3x🐊 | building @zencap_ai
West Palm Beach, FL
Joined April 2012
- Tweets1.2K
- Following2.5K
- Followers352
- Likes4.5K
RSI is good for the soul
Today we're releasing data on models accelerating research at OpenAI.
Recursive self-improvement could be the most important contributor to AI capabilities over the next few years, but by default it will only be seen inside a few frontier AI labs. Being transparent is more urgent than ever, so we can inform the public discussion on whether and how to pace model development. I ask other AI companies to do the same.
openai.com/index/research-ac…
a win for opensource
Exciting day for NVIDIA and @huggingface.
Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. They allow every developer, startup, university, industry and country to build with, customize and benefit from AI.
Thank you @ClementDelangue for coming to me.
NVIDIA is going to be a great home for Hugging Face, its community and the future of open models. 🤗
blogs.nvidia.com/blog/nvidia…
Max retweeted
I don't know if you really understand what these people just did...
They built an AI that looks at a few casual real-world videos or photos and instantly turns them into a full 3D playground.
You can MOVE THE CAMERA HOWEVER YOU WANT, change the lighting, move objects around, and even let physics play out like in the real world.
View events from impossible angles because Atlas knows what was there.
This is massive!
Let's take robotics. Until now training robots meant collecting tons of expensive real-world data. Now a handful of recordings can become thousands of different simulated situations.
Robots can practice the same task in new rooms, with different objects and lighting without anyone having to film it all again.
For the industry it means we can train robots way faster and cheaper. Home robots, warehouse robots, whatever... they can learn new environments at scale instead of staying stuck in the lab.
This is one of the first real bridges between “cool AI video models” and actual physical machines that have to work in the real world.
Blog: worldlabs.ai/blog/atlas
The team behind @theworldlabs: @drfeifei, @jcjohnss, @BenMildenhall and @YunzhuLiYZ. The video is from @davidpantera_.
Follow for more insights into robotics and AI.
——
Weekly robotics and AI insights.
Subscribe free: 22astronauts.com
Max retweeted
How our data shaped neural architecture discovery, and how automation can reshape the future
coreauto.com/blog/how-our-da…
looks pretty elite
I'm so excited that our @theworldlabs team has achieved a major milestone today! Introducing Atlas - a first of its kind multimodal world model trained from scratch! 🚀
Atlas is capable of generating frames with pixel-perfect camera control, reconstructing large scenes from as few as one single input image, simulating space-time by reframing videos, natively outputting 3D spaces from one or more input images, composing multiple posed images into a consistent 3d world, and more! This is the best camera conditioned world model ever, opening doors to many possible use cases from VFX to robotics. I'm so so so proud of our team!♥️
Max retweeted
A great read that, in my experience, very accurately captures the challenge of diffusing AI into investment firms.
Investment firms are complicated, multi-player networks and even as model intelligence has reached remarkable levels, the challenge of wiring that intelligence into the firm remains a key inhibitor to adoption.
AI remains mostly a one off tool “Hey Claude, summarize this press release and tell me what it means” or the enabler of simple tools like a morning inbox summary.
AI is capable of so much more, however, and is steadily marching towards “Digital Analyst” capabilities with much better Excel read/write and the ability to navigate a much more extensive internal & external Context graph.
More firms are at the point of realizing that “Claude Code and a stack of Skill files” isn’t the right playbook and I agree with Aatish that the vertical layer is set to become very interesting, particularly as the pre-training curve has flattened and the frontier ecosystem will likely support a “slingshot” approach of extremely good, domain specific, post-trained models, wired into investment firms with an FDE/FDI approach.
Max retweeted
Read this repo. Start from:
1. packages/ai
2. packages/agent
3. packages/coding-agent
Will teach you more about this technology than most other repos.
To learn to train a model: github.com/karpathy/nanoGPT
To learn research: github.com/karpathy/autorese…
Learn the craft well
A harness turns a model into an agent. At it’s core it provides 4 things:
- a system prompt
- tools
- an agentic loop
- a translation layer across models
New blog post from Earendil co-founder @colindaymond on what a harness is, and how you can own yours. Full post below
Some awesome work here from @d33v33d0. The use-cases for industrial real estate are endless. Will be thinking a lot more about this soon.
Max retweeted
Elon's playbook: build the datacenter first, buy the GPUs last. Get close enough to powering on, and suddenly you're holding all the leverage.
Distillation by open weight model providers is becoming more difficult. Reasoning traces are now being obfuscated by frontier labs for this very purpose - to stop others from reverse engineering their models via reasoning traces. After all, frontier labs spent the money to train the models in the first place, and those capex moats must be protected and harvested now. Telemetry data contains rich insight so it’s important to utilize it while you still can.
there are some really interesting rumors going around related to the distillation of open-weights models (Kimi, Qwen, Minimax, etc.) and they're very related to my PhD work
The narrative [speculative]:
• good distillation relies on reasoning traces, normally hidden from users
• Chinese labs figured out in early 2026 how to reverse-engineer reasoning from Claude Code and Codex
• they were able to collect large amounts of long-horizon data *with reasoning traces included* this way
• this jailbreak led to a new wave of OSS models we've enjoyed over the past few months
I'm not sure how true it is, but reasoning extractability seems like a huge uncertainty around the future of open models
in particular i'm curious how much having the reasoning matters. this (plus Anthropic's messaging around distillation attacks) indicates that reasoning chains are crucial for distilling model capabilities.
our research (arxiv.org/abs/2603.07267) found something different: if you train a high-quality reasoning inverter, it's often pretty easy to reconstruct useful traces from frontier models given their outputs.
figuring out how to approximate frontier model reasoning traces might turn out to be an existential problem for open weights models
Max retweeted
I think the market ultimately has no choice but to go sell memory, long optical in the "short term." Actually, some hedge funds already seem to have this position on.
There are three main reasons.
1. With Korean leveraged ETFs effectively dead, LPs are in a redemption rush, which could bring out additional sell on flow.
2. Nvidia is nerfing Rubin Ultra's HBM and responding with optics, tying multiple racks together, so that even if Rubin Ultra's per rack performance is not superior to Rubin, at the cluster level optics let the Rubin Ultra cluster hold an edge over the Rubin cluster. This holds even if Rubin Ultra's HBM nerf is a supply problem rather than a demand problem.
3. Consensus is forming that memory prices will peak within the next two quarters.
Medium to long term I am still a memory bull, but short term I am somewhat bearish on memory. I currently have no memory position.
Max retweeted
FEL = Free-Electron Laser
Is this even possible?
Next-gen EUV light source for chipmaking: a particle accelerator shoots high-speed electrons through magnets, generating ultra-bright, clean EUV light.
At Terafab scale it could sit as a central “light utility,” powering many scanners at once with way more photons and less mess than today’s tin-plasma sources.
Not day-one tech, more like the upgrade path once volume ramps hard.
Max retweeted
Looks like a synchrotron - Elon going to build free electron laser based EUV machine?
Holy Giga based 😳⚛️🤯
Max retweeted
NEWS: The Boring Company introduced Vertical Cassette 7, a huge conveyor belt that carries dirt and rock out of the tunnel as the digging machine works.
It stands 37 feet tall and weighs about 220,000 lb.
It can move up to 990 tons of dirt an hour, enough to keep the digging machine pushing ahead 4 miles a week.
It can bend around tight 500 foot turns, which matters for digging under crowded cities.
It can support a tunnel of any length. Small "Belt Boosters" every 1.5 miles keep it moving.
It has more than 650 upgrades over the last version, for faster setup, better reliability and less power use.
Fun fact: every 15 inches the machine digs makes about one dump truck of material.
Source: The Boring Company, August 2026
Max retweeted
My seventh conversation with @GavinSBaker.
It's about the gap between what the market is doing and what companies are seeing. It's been a tough month or so for public AI names, but there's no sign of a slowdown on the ground in Silicon Valley.
We discuss:
- Why old GPUs are getting more expensive
- Open source vs the frontier
- Claude as Wall Street's Walter Cronkite
- Whether the buildout gets funded out of cash flow or debt
- SpaceX as a data center company
- And more
Enjoy!
TIMESTAMPS
0:00 Intro
1:17 AI Selloff vs. Fundamentals
10:20 Financing the AI Buildout
18:18 GPU Prices Keep Rising
24:23 Claude Moves Markets
29:01 What Could Break the Thesis
36:59 The Memory Supply War
42:08 Nvidia’s New Playbook
50:14 China and Open Source
61:51 Data Centers and Regulation
71:11 SpaceX and Orbital Compute
Max retweeted
Our general approach is to automate the experimental loop. We think this approach is broadly applicable across many different fields of science and engineering. We’ll initially focus on ML research and engineering, but believe the approach can help with important subproblems in nearly every one of the fourteen NAE Grand Challenge problems. We think doing this well requires strong expertise in machine learning as well as large-scale systems.
Max retweeted
I’ve been working towards AGI my whole life, and as we enter this pivotal moment, I’m stepping into a new role as Chair of Google DeepMind & Chief Scientist of Alphabet. This will allow me to focus on long-term strategy, and accelerating scientific breakthroughs, including leaning into my work at Isomorphic to help cure disease.
I’m excited that @koraykv will be stepping up to lead GDM as SVP, alongside @joshwoodward and our exec team. I could not be more excited and confident about our amazing next chapter! 🚀
blog.google/company-news/ins…
Max retweeted
it's worth noting that due to rising costs of pretraining, the frontier isn't *that* accessible. training a 5T-A100B model on 50T tokens on GB300s costs like $10 million for the hero run. double it for RL. there's only a few thousand companies in the world that could justify it