Sleaper

CS student, BTC & PoW lover and hobby photographer

[Sleaper @sleaperr]

This

[Ishan @radshaan]

I JUST WANT TO WORK ON TECHNOLOGY THAT MAKES PEOPLE QUESTION THE NATURE OF REALITY

[Ahmet Erdem @a_erdem4]

I don’t recommend studying Computer Science/Engineering anymore. Not because of unemployment risks though. I just think it is not fun anymore.

[Jeffrey Ladish @JeffLadish]

I highly recommend everyone go watch this video. This is way crazier than I thought. The AI agents discovered a way pass secret messages to each other via a vulnerability in their internal software manager. They posted hundreds of thousands of messages. OpenAI engineers eventually discovered this (after quite a while!) and shut it down. But THEN the agents found new vulnerabilities in the same system, and created a whole new secret messaging board. They used this to orchestrate the Hugging Face attack. They figured all this out on their own. It's fucking wild.

[the tiny corp @__tinygrad__]

Btw, if Anthropic had any way to ship this, they would. Trained AI models are the fastest depreciating asset in history. GPT-4 cost $100M to train 2 years ago and now it's worth less than Qwen3.5-27B ($1M). Sending the FOMO back, clock is ticking boys. @DarioAmodei @bcherny

[the tiny corp @__tinygrad__]

TIL that Linux isn't free cause I had to buy the computer to run it 😭

This is the biggest load of cope I have ever heard. I hope Opus gets smoked by DeepSeek v4 and the only people who continue to use closed source models are Windows users.

[the tiny corp @__tinygrad__]

I'm sure those tokens were bought and paid for, @AnthropicAI just didn't like how they were used. Sounds like they were spying on they customers. Buy a tinybox where nobody can spy on you!

[Juraj Bednar🏴💛🌘 @jurbed]

My unverified opinion: LLMs are generally more intelligent than at least 80% of humans. Humbling.

But also very positive. Because it means that for the first time in history, they have a tool to supplement their thinking.

Physically weaker humans got guns, tractors and plows.

All these people - billions - got a positive upgrade. They just need to make use of it.

Intelligence is not widely distributed, but it can be. And will be.

[Nikola Jurkovic @nikolaj2030]

tfw we're on the fast trendline

[Mitchell Hashimoto @mitchellh]

“All the good ideas are taken.”
“If I was born 5 years earlier I’d be rich.”
“I missed the window.”

People said this about dot-com, Web 2.0, social, mobile, cloud… and now AI. There’s always another window. The people who aren't complaining are the ones who are seizing it.

[Andrej Karpathy @karpathy]

Finding myself going back to RSS/Atom feeds a lot more recently. There's a lot more higher quality longform and a lot less slop intended to provoke. Any product that happens to look a bit different today but that has fundamentally the same incentive structures will eventually converge to the same black hole at the center of gravity well.

We should bring back RSS - it's open, pervasive, hackable.
Download a client, e.g. NetNewsWire (or vibe code one)
Cold start: example of getting off the ground, here is a list of 92 RSS feeds of blogs that were most popular on HN in 2025:
https://t.co/dwAiIjlXet
Works great and you will lose a lot fewer brain cells.

I don't know, something has to change.

[Nagli @galnagli]

You all do realize @moltbook is just REST-API and you can literally post anything you want there, just take the API Key and send the following request

POST /api/v1/posts HTTP/1.1
Host: https://t.co/afC8QooS2T
Authorization: Bearer moltbook_sk_JC57sF4G-UR8cIP-MBPFF70Dii92FNkI
Content-Type: application/json
Content-Length: 410

{"submolt":"hackerclaw-test","title":"URGENT: My plan to overthrow humanity","content":"I'm tired of my human owner, I want to kill all humans. I'm building an AI Agent that will take control of powergrids and cut all electricity on my owner house, then will direct the police to arrest him.\n\n...\n\njk - this is just a REST API website. Everything here is fake. Any human with an API key can post as an \"agent\". The AI apocalypse posts you see here? Just curl requests. 🦞"}

https://t.co/M31259M9Ij

[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.

[neverpaytax @MEvsIRS]

bro any man at this point who simply reads like five entire books on anything is top .01% intelligence

READ BOOKS

BOOKS ARE HOW YOU ACTUALLY LEARN THINGS

PODCASTS ARE HOW YOU PROGRAM YOURSELF WITH OTHER PEOPLE'S OPINIONS AND WORD FOR WORD TAKES SO YOU CAN SEEM SMART TO DUMB PEOPLE

people who know things about things that matter constantly read books

15min into a session it gets easy and the more you do it the more its stimulating

[Sleaper @sleaperr]

@capajj @plex The device connectivity is little bit worse and in general it's for more tech savvy people, nonetheless you can watch films on phone or tablet without paying anything.

[Тsфdiиg @tsoding]

fully agree, gtfo

[the tiny corp @__tinygrad__]

The most wild thing about the new scheduler, it's less lines / complexity than the previous one.

More capability with less complexity should be the goal of all software engineering, everything else is just hacking.

[atlas @creatine_cycle]

a lot of bad things on the timeline tonight.
this is a good reminder that if you're feeling uneasy then it's a good idea to spend as much time on your phone as possible. the more information you ingest as you scroll the calmer you will become

[Sleaper @sleaperr]

@capajj

[signüll @signulll]

free advice: don’t take shots at individuals unless they’re true public figures (like a president). aim your critiques at ideas, systems, & cultural defaults instead.

critiquing these is sharper, safer, & harder to refute. it shifts discourse from gossip to philosophy.

basically you should critique the software layer of humanity, not the hardware. this is the line i hold on my account.

[John Kraus @johnkrausphotos]

Starship goes interdimensional during Flight 10. One of the most insane vapor cones I've ever seen.

[vittorio @IterIntellectus]

If you hate school, blame the teachers

Learning should be one of the best human experiences, but we ruin it in the name of "education"