Filip Ryzner 📌

ML @Pinterest | CS @MIT 22'

[Filip Ryzner 📌 @Ryzny]

@JohnLeFevre Ignoring all else, this calculation completely ignores dilution.

[Filip Ryzner 📌 @Ryzny]

@LukasKovanda Momentalne jiz ctvrta

[Filip Ryzner 📌 @Ryzny]

@maxstephhh It’s not like it was by choice

[Ethan Jiang @ethanjyx]

We just launched Tight Studio 3.0.

It’s a powerful Screen Studio alternative with a generous free tier.

The core features you expect from a polished screen recorder are free:
• Smooth zoom-ins
• Cursor animations
• Trimming and cutting
• Beautiful recording effects

The biggest difference: you don’t have to keep restarting your entire demo when you mess up.
Record in clips, edit everything in one app, and turn it into a polished video.

We won Product Hunt Product of the Day last October, and thousands of users have tried Tight Studio since.

Perfect for product demos, tutorials, launch videos, and explainers.

[Matthias Schmidt @eurofounder]

If the EU built Claude

[Filip Ryzner 📌 @Ryzny]

@PIE_RRE_O @Kj_lite @Arsenal They are rotation players it’s fine. You cant have mucu better players on the bench

[Aidan McLaughlin @aidan_mclau]

one of my all-time favorite plots

[Carole Cadwalladr @carolecadwalla]

This speech *is* one for the history books. But that’s less a compliment, than a coda.

Carney has given us the words to mark the end of the ‘rules-based order’…by acknowledging it never really existed.

It was a collective illusion. That now is over.

[legal YIMBY alien @YIMBYPoland]

I don't recall Czechoslovakia owning any colonies in 1930s, when they were among the wealthiest countries in the world, with advanced manufacturing often beating that of Germany, UK and the US. Somehow by 1989, that same nation was decades behind the West on every metric.

[MIT CSAIL @MIT_CSAIL]

Andrej Karpathy, Eureka Labs founder & former Director of AI at Tesla, breaks down how LLMs like ChatGPT "download the Internet."

Watch the full beginner-friendly breakdown here: https://t.co/doph3alGnE

[Filip Ryzner 📌 @Ryzny]

@faelpontopt Ragebait

[Filip Ryzner 📌 @Ryzny]

@wesbos For most airlines the answer is “in hell”

[Filip Ryzner 📌 @Ryzny]

@ns123abc Ofc it is haha

[Filip Ryzner 📌 @Ryzny]

@alexfinn It’s built on chromium. What is the point you are making?

[Andrej Karpathy @karpathy]

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.

[Casey Neistat @Casey]

so cooked.

[Tom Krcha @tomkrcha]

Design Mode for Cursor

Introducing a new way to design right where you code.

[Filip Ryzner 📌 @Ryzny]

@Andr3jH 🤣🤣🤣🤣

[Filip Ryzner 📌 @Ryzny]

@mehdirhasan I mean, he’s not wrong lol

[Filip Ryzner 📌 @Ryzny]

@ChristianHeiens I mean that account claims to be based in Germany.