the neural network architecture, the optimizer, all the hyperparameters, etc. You can imagine comparing the research progress of different prompts, different agents, etc. github.com/karpathy/autor… Part code, part sci-fi, and a pinch of psychosis :)
English
AI vs human code gen report: AI code creates 1.7x more issues coderabbit.ai/blog/state-of-…
Català
The venue (@lotto_arena) decided to change my @ElectricCallboy concert ticket from the pit to the seats, unilaterally. Really not cool! What's the point of booking my ticket year in advance if I can't enjoy the show the way I'd prefer?
English
Surfer-H Meets Holo1: Cost-Efficient Web Agent Powered by Open Weights arxiv.org/abs/2506.02865
English
I'm leaving Twitter, find me on bluesky: bsky.app/profile/philmo…
English
@eschnou Agreed! (Although “retarded” is not a good word to use)
English
@todoist Feature idea: ability to automatically group items in a specific list. For example for the groceries, I would love to have them organized, so it's easier when I go through the store.
English
TIL: Python's tarfile is slower than the tar CLI because it uses a compression level of 9 vs 6 for the CLI.
English
An Open-Ended Embodied Agent with Large Language Models buff.ly/3BWIDsk
English
Interesting write-up about a path to AGI and Superintelligence, and the risks associated buff.ly/4ea1UIE
English
Philippe Modard me-retweet
Here's what I see as a likely AGI trajectory over the next decade.
I claim that later parts of the path present the biggest alignment risks/challenges. The alignment world has been focusing a lot on the lower left corner lately, which I'm worried is somewhat of a Maginot line.
English
EV range metric should be given for a highway usage. Having 500+km in a citadine environment is not useful, since recharging is more convenient in urban environments.
English