𓀨 @heat21_uk I https://teapawt.net/t.co/VC6ruLqZsd I https://teapawt.net/t.co/hEaRYZvtPb I looking for incunabula
Luxembourg, London
Joined May 2013
- Tweets2.3K
- Following518
- Followers192
- Likes4.7K
Pavel Grachev retweeted
open-sourced cutcraft: a video editor where the file is the video. a model writes one TSX file, the browser renders it, and the timeline edits that same file. no project doc, no tool catalog, no second renderer for export
github.com/prefix-tree/cutcr…
Pavel Grachev retweeted
Wind barely blew yesterday (62,711.0 MWh total gen), leading to low curtailment (1,439.9 MWh). But even on this quiet day, wind farms collectively PAID the UK grid -£59,993. Moray Offshore Windfarm East L alone paid -£52,653. 💸💨📊 curtailcoin.com
Pavel Grachev retweeted
Some data we recently assembled on entrepreneurship/compute in Europe: eudata.vercel.app.
We hope that one of the useful roles that Stripe can play is in collecting and publishing empirical data pertaining to entrepreneurship and industry in Europe. There's growing appetite to get Europe on a better footing, and cross-sectional comparisons can often shine light on where opportunities lie. If you're interested in this kind of thing, we publish more at stripeeconomics.substack.com.
Pavel Grachev retweeted
Vesuvius Challenge has been quietly making a lot of progress; we read a full scroll, developed a new unwrapping algorithm with a global spiral prior, and we just launched a new $1M grand prize for reading one of the scrolls we've scanned.
There are also 10 $50k prizes available for finding text in a scroll we haven't found text in before, plus monthly progress prizes.
Join the discord!
A new phase BEGINS! 🚀
🏆 $1M Grand Prize
💰 $2M+ in open prizes
👑 $20K awarded every month to the best submission
scrollprize.org/
Will you read the next scroll? 👀
Spread the word and repost! 🔁
Pavel Grachev retweeted
Some cool visuals.
Dell Delivers world's first Nvidia Vera Rubin NVL72 rack to CoreWeave.
It packs 72 Rubin GPUs, 36 Vera CPUs, 3.6 exaFLOPS of FP4 inference, 75 TB of fast memory, and 260 TB/s NVLink bandwidth
Pavel Grachev retweeted
Dr. Lawrence White’s eight-hour course: Monetary Theory, is available now.
In this course, @lawrencehwhite1 explores the debate between market-based and government-controlled monetary systems, combining theory with historical evidence from gold standards to modern fiat currencies. We examine how money can emerge through market forces, the mechanics of fractional-reserve banking, and how central banks influence inflation.
The course also addresses banking stability, market failure arguments, and the causes of inflation, including seigniorage. It concludes by evaluating monetary policy rules, inflation targeting, and alternatives like Bitcoin, and asks whether sound money requires commodity backing or if proper institutional design can achieve stability in fiat systems.
Pavel Grachev retweeted
Tech sovereignty is so much bigger than AI. Congratulations to @harveyhodd and Rivan Industries - building synthetic natural gas right here in the UK!
(They're hiring for some excellent roles)
Pavel Grachev retweeted
Introducing Claude Managed Agents: everything you need to build and deploy agents at scale.
It pairs an agent harness tuned for performance with production infrastructure, so you can go from prototype to launch in days.
Now in public beta on the Claude Platform.
Pavel Grachev retweeted
296 hackers. 18,432 B200 GPU hours. $180K in compute credits. 60,000mg of caffeine.
We kicked off GTC with @fluidstack for From Silicon to Scale — the most full-stack AI hackathon ever built. 48 hours. 1 DGX Spark signed by Jensen. Here's what happened.(1/7)🧵
Pavel Grachev retweeted
Data Centers: Financing the AI Buildout with Stijn Van Nieuwerburgh | Ma... youtu.be/e5YuA53OFQ0?si=v8kx… via @YouTube
markusacademy.substack.com
Pavel Grachev retweeted
Why you should work much harder RIGHT NOW marginalrevolution.com/margi…
Pavel Grachev retweeted
I (finally) put together a new LLM Architecture Gallery that collects the architecture figures all in one place!
sebastianraschka.com/llm-arc…
Pavel Grachev retweeted
OK this thing is totally insane. Before going to bed I...
* used try to make a new qmdresearcher directory
* told my pi to read this github repo and make a version of that for the qmd query-expansion model with the goal of highest quality score and speed. Get training data from tobi/qmd github.
* woke up to +19% score on a 0.8b model (higher than previous 1.6b) after 8 hours and 37 experiments.
I'm not a ML researcher of course. I'm sure way more sophisticated stuff is being done by real researchers. But its mesmerizing to just read it reasoning its way through the experiments. I learned more from that than months of following ml researchers.
I just asked it to also make a new reranker and its already got higher base than the previous one. Incredible.
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/autorese…
Part code, part sci-fi, and a pinch of psychosis :)
Pavel Grachev retweeted
My conversation with John Arnold (@johnarnold).
Few people I've spoken with have as wide a view of the global system as John.
He was one of the most successful energy traders of all time, and after stepping away from markets he built a foundation devoted to solving America's most critical systemic problems in a principled way.
John's recent trip to China was the catalyst for this conversation, and I feel lucky we all get to learn from him.
We discuss:
- His trip to China and what it taught him about robotics, AI, and EVs
- What it takes to be the best (and what it costs)
- Building the best seat in the market
- The state of energy markets today
- NIMBYism as the impediment to progress
- What he thinks about the wave of nuclear startups
- Fixing America's broken systems: healthcare, criminal justice, education, and journalism
Enjoy!
Timestamps:
0:00 intro
0:45 China’s Rapid Transformation
3:53 Lessons from the Chinese EV Market
6:12 Robotics
11:22 The Discipline of an Elite Trader
15:42 Leveraging Scale and Proprietary Data
17:36 Lessons from the Baseball Cards
21:15 Trading Natural Gas and Market Dynamics
25:34 Innovation in the Modern Energy Sector
27:02 High-Level Goals of the U.S. Energy System
32:59 Overcoming NIMBYism
36:10 The Challenges of U.S. Transmission Lines
37:55 The Future of Nuclear, Fusion, and SMRs
44:00 The Economics of Solar and Battery Storage
48:28 Data Center Demand
50:28 Housing Reform
53:32 Rethinking the Role of Philanthropic Foundations
57:05 Improving the Criminal Justice System
1:01:58 Privacy and Security
1:05:03 Education and Life Outcomes
1:06:41 The Promise and Pitfalls of EdTech and AI
1:09:12 Identifying Market Failures in Healthcare
1:12:10 The Role of Regulation Across Different Systems
1:14:06 Journalism as the Fourth Estate
1:16:41 The Kindness of Hard Truths
Pavel Grachev retweeted
I started Boom knowing literally zero people in aerospace. Networks are built naturally when doing something interesting—they are not a precondition.
Naval Ravikant: “Networking is overrated… Do something great and your network will instantly emerge"
Naval offers the following advice to startup founders:
“Don’t spend your time doing meetings unless you really, really have to. I really think networking is overrated. There’s all these articles about how you’ve got to network more, and it makes me want to vomit.”
Instead he suggests:
“Go do something great and your network will instantly emerge. If you build a great product or if you get a good customer base, I guarantee you will get funded.”
Recruiting (customers and employees) and learning from really smart people are two exceptions. But don’t worry about building relationships with VCs or going to conferences early on. Just focus on your product, your team, and your users.