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.

JUST IN: $ZEC reclaims $8B market cap, up 21.3% today, as privacy tokens continue to trend. View chart: coingecko.com/en/coins/zcash

Ripple announced a $500 million funding round at a $40 billion valuation, with participation from Wall Street firms like Citadel Securities, Fortress Investment Group, and Brevan Howard.

$BTC (yellow) -5%, $ liq (white) -8% since US debt ceiling raised in July. TGA build up sucked $ out of the system. When US gov shutdown ends, TGA will fall +ve for $ liq, and $BTC will rise ... and $ZEC will go up MOAR!

Bitcoin Drops Below $100K Amid Ongoing U.S. Government Shutdown Bitcoin briefly fell below the $100K mark before rebounding, as broader markets declined amid the continued U.S. government shutdown.

🎙️ Session reveal #DuneCon25 Crypto-native products aren’t for everyone; they’re for power users. @richardchen39 from @VarrockVC joins us to talk about building products that go deep, not broad.

🔓 Top 7 Token Unlocks of the Upcoming Week The following tokens with the largest unlock amount will be unlocked next week: $KITE - $145.91M $ENA - $60.73M $TEA - $15.11M $JTO - $9.69M $COCA - $8.27M $AVNT - $7.05M $BERA - $5.80M

JUST IN: DeFi protocol @Balancer appears to have been exploited for ~$70M. TX Hash: etherscan.io/tx/0xd155207261…

Balancer, a long-standing DeFi exchange, suffers a major exploit resulting in $129 million in losses across multiple blockchains, including Ethereum, Arbitrum, and Polygon, within just over two hours of the attack starting on November 3, 2025. This incident marks one of the larger DeFi exploits in recent months and is Balancer's third security incident.