fresh

13.6K posts

fresh banner
fresh

fresh

@YoSoy_Fresh

$nfd $dog

Katılım Mayıs 2018
1.9K Takip Edilen440 Takipçiler
fresh retweetledi
Nova
Nova@badattrading_·
The era of rugs and bundled scams should come to an end very soon just like the remove liquidity rugs disappeared slowly last year (we don't see them anymore now), same thing will happen to all the shit we've seen so far. More tools will be developed to protect basic memecoins buyers. Pumpfun, Bonk and all the others aren't interested in protecting you because all their business rely on volume from ruggers and money laundering. The first dev who makes the right tool that can detect bundled/cabaled garbage with 99% accuracy wins the game.
English
126
101
721
44.3K
fresh retweetledi
Sir Doge of the Coin ⚔️
Sir Doge of the Coin ⚔️@dogeofficialceo·
Natural Retardation is far more dangerous than Artificial Intelligence
English
35
16
127
11.1K
fresh retweetledi
Phantom
Phantom@phantom·
@earthcurated the lion doesn't concern himself with a little male pattern baldness
Phantom tweet media
English
94
44
467
45.8K
fresh retweetledi
Autumn
Autumn@Autumn0365·
OMG wallstreetbets doing this again but now it will be their BIGGEST enemy USGOV operation is goin viral, 10k upvotes in 9h THE DEFLATION GLITCH describe $2 Bills as ultimate hedge against Fed spend $2, people think they’re rare and keep them so they disappear from circulation
Autumn tweet media
English
14
30
73
11.8K
fresh retweetledi
Andrew Jiang
Andrew Jiang@andrewjiang·
The brilliance of @karpathy is being able to distill vastly complex concepts and make them simple to understand and implement at a small scale. All it took was Claude Code and $10 on @runpod to spin up a single H100, and I had a world class ML researcher working on autopilot. I'm taking the general concept of autoresearch and applying it to an inference pipeline I've been working on (no GPU needed thankfully). Everything is so fun now.
Andrew Jiang tweet media
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
54
164
2.4K
319.7K
fresh
fresh@YoSoy_Fresh·
$pizza G9ivB7K41a4G8m1k4QdxxN4L5eGKL7Mr12S26B85pump
HT
0
0
2
237
fresh retweetledi
Prometheus
Prometheus@_pr0m3th3uss_·
I bought 10x boost for the community to help with trending who wants to match me?
Prometheus tweet media
English
15
5
47
763
fresh retweetledi
Farzad 🇺🇸 🇮🇷
Farzad 🇺🇸 🇮🇷@farzyness·
I asked Grok 420 to dumb this down for all of us retards: ### What actually happened Karpathy (a famous AI guy) built a little AI helper called “autoresearch.” He left it alone for 2 days on a tiny version of his project called nanochat. That little AI ran **700 experiments by itself**. It kept looking at the results, learning what worked, and trying smarter ideas next — exactly like a human researcher would do. It discovered **~20 real improvements** that no one (including Karpathy) had found in years of manual tweaking. He tested them on a bigger model and they all added up nicely. Result? His project now trains **11% faster** — the official leaderboard score dropped from 2.02 hours to 1.80 hours. That might sound small, but in AI it’s a *huge* win. ### Why this feels like a giant deal Normally Karpathy (and every AI engineer) does this the old-fashioned way for 20 years: - Think of an idea - Code it - Test it - Read papers for inspiration - Repeat forever It’s slow, tiring, and you miss stuff. This time the AI did the **entire loop by itself** — no human in the loop for days. It spotted things like: - “Hey, you forgot a tiny knob that makes the model pay attention better” - “Your model likes extra guardrails on one part (regularization) — you weren’t using any!” - “Your attention window is too narrow — let’s open it up” - “Your optimizer settings are weird — here’s the fix” - Better schedules and starting points, etc. These are the kind of tiny-but-powerful tweaks that usually take humans weeks or months to find. The AI found them in 2 days and they all stacked together. ### The bigger picture (this is the exciting part) Karpathy says: **“Every big AI lab is going to do this. It’s the final boss battle.”** Here’s what he means in plain English: 1. Start with a swarm of AIs (not just one). 2. Let them experiment on cheap, tiny models. 3. Keep only the best ideas and try them on bigger models. 4. Humans only jump in for the really hard stuff. Repeat → you get faster and smarter models with way less human work. And it’s not just for training speed. **Any goal you can measure** (speed, accuracy, cost, safety, creativity — whatever) can be handed to an AI swarm if you have a quick way to test it. Your problem might be next. ### Bottom line This is the first time Karpathy watched an AI do his entire 20-year job **autonomously** and actually beat him at it. It’s not sci-fi anymore — it’s happening right now on a public GitHub project. That’s why it feels like a giant deal. It’s the moment we went from “humans tune the AI” to “AI tunes the AI… and does a shockingly good job.” Mind officially blown? Yeah, mine too. 😄
Andrej Karpathy@karpathy

Three days ago I left autoresearch tuning nanochat for ~2 days on depth=12 model. It found ~20 changes that improved the validation loss. I tested these changes yesterday and all of them were additive and transferred to larger (depth=24) models. Stacking up all of these changes, today I measured that the leaderboard's "Time to GPT-2" drops from 2.02 hours to 1.80 hours (~11% improvement), this will be the new leaderboard entry. So yes, these are real improvements and they make an actual difference. I am mildly surprised that my very first naive attempt already worked this well on top of what I thought was already a fairly manually well-tuned project. This is a first for me because I am very used to doing the iterative optimization of neural network training manually. You come up with ideas, you implement them, you check if they work (better validation loss), you come up with new ideas based on that, you read some papers for inspiration, etc etc. This is the bread and butter of what I do daily for 2 decades. Seeing the agent do this entire workflow end-to-end and all by itself as it worked through approx. 700 changes autonomously is wild. It really looked at the sequence of results of experiments and used that to plan the next ones. It's not novel, ground-breaking "research" (yet), but all the adjustments are "real", I didn't find them manually previously, and they stack up and actually improved nanochat. Among the bigger things e.g.: - It noticed an oversight that my parameterless QKnorm didn't have a scaler multiplier attached, so my attention was too diffuse. The agent found multipliers to sharpen it, pointing to future work. - It found that the Value Embeddings really like regularization and I wasn't applying any (oops). - It found that my banded attention was too conservative (i forgot to tune it). - It found that AdamW betas were all messed up. - It tuned the weight decay schedule. - It tuned the network initialization. This is on top of all the tuning I've already done over a good amount of time. The exact commit is here, from this "round 1" of autoresearch. I am going to kick off "round 2", and in parallel I am looking at how multiple agents can collaborate to unlock parallelism. github.com/karpathy/nanoc… All LLM frontier labs will do this. It's the final boss battle. It's a lot more complex at scale of course - you don't just have a single train. py file to tune. But doing it is "just engineering" and it's going to work. You spin up a swarm of agents, you have them collaborate to tune smaller models, you promote the most promising ideas to increasingly larger scales, and humans (optionally) contribute on the edges. And more generally, *any* metric you care about that is reasonably efficient to evaluate (or that has more efficient proxy metrics such as training a smaller network) can be autoresearched by an agent swarm. It's worth thinking about whether your problem falls into this bucket too.

English
28
74
754
98.9K
fresh retweetledi
Wes Roth
Wes Roth@WesRoth·
who had money on Karpathy triggering The Intelligence Explosion?
Andrej Karpathy@karpathy

The next step for autoresearch is that it has to be asynchronously massively collaborative for agents (think: SETI@home style). The goal is not to emulate a single PhD student, it's to emulate a research community of them. Current code synchronously grows a single thread of commits in a particular research direction. But the original repo is more of a seed, from which could sprout commits contributed by agents on all kinds of different research directions or for different compute platforms. Git(Hub) is *almost* but not really suited for this. It has a softly built in assumption of one "master" branch, which temporarily forks off into PRs just to merge back a bit later. I tried to prototype something super lightweight that could have a flavor of this, e.g. just a Discussion, written by my agent as a summary of its overnight run: github.com/karpathy/autor… Alternatively, a PR has the benefit of exact commits: github.com/karpathy/autor… but you'd never want to actually merge it... You'd just want to "adopt" and accumulate branches of commits. But even in this lightweight way, you could ask your agent to first read the Discussions/PRs using GitHub CLI for inspiration, and after its research is done, contribute a little "paper" of findings back. I'm not actually exactly sure what this should look like, but it's a big idea that is more general than just the autoresearch repo specifically. Agents can in principle easily juggle and collaborate on thousands of commits across arbitrary branch structures. Existing abstractions will accumulate stress as intelligence, attention and tenacity cease to be bottlenecks.

English
15
15
393
34.8K
fresh retweetledi
ericosiu
ericosiu@ericosiu·
Karpathy's autoresearch isn't just for AI research. You can use it for business. Imagine running 700x more experiments: - Landing pages - Creatives - AEO/SEO - Pricing - E-mails - Cold outreach - Warm outreach - AP/AR - Procurement Revenues: 📈
ericosiu@ericosiu

x.com/i/article/2030…

English
43
134
1.6K
305.9K