Dario says stupid stuff. But, it doesn’t really matter. You expect regulators to be apolitical and set clear policy. Statements like this make it clear the decision was political and had little to do with security or policy.

2019: "What if we put the AI in a sealed box with no internet and it STILL manipulates someone into letting it out?!" 2026: We connected it to a biolab, gave it a credit card, and 50 million people volunteered to give it root access to their laptops for $20/mo

Excited to release new repo: nanochat! (it's among the most unhinged I've written). Unlike my earlier similar repo nanoGPT which only covered pretraining, nanochat is a minimal, from scratch, full-stack training/inference pipeline of a simple ChatGPT clone in a single, dependency-minimal codebase. You boot up a cloud GPU box, run a single script and in as little as 4 hours later you can talk to your own LLM in a ChatGPT-like web UI. It weighs ~8,000 lines of imo quite clean code to: - Train the tokenizer using a new Rust implementation - Pretrain a Transformer LLM on FineWeb, evaluate CORE score across a number of metrics - Midtrain on user-assistant conversations from SmolTalk, multiple choice questions, tool use. - SFT, evaluate the chat model on world knowledge multiple choice (ARC-E/C, MMLU), math (GSM8K), code (HumanEval) - RL the model optionally on GSM8K with "GRPO" - Efficient inference the model in an Engine with KV cache, simple prefill/decode, tool use (Python interpreter in a lightweight sandbox), talk to it over CLI or ChatGPT-like WebUI. - Write a single markdown report card, summarizing and gamifying the whole thing. Even for as low as ~$100 in cost (~4 hours on an 8XH100 node), you can train a little ChatGPT clone that you can kind of talk to, and which can write stories/poems, answer simple questions. About ~12 hours surpasses GPT-2 CORE metric. As you further scale up towards ~$1000 (~41.6 hours of training), it quickly becomes a lot more coherent and can solve simple math/code problems and take multiple choice tests. E.g. a depth 30 model trained for 24 hours (this is about equal to FLOPs of GPT-3 Small 125M and 1/1000th of GPT-3) gets into 40s on MMLU and 70s on ARC-Easy, 20s on GSM8K, etc. My goal is to get the full "strong baseline" stack into one cohesive, minimal, readable, hackable, maximally forkable repo. nanochat will be the capstone project of LLM101n (which is still being developed). I think it also has potential to grow into a research harness, or a benchmark, similar to nanoGPT before it. It is by no means finished, tuned or optimized (actually I think there's likely quite a bit of low-hanging fruit), but I think it's at a place where the overall skeleton is ok enough that it can go up on GitHub where all the parts of it can be improved. Link to repo and a detailed walkthrough of the nanochat speedrun is in the reply.

I got to Bangkok yesterday and already won a mini in-house tournament in @LootSurvivor game @PlayerHouse_gg @ohayo_dojo

It’s a Wrap!! Congratulations to @adrdhots and @Elahzul_ for your top 3 placements. New Death Pit is live now! #gamingcommunity #whatthefunge #web3game $PRIME

Congratulations to our Top Jumper, @adrdhots, for securing first place by reaching level 1112 during First Jump. - Run time: 15h 22m - Total kills: 9,094 - Shots fired: 20,722 You have received a 1/1 "Quantum Jumper" NFT, certifying you as a Wormhole Hero.

I have won the leaderboard in @Wanderers , reaching lvl 1111 and winning a custom 1/1 avatar NFT - linked is the video of some of the gameplay. The game is fun and runs really smooth (opposite of most other crypto games), the gameplay is interesting and has potential, for now it's pretty simple, but only a small part of the game was available. Additional speed run attempt for the first 10 lvls at the end of the video at youtu.be/zySCeynAvrk?t=3205 (afaik it was also the highest score for this beta phase) youtube.com/watch?v=zySCeynA…

About to play the finals against the rank 1 player @ManosTCG , wish me luck, he destroyed me yesterday so let's see if I can comeback and get my first lan win as a player 🫸🫷

Coming in hot as the season 6 champion, we're happy to welcome @NMZ_0702 as the newest addition to our esport roster. 🔥

Started the season 2 days ago, let's climb

Small thread about the recent crypto games I tested/competed in etc recently Focusing mainly on gameplay competitive earnings.

The main reason why people are complaining about rage now is because they remove the cancel queue function. Widening the mmr range allows strong aggro comp to thrive and "counter comp" to be less viable.

Although some couldn't believe it was even possible, my art is getting better and better, here's the more aerodynamic version of @argRALO's Strawberry aqua

Congratulations to @Elahzul_ on winning - Paragons Showdown #10 Lock and load, gamers! Initiating our run in @ParallelTCG

Gonna finish at 2000, not enough time to get top10, climbed about 600 points in ~10 hours

Time to start epic era, 20 hours to go, let's see how much I can climb

Just showing off my best pulls @CyberKongz #OOH