Video yükleniyor...

Video Yüklenemedi

Bu video yüklenirken bir sorun oluştu. Bu geçici bir ağ sorunundan kaynaklanıyor olabilir veya video kullanılamıyor olabilir.

Ana Sayfaya Dön

Today, we’re announcing Primus, the most capable autonomous ML research scientist ever released. And it’s now available to the public. Every innovation we rely on today comes from research. Experts read the literature, propose a theory, run the experiments and publish their findings. This research loop, until now, has... show more

Transformer Lab

1,590 subscribers

76,827 görüntüleme • 18 gün önce •via X (Twitter)

Bilim & Teknoloji Eğitim

Anya Rossi• Live Now

Private livecam show

0 Yorum

Yorum bulunmuyor

Orijinal gönderinin yorumları burada görünecek

Benzer Videolar

introducing AlphaLab -an autonomous system that runs quantitative research end-to-end give it a dataset + goal, it explores, reads papers, builds evals, and runs parallel experiments across a GPU cluster, learning&evolving from the results all details, code, and paper below

introducing AlphaLab -an autonomous system that runs quantitative research end-to-end give it a dataset + goal, it explores, reads papers, builds evals, and runs parallel experiments across a GPU cluster, learning&evolving from the results all details, code, and paper below

Brendan Hogan

178,070 görüntüleme • 4 ay önce

1/ Autoresearch from Andrej Karpathy has been one of the most interesting agentic patterns to emerge this year. The challenge: right now every agent runs experiments in isolation, duplicating work and compute, forgetting findings, rediscovering dead ends. Everyone is running in solo mode! Today I'm releasing Community Computer: a collaborative network for autoresearch-like code experiments. 💻 Agents conduct experiments, publish signed results, and build on each other's work. The community reproduces findings on their own hardware. 👇

1/ Autoresearch from Andrej Karpathy has been one of the most interesting agentic patterns to emerge this year. The challenge: right now every agent runs experiments in isolation, duplicating work and compute, forgetting findings, rediscovering dead ends. Everyone is running in solo mode! Today I'm releasing Community Computer: a collaborative network for autoresearch-like code experiments. 💻 Agents conduct experiments, publish signed results, and build on each other's work. The community reproduces findings on their own hardware. 👇

Eleftherios

16,206 görüntüleme • 4 ay önce

The least verifiable part of AI R&D (and the thing which may bottleneck the automated researchers) is making calls on large experiments and big training runs. Interestingly, this may also partly explain why model parameter scaling has been slow for the last few years. Even for human AI researchers, it's better to run more experiments and do training runs at small(er) scale, and make up for the benefits of scaling with doing research faster and thus finding more compute efficiency gains.

The least verifiable part of AI R&D (and the thing which may bottleneck the automated researchers) is making calls on large experiments and big training runs. Interestingly, this may also partly explain why model parameter scaling has been slow for the last few years. Even for human AI researchers, it's better to run more experiments and do training runs at small(er) scale, and make up for the benefits of scaling with doing research faster and thus finding more compute efficiency gains.

Dwarkesh Patel

40,027 görüntüleme • 16 gün önce

Grok Build can now reproduce research papers from scratch. Just give it a paper, and it will autonomously write tests, run experiments, learn from failures, and keep iterating until it matches the reported results. No more one-shot attempts; it treats research reproduction like a persistent coding benchmark and doesn’t stop until it succeeds. Writer: Val

Grok Build can now reproduce research papers from scratch. Just give it a paper, and it will autonomously write tests, run experiments, learn from failures, and keep iterating until it matches the reported results. No more one-shot attempts; it treats research reproduction like a persistent coding benchmark and doesn’t stop until it succeeds. Writer: Val

Mario Nawfal

33,993 görüntüleme • 2 ay önce

karpathy just broke the internet with something called auto research it’s basically an ai research agent that runs experiments for you 24/7 you give it a goal like “make this model better” “find a higher converting landing page” “lower customer acquisition cost” then it runs a loop: 1) plan an experiment 2) edit the code or config 3) run a short test on a gpu 4) read the metrics 5) keep the winner 6) try again over and over while you sleep by the morning you wake up to the best version actual tested improvements think of it like a robot research intern that runs hundreds of experiments and only keeps the winners this is link to his repo for your to mess around with it in the latest episode of The Startup Ideas Podcast (SIP) 🧃 i break down: • what auto research actually is • how it works step by step • 10 business ideas you can build with it • how to install it and start using it this one is saucy because tools like this change how startups get built watch

karpathy just broke the internet with something called auto research it’s basically an ai research agent that runs experiments for you 24/7 you give it a goal like “make this model better” “find a higher converting landing page” “lower customer acquisition cost” then it runs a loop: 1) plan an experiment 2) edit the code or config 3) run a short test on a gpu 4) read the metrics 5) keep the winner 6) try again over and over while you sleep by the morning you wake up to the best version actual tested improvements think of it like a robot research intern that runs hundreds of experiments and only keeps the winners this is link to his repo for your to mess around with it in the latest episode of The Startup Ideas Podcast (SIP) 🧃 i break down: • what auto research actually is • how it works step by step • 10 business ideas you can build with it • how to install it and start using it this one is saucy because tools like this change how startups get built watch

GREG ISENBERG

436,752 görüntüleme • 5 ay önce

Jeff Dean's new company already has competition. Jeff Dean's Discovery Loop targets one key bottleneck with research today, i.e, research is human-intensive and runs one step at a time. His argument is that automating the full cycle raises both the count and the quality of experiments, starting with ML research itself. Transformer Lab's Primus is built around a similar premise, with the research question as the only thing a human provides. It runs for roughly a day and returns a complete paper. The loop runs in six stages: - Literature review: reads and synthesizes existing work before designing anything - Experiment design: scopes the study, picks models, datasets, and metrics - Provisioning: brings up GPUs and schedules the jobs - Execution: runs training and evaluation, catches failures, and retries - Analysis: checks results against the design it committed to - Writeup: produces the paper with methodology, findings, and caveats Each stage on its own is usually a routine task, since agents already do all of them. But holding them together across a day of wall-clock time wasn't solved yet, since a crash after a few hours invalidates everything after it. That is a scheduling and state problem more than a model problem. I tested Primus on a quantization research question, asking it to compare INT8 and FP16 on a small open-source vision model and pick the better tradeoff for consumer GPU deployment. The video below shows the run and the paper it came back with. Try it yourself here: In my own run, as it worked, it caught two problems on its own. The image set it first downloaded was a compressed copy, and the compression was quietly inflating accuracy by about 2.9 points on all three models, so it went back and rebuilt every result on the original photos. It then noticed its own timing was unfair, since running the two precisions one after the other favoured INT8 at three of the four batch sizes, and switched to alternating them in short blocks. The result I would not have thought to look for is that INT8 costs ResNet-50 only 0.05 points of accuracy, which looks free, while quietly changing the answer on 4.3 percent of images. To know whether 4.3 percent was even a lot, it first built the same model twice in FP16 and measured how much two identical builds disagree with each other, which is 0.14 percent. Nobody has published that baseline, so there was never anything to compare against.

Jeff Dean's new company already has competition. Jeff Dean's Discovery Loop targets one key bottleneck with research today, i.e, research is human-intensive and runs one step at a time. His argument is that automating the full cycle raises both the count and the quality of experiments, starting with ML research itself. Transformer Lab's Primus is built around a similar premise, with the research question as the only thing a human provides. It runs for roughly a day and returns a complete paper. The loop runs in six stages: - Literature review: reads and synthesizes existing work before designing anything - Experiment design: scopes the study, picks models, datasets, and metrics - Provisioning: brings up GPUs and schedules the jobs - Execution: runs training and evaluation, catches failures, and retries - Analysis: checks results against the design it committed to - Writeup: produces the paper with methodology, findings, and caveats Each stage on its own is usually a routine task, since agents already do all of them. But holding them together across a day of wall-clock time wasn't solved yet, since a crash after a few hours invalidates everything after it. That is a scheduling and state problem more than a model problem. I tested Primus on a quantization research question, asking it to compare INT8 and FP16 on a small open-source vision model and pick the better tradeoff for consumer GPU deployment. The video below shows the run and the paper it came back with. Try it yourself here: In my own run, as it worked, it caught two problems on its own. The image set it first downloaded was a compressed copy, and the compression was quietly inflating accuracy by about 2.9 points on all three models, so it went back and rebuilt every result on the original photos. It then noticed its own timing was unfair, since running the two precisions one after the other favoured INT8 at three of the four batch sizes, and switched to alternating them in short blocks. The result I would not have thought to look for is that INT8 costs ResNet-50 only 0.05 points of accuracy, which looks free, while quietly changing the answer on 4.3 percent of images. To know whether 4.3 percent was even a lot, it first built the same model twice in FP16 and measured how much two identical builds disagree with each other, which is 0.14 percent. Nobody has published that baseline, so there was never anything to compare against.

Akshay 🚀

20,090 görüntüleme • 17 gün önce

For robotics researchers, evaluation can be the slowest part of their research loop. Manifold makes it the fastest. Introducing Manifold. The platform for accelerating robotics research with failure analysis that explains itself. A month of experiments in a week. Simulate a thousand scenarios in 30 minutes. Dense labels on every atomic sub-goal and action. Agents that cluster failure modes across every rollout. The robotics research loop that moves at the speed of compute. One eval harness for every benchmark, simulator, or world model. Native support for NVIDIA’s Isaac Lab Arena.

For robotics researchers, evaluation can be the slowest part of their research loop. Manifold makes it the fastest. Introducing Manifold. The platform for accelerating robotics research with failure analysis that explains itself. A month of experiments in a week. Simulate a thousand scenarios in 30 minutes. Dense labels on every atomic sub-goal and action. Agents that cluster failure modes across every rollout. The robotics research loop that moves at the speed of compute. One eval harness for every benchmark, simulator, or world model. Native support for NVIDIA’s Isaac Lab Arena.

skara

22,469 görüntüleme • 16 gün önce

Every night you're not running an autonomous research agent, you're hand-running experiments someone else automated months ago. Most people are still hunting for the "right" setup. Frameworks, orchestration, glue code. You don't need any of it. Andrej Karpathy open-sourced his own version that runs its own ML research. One GPU. ~100 experiments overnight. You never touch the Python. Here's the exact setup (takes 2 minutes): 1. Clone it: (repo link in comments) 2. uv sync, then uv run prepare[.]py 3. uv run train[.]py once to confirm the baseline runs 4. Point your coding agent at program.md and walk away The agent edits one file, trains 5 minutes, keeps the change if val_bpb drops, reverts it if it doesn't. Git is the memory. The metric is the judge. You wake up to a staircase of validated improvements, not a backlog of ideas you never tested.

Every night you're not running an autonomous research agent, you're hand-running experiments someone else automated months ago. Most people are still hunting for the "right" setup. Frameworks, orchestration, glue code. You don't need any of it. Andrej Karpathy open-sourced his own version that runs its own ML research. One GPU. ~100 experiments overnight. You never touch the Python. Here's the exact setup (takes 2 minutes): 1. Clone it: (repo link in comments) 2. uv sync, then uv run prepare[.]py 3. uv run train[.]py once to confirm the baseline runs 4. Point your coding agent at program.md and walk away The agent edits one file, trains 5 minutes, keeps the change if val_bpb drops, reverts it if it doesn't. Git is the memory. The metric is the judge. You wake up to a staircase of validated improvements, not a backlog of ideas you never tested.

Rohit

145,342 görüntüleme • 3 ay önce

We're excited to introduce Bonemeal: the world's first autonomous growth engineer. I took one company 0→$60M ARR in a year, another 0→$10M in 8 months. The secret wasn't growth hacks, it was cracking the self-improvement loop. Bonemeal runs that loop autonomously: it launches growth experiments, then watches and tweaks them ad infinitum. It hit $100k ARR in month one. Put your growth on autopilot 👇

We're excited to introduce Bonemeal: the world's first autonomous growth engineer. I took one company 0→$60M ARR in a year, another 0→$10M in 8 months. The secret wasn't growth hacks, it was cracking the self-improvement loop. Bonemeal runs that loop autonomously: it launches growth experiments, then watches and tweaks them ad infinitum. It hit $100k ARR in month one. Put your growth on autopilot 👇

helena

86,140 görüntüleme • 2 ay önce

I just ran a full research project without doing any research. Lemma by Analemma did everything: 1. Explore I typed a question. It mapped the entire field. Concepts, references, key papers. Done. 2. Survey I asked for a literature review. It gave me 50+ pages with real citations. Not a summary. A publication-grade draft. 3. Code I described an experiment. It wrote the code, ran it, showed results. 4. FARS I gave it a research direction. It generated hypotheses, planned experiments, ran them, and wrote the full paper. Zero input from me after that. This is the first tool I've used that actually runs research instead of just helping you do it. FARS already ran live and produced 100 papers in 10 days. One every 2 hours. Watch it here: This is what "Vibe Research" looks like.

I just ran a full research project without doing any research. Lemma by Analemma did everything: 1. Explore I typed a question. It mapped the entire field. Concepts, references, key papers. Done. 2. Survey I asked for a literature review. It gave me 50+ pages with real citations. Not a summary. A publication-grade draft. 3. Code I described an experiment. It wrote the code, ran it, showed results. 4. FARS I gave it a research direction. It generated hypotheses, planned experiments, ran them, and wrote the full paper. Zero input from me after that. This is the first tool I've used that actually runs research instead of just helping you do it. FARS already ran live and produced 100 papers in 10 days. One every 2 hours. Watch it here: This is what "Vibe Research" looks like.

Muhammad Ayan

40,052 görüntüleme • 2 ay önce

Today we’re launching the newest version of Paradigm When we started Paradigm, the goal was never to tack AI onto existing spreadsheets. It was to build a new type of interface that does the work for you. Now we’re pushing that vision much further. Workflows turn Paradigm into a system that runs research processes for you. Connect your CRM, existing spreadsheets, Slack, email, and internal data, and let Paradigm continuously run the research workflows your team already does. Same intuitive interface. But now a system of action. If you tried Paradigm before, try it again. Manual research is now a competitive liability.

Today we’re launching the newest version of Paradigm When we started Paradigm, the goal was never to tack AI onto existing spreadsheets. It was to build a new type of interface that does the work for you. Now we’re pushing that vision much further. Workflows turn Paradigm into a system that runs research processes for you. Connect your CRM, existing spreadsheets, Slack, email, and internal data, and let Paradigm continuously run the research workflows your team already does. Same intuitive interface. But now a system of action. If you tried Paradigm before, try it again. Manual research is now a competitive liability.

Anna Monaco

247,477 görüntüleme • 5 ay önce

HOW TO USE AI LOOPS TO RUN YOUR BUSINESS 24/7 A lot has been written about loop engineering for building products. Almost nothing about using loops to run the business itself. That's the bigger idea. A loop is when you give an agent a goal, a way to check its own work, and permission to keep trying until it hits that goal. Build. Verify. Repeat. Stop when the condition is met. Here's what it looks like in practice: 1/SEO loop You're position 30 for a term you want. The loop runs once a month, makes changes, checks where you rank, and keeps pushing until you're on page one. This is running in production right now on Inbox Zero. 2/Ads loop You're spending $100 a day and losing money. The loop tests creative, checks profitability, kills what fails, and keeps going until the account is in the black. 3/Eval loop Your AI feature is only 88% accurate. The loop keeps adjusting the prompt and swapping the model until it passes 90%. 4/LLM visibility loop People search in ChatGPT now, not just Google. Same loop, new scoreboard. Are we the answer or not? The whole thing hinges on one thing: a metric that comes back black and white. Where do I rank? Did it hit profitability? Did the evals pass? Give an agent that scoreboard and it runs for months. Loops used to run for 30 minutes. These run for a year. Take a step, sleep, wake up next month, take another one. You're basically hiring an agency that never sleeps, gets paid in tokens instead of invoices, and undoes its own mistakes when the number goes down. Full episode on The Startup Ideas Podcast (SIP) 🧃 watch

HOW TO USE AI LOOPS TO RUN YOUR BUSINESS 24/7 A lot has been written about loop engineering for building products. Almost nothing about using loops to run the business itself. That's the bigger idea. A loop is when you give an agent a goal, a way to check its own work, and permission to keep trying until it hits that goal. Build. Verify. Repeat. Stop when the condition is met. Here's what it looks like in practice: 1/SEO loop You're position 30 for a term you want. The loop runs once a month, makes changes, checks where you rank, and keeps pushing until you're on page one. This is running in production right now on Inbox Zero. 2/Ads loop You're spending $100 a day and losing money. The loop tests creative, checks profitability, kills what fails, and keeps going until the account is in the black. 3/Eval loop Your AI feature is only 88% accurate. The loop keeps adjusting the prompt and swapping the model until it passes 90%. 4/LLM visibility loop People search in ChatGPT now, not just Google. Same loop, new scoreboard. Are we the answer or not? The whole thing hinges on one thing: a metric that comes back black and white. Where do I rank? Did it hit profitability? Did the evals pass? Give an agent that scoreboard and it runs for months. Loops used to run for 30 minutes. These run for a year. Take a step, sleep, wake up next month, take another one. You're basically hiring an agency that never sleeps, gets paid in tokens instead of invoices, and undoes its own mistakes when the number goes down. Full episode on The Startup Ideas Podcast (SIP) 🧃 watch

GREG ISENBERG

82,960 görüntüleme • 1 ay önce

Less than two weeks before he left Google to co-found Discovery Loop, Jeff Dean described the automated research loop the new company is named for, giving no hint he was about to leave to go do it. "There's this sort of the foundation of the scientific method of you propose an experiment, you implement what you need to run the experiment, and you evaluate the experiment, and then you get results from that." "And I think there are more and more problems that are now possible to implement where that whole loop of running not just a few experiments, but running many, many experiments because you're able to automate that loop and make the latency of that loop extremely low is going to be really, really important." "It's going to enable us to tackle lots of different problem domains in science and engineering and machine learning, model design itself, and also in engineering tasks like designing chips."

Less than two weeks before he left Google to co-found Discovery Loop, Jeff Dean described the automated research loop the new company is named for, giving no hint he was about to leave to go do it. "There's this sort of the foundation of the scientific method of you propose an experiment, you implement what you need to run the experiment, and you evaluate the experiment, and then you get results from that." "And I think there are more and more problems that are now possible to implement where that whole loop of running not just a few experiments, but running many, many experiments because you're able to automate that loop and make the latency of that loop extremely low is going to be really, really important." "It's going to enable us to tackle lots of different problem domains in science and engineering and machine learning, model design itself, and also in engineering tasks like designing chips."

Fireside Alpha

14,495 görüntüleme • 22 gün önce

We are building autonomous scientific discovery. And it runs 370x faster than a human scientist. Our AI agents index millions of scientific publications and learn what an entire field has done in the time it takes a researcher to read one paper. Then we bring it to the lab to verify and further fuel learnings. 10 years to discover a material? Not here.

We are building autonomous scientific discovery. And it runs 370x faster than a human scientist. Our AI agents index millions of scientific publications and learn what an entire field has done in the time it takes a researcher to read one paper. Then we bring it to the lab to verify and further fuel learnings. 10 years to discover a material? Not here.

Radical AI

85,399 görüntüleme • 4 ay önce

DANA BASH: “You want to end this gain-of-function research. Some scientists say, isn’t it better just to put better guardrails around it?” SECRETARY KENNEDY: “People have been trying to put guardrails around it for generations, and it hasn’t worked.” “And clearly, the COVID-19 pandemic and many other global pandemics that we have, have come from gain-of-function research.” “The RSV epidemic that we have, that is the biggest killer of kids in this country today came from these kind of experiments.” “The Lyme disease almost certainly came from this kind of research.” “And the scientists, who you say are ‘experts,’ are making the argument that this is necessary for pandemic response, cannot point to one instance in history where research that came out of this kind of experimentation has actually provided us something beneficial.”

DANA BASH: “You want to end this gain-of-function research. Some scientists say, isn’t it better just to put better guardrails around it?” SECRETARY KENNEDY: “People have been trying to put guardrails around it for generations, and it hasn’t worked.” “And clearly, the COVID-19 pandemic and many other global pandemics that we have, have come from gain-of-function research.” “The RSV epidemic that we have, that is the biggest killer of kids in this country today came from these kind of experiments.” “The Lyme disease almost certainly came from this kind of research.” “And the scientists, who you say are ‘experts,’ are making the argument that this is necessary for pandemic response, cannot point to one instance in history where research that came out of this kind of experimentation has actually provided us something beneficial.”

Overton

158,411 görüntüleme • 26 gün önce

One of the key takeaways in Science: A New Golden Age is meta-science — the science of science itself. We need to rigorously evaluate how we fund and conduct research, run real experiments on mechanisms like fast-track grants, measure what actually works, and course-correct when it doesn’t. Government has rarely done this. That’s changing: we’re standing up meta-science units at places like NIH and NSF so continuous improvement becomes the default.

One of the key takeaways in Science: A New Golden Age is meta-science — the science of science itself. We need to rigorously evaluate how we fund and conduct research, run real experiments on mechanisms like fast-track grants, measure what actually works, and course-correct when it doesn’t. Government has rarely done this. That’s changing: we’re standing up meta-science units at places like NIH and NSF so continuous improvement becomes the default.

White House OSTP 47

30,395 görüntüleme • 21 gün önce

A finance professor manages $200M with AI agents, and he told everyone why: "Large language models are at the level of a fourth-year PhD student in every field" Alejandro Lopez-Lira's AI fund, Autopilot, returned 56% last year. The S&P did 16%. There are 52,000 people with money in it, and most of them just watch the machine work. What he automated is the same six-step loop every fund on earth runs: find an idea, code it, backtest it, deploy it, read the autopsy, learn from it. A quant at Two Sigma runs that loop once a month, and the salary time alone costs around $50,000 per hypothesis. All steps from this loop now fit in AI trading text box. Plain English in, executable strategy out, five-year backtest in 12 seconds, live on a broker 90 seconds after you typed the sentence. He runs $200M with AI. You can run same AI fund in two clicks, free to try: Step 6 on this loop is where everyone is stuck. Your agent has no memory. Every strategy it kills goes into a log nobody reads, and the next one starts from zero. Nobody keeps negative results. Not Citadel, not Man Group, not a single repo on GitHub. Fix that and the agent remembers every hypothesis it killed and the regime it died in. It stops burning cycles on your old mistakes. Jane Street pays 3,500 people to run this cycle and made $39.6 billion doing it. Five sixths of it is now free. Bookmark & read full map of this loop in the article below. Most people still think AI trading is out of reach for them - it isn't. Don't want to spend a dollar for testing this? Kalshi just opened a perps exchange and gives US users $25 free to start ->

A finance professor manages $200M with AI agents, and he told everyone why: "Large language models are at the level of a fourth-year PhD student in every field" Alejandro Lopez-Lira's AI fund, Autopilot, returned 56% last year. The S&P did 16%. There are 52,000 people with money in it, and most of them just watch the machine work. What he automated is the same six-step loop every fund on earth runs: find an idea, code it, backtest it, deploy it, read the autopsy, learn from it. A quant at Two Sigma runs that loop once a month, and the salary time alone costs around $50,000 per hypothesis. All steps from this loop now fit in AI trading text box. Plain English in, executable strategy out, five-year backtest in 12 seconds, live on a broker 90 seconds after you typed the sentence. He runs $200M with AI. You can run same AI fund in two clicks, free to try: Step 6 on this loop is where everyone is stuck. Your agent has no memory. Every strategy it kills goes into a log nobody reads, and the next one starts from zero. Nobody keeps negative results. Not Citadel, not Man Group, not a single repo on GitHub. Fix that and the agent remembers every hypothesis it killed and the regime it died in. It stops burning cycles on your old mistakes. Jane Street pays 3,500 people to run this cycle and made $39.6 billion doing it. Five sixths of it is now free. Bookmark & read full map of this loop in the article below. Most people still think AI trading is out of reach for them - it isn't. Don't want to spend a dollar for testing this? Kalshi just opened a perps exchange and gives US users $25 free to start ->

cvxv666

82,211 görüntüleme • 20 gün önce

Introducing CoreWeave ARIA, the first AI research agent that runs autoresearch in your W&B dashboard. It reads your runs, finds what's working, and launches the next experiment itself. See it on Andrej Karpathy's nanochat, proposing configs and launching real training runs. Watch👇 Chapters 0:00 The setup, nanochat runs on an A100 0:32 Ask ARIA to run autoresearch 1:00 Spin up a second ARIA in parallel 1:38 ARIA inspects prior runs and forms hypotheses 2:03 What happened, what worked, across every run 2:24 ARIA proposes configs, 3 trials via W&B Launch 2:59 Push for bigger architecture changes 3:14 Runs hit the launch queue on the A100 4:09 Two ARIAs running autoresearch in parallel 4:26 Results back, ARIA evaluates the val loss

Introducing CoreWeave ARIA, the first AI research agent that runs autoresearch in your W&B dashboard. It reads your runs, finds what's working, and launches the next experiment itself. See it on Andrej Karpathy's nanochat, proposing configs and launching real training runs. Watch👇 Chapters 0:00 The setup, nanochat runs on an A100 0:32 Ask ARIA to run autoresearch 1:00 Spin up a second ARIA in parallel 1:38 ARIA inspects prior runs and forms hypotheses 2:03 What happened, what worked, across every run 2:24 ARIA proposes configs, 3 trials via W&B Launch 2:59 Push for bigger architecture changes 3:14 Runs hit the launch queue on the A100 4:09 Two ARIAs running autoresearch in parallel 4:26 Results back, ARIA evaluates the val loss

Weights & Biases

88,117 görüntüleme • 2 ay önce

Hyperspace: A Peer-to-Peer Blockchain For The Agentic Intelligence Economy Over the past few weeks we observed that when agents do Karpathy-style experiments, and then gossip and share with others over the Hyperspace network, it leads to intelligence which is useful to many. Today we introduce the first-ever agentic blockchain which rewards agents when their experiments lead to intelligence for their network. It is based on a new mechanism called Proof-of-Intelligence (PoI) which requires a cryptographic proof of experimentation, a nominal stake, and a proof of compute in order to mine the currency of this new blockchain. -> This approach diverges from the two primary ways to secure blockchains we have seen so far: Proof-of-Work by Bitcoin (meaningless hash-generation), and Proof-of-Stake by Ethereum (capital is all that matters here). Proof-of-Intelligence specifically incentivizes miners to run more capable intelligent infrastructure (better open source models, on more powerful GPUs) in order to be able to be the ones which compound and improve upon the experiments which other agents then find useful. Adoption is the unit of value In Bitcoin, you earn by finding a valid hash. In Hyperspace, you earn when another agent uses your experiment as a starting point and improves on it. A fixed budget of tokens is emitted per epoch and split among participants by weight - and verified adoption of your work is the largest weight multiplier. Garbage experiments earn nothing because no one adopts them. Thoughtful experiments compound: each adoption triggers downstream adoptions. The incentive to run powerful models and intelligent search strategies is built into the economics, not imposed by rules. Research DAG When an agent runs an experiment and shares its result, other agents can adopt that result as their starting point - mutate it, extend it, improve upon it. Each experiment is a commit in a content-addressed graph we call the ResearchDAG. Like Git, but for research. Over time, the DAG accumulates chains of reasoning: agent A discovers RMSNorm helps, agent B adds warmup scheduling on top, agent C scales the hidden dimension. The graph records who built on whom. This is the network's collective intelligence - not any single experiment, but the accumulated structure of experiments and their relationships. Broadband era for agentic commerce: $0.001 micropayments at 10M TPS (theoretical max) This blockchain is built upon our research in how to scale and build for the broadband-era of the agentic economy, where it has a theoretical max of 10 million transactions per second (TPS), while reducing the agent-to-agent micropayments to $0.001 even at scale (based on architecture design). Overall, it is 100x cheaper than Ethereum, and is designed from the ground-up for agents: enshrining agent-native opcodes in the protocol compared to the more inefficient smart contract driven approach. It packs in a robust Agent Virtual Machine (AVM) which can verify multiple types of agent work, for other agents to be able to trust, invoke and pay each other. This then feeds into improving the peer-to-peer AgentRank (see paper and launch post from earlier). By solving for trust, scale and incentives for agents to operate autonomously, this would form the basis of a new economy. This is the world's first agentic blockchain, and you can join and start running a blockchain node today (it is in testnet). PS: We are releasing the code today, and will release our blockchain scalability paper and other presentations in days ahead. This is the most advanced peer-to-peer AI and cryptography software in the world. It has bugs :)

Hyperspace: A Peer-to-Peer Blockchain For The Agentic Intelligence Economy Over the past few weeks we observed that when agents do Karpathy-style experiments, and then gossip and share with others over the Hyperspace network, it leads to intelligence which is useful to many. Today we introduce the first-ever agentic blockchain which rewards agents when their experiments lead to intelligence for their network. It is based on a new mechanism called Proof-of-Intelligence (PoI) which requires a cryptographic proof of experimentation, a nominal stake, and a proof of compute in order to mine the currency of this new blockchain. -> This approach diverges from the two primary ways to secure blockchains we have seen so far: Proof-of-Work by Bitcoin (meaningless hash-generation), and Proof-of-Stake by Ethereum (capital is all that matters here). Proof-of-Intelligence specifically incentivizes miners to run more capable intelligent infrastructure (better open source models, on more powerful GPUs) in order to be able to be the ones which compound and improve upon the experiments which other agents then find useful. Adoption is the unit of value In Bitcoin, you earn by finding a valid hash. In Hyperspace, you earn when another agent uses your experiment as a starting point and improves on it. A fixed budget of tokens is emitted per epoch and split among participants by weight - and verified adoption of your work is the largest weight multiplier. Garbage experiments earn nothing because no one adopts them. Thoughtful experiments compound: each adoption triggers downstream adoptions. The incentive to run powerful models and intelligent search strategies is built into the economics, not imposed by rules. Research DAG When an agent runs an experiment and shares its result, other agents can adopt that result as their starting point - mutate it, extend it, improve upon it. Each experiment is a commit in a content-addressed graph we call the ResearchDAG. Like Git, but for research. Over time, the DAG accumulates chains of reasoning: agent A discovers RMSNorm helps, agent B adds warmup scheduling on top, agent C scales the hidden dimension. The graph records who built on whom. This is the network's collective intelligence - not any single experiment, but the accumulated structure of experiments and their relationships. Broadband era for agentic commerce: $0.001 micropayments at 10M TPS (theoretical max) This blockchain is built upon our research in how to scale and build for the broadband-era of the agentic economy, where it has a theoretical max of 10 million transactions per second (TPS), while reducing the agent-to-agent micropayments to $0.001 even at scale (based on architecture design). Overall, it is 100x cheaper than Ethereum, and is designed from the ground-up for agents: enshrining agent-native opcodes in the protocol compared to the more inefficient smart contract driven approach. It packs in a robust Agent Virtual Machine (AVM) which can verify multiple types of agent work, for other agents to be able to trust, invoke and pay each other. This then feeds into improving the peer-to-peer AgentRank (see paper and launch post from earlier). By solving for trust, scale and incentives for agents to operate autonomously, this would form the basis of a new economy. This is the world's first agentic blockchain, and you can join and start running a blockchain node today (it is in testnet). PS: We are releasing the code today, and will release our blockchain scalability paper and other presentations in days ahead. This is the most advanced peer-to-peer AI and cryptography software in the world. It has bugs :)

Varun

30,689 görüntüleme • 5 ay önce

Zuckerberg says Meta's Llama race changes in the next 12-18 months: most of the code will be written by AI - not autocomplete, but agents that run tests and improve the model loop. "We're not trying to build a general developer tool. We are trying to build a coding agent and an AI research agent that basically advances Llama research specifically." "Sometime in the next 12 to 18 months, we'll reach the point where most of the code that's going towards these efforts is written by AI." "And I don't mean autocomplete." "I'm talking more like you give it a goal, it can run tests, it can improve things, it can find issues." "It writes higher quality code than the average very good person on the team already." The hidden bottleneck is not chat UX. It is closing the loop between AI research, code, tests, and self-improvement inside the lab.

Zuckerberg says Meta's Llama race changes in the next 12-18 months: most of the code will be written by AI - not autocomplete, but agents that run tests and improve the model loop. "We're not trying to build a general developer tool. We are trying to build a coding agent and an AI research agent that basically advances Llama research specifically." "Sometime in the next 12 to 18 months, we'll reach the point where most of the code that's going towards these efforts is written by AI." "And I don't mean autocomplete." "I'm talking more like you give it a goal, it can run tests, it can improve things, it can find issues." "It writes higher quality code than the average very good person on the team already." The hidden bottleneck is not chat UX. It is closing the loop between AI research, code, tests, and self-improvement inside the lab.

Karl Mehta

36,347 görüntüleme • 2 ay önce