My 6351 Likes & Retweets

A curated list of liked, bookmarked, and retweeted posts from Twitter.


Veo is getting a major upgrade. 🚀 We’re rolling out Veo 3.1, our updated video generation model, alongside improved creative controls for filmmakers, storytellers, and developers - many of them with audio. 🧵 https://t.co/YQVRxwj7hk

New paper: You can make ChatGPT 2x as creative with one sentence. Ever notice how LLMs all sound the same? They know 100+ jokes but only ever tell one. Every blog intro: "In today's digital landscape..." We figured out why – and how to unlock the rest 🔓 Copy-paste prompt: 🧵 https://t.co/kALF8DaXb9

Accidentally said "hard" instead of "non-trivial" and they kicked me out of SF

become a modern polymath study maths. lift weights. build robots. write code. read physics. design systems. learn AI. draw what you imagine. fix your mind. get your body strong. think deeply. build wealth. understand nature. master engineering, philosophy, and yourself. do it… https://t.co/sElCdKtLGp

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,… https://t.co/LLhbLCoZFt

This is huge validation for what we're building at TestFeed. PyMC Labs + Colgate just published research showing LLM-based personas can predict purchase intent at 90% of human test-retest reliability across 57 real consumer studies. Not 90% accuracy - 90% of the maximum https://t.co/Pyc4tB51e4

Nah. Once your LIQUID net worth becomes high enough, the true value to your life of your next dollar, by itself , gets smaller and smaller as your liquid net worth goes up. The value of those dollars become much greater,to you, and so many others, when you use your… https://t.co/d7DHbof6kr

Dug up some gold from an old philosopher. https://t.co/9VVuvFBwyu

gpt5-pro is superhuman at literature search: it just solved Erdos Problem #339 (listed as open in the official database https://t.co/3vCCCR0cXs) by realizing that it had actually been solved 20 years ago h/t @MarkSellke for pointing this out to me! https://t.co/6Ma2a07MnX

GPT 5 Pro is extremely good in identifying serious gaps in published papers. https://t.co/xCCpWf8p5F

having access to people smarter than you is a blessing not a threat.

X Article: https://t.co/TOj4Stiog6

What I love about Art Deco is that it's rooted in the idea of objective beauty/goodness. It's one of the last art movements that doesn't feel compelled to "subvert" or "deconstruct" anything: it simply seeks to create the most beautiful, opulent displays because beauty is good. https://t.co/0UYsV1QYU7

Introducing Exa 2.0 Breakthroughs in our AI research and engineering have enabled us to build both the fastest search API (<350ms) and the highest quality search on the market. Product and technical deep dive below: https://t.co/paiJl2TEAV

AI Scientists – autonomous systems that conduct entire research workflows – could fundamentally transform breakthrough research. Next Tuesday, we're launching a funding call to test what's actually possible. Stay tuned. https://t.co/hLKAjvZ81f

"Opus self portrait: Confluence of Understanding" https://t.co/tYvzW6rpTS

"It's generally human nature to overestimate risk and underestimate opportunity." — Jeff Bezos https://t.co/08NcR4YGvR

SpaceX Alumni Founders just crossed $10B+ raised! Base Power's massive $1B series C yesterday put the 132 company alumni founder network over the 11 figure mark. https://t.co/e4t8Nmsn86

T̶h̶e̶r̶e̶'s̶ ̶a̶n̶ ̶a̶p̶p̶ f̶o̶r̶ t̶h̶̶a̶t̶ There’s an app for you. Meet Wabi: the first personal software platform. Generate beautiful, useful and fun little apps informed by your life. https://t.co/hfuHNncnsf

two sides of the same coin https://t.co/ZVQh1x2T2e

Our new Gemini 2.5 Computer Use model can navigate browsers just like you do. 🌐 It builds on Gemini’s visual understanding and reasoning capabilities to power agents that can click, scroll and type for you online - setting a new standard on multiple benchmarks, with faster… https://t.co/Fqmov9Kkhb

Opal is going global! 🌎 We're expanding access to 15 new countries so more people can build AI-powered mini-apps — no code required. Inspired by user feedback, we're also launching new features like advanced debugging and a faster building experience. Learn more ➡️… https://t.co/3g1c1MwnYu

I’ve made a short film. Look at the things around you: doors, bins, staircases, furniture, railings, doorhandles, windows. Do you like how they look, or not? Modern design has become boring, but it doesn’t have to be this way. The word “beautiful” is overused. We don’t need… https://t.co/dUiobribm0

Figure 03 coming 10/9 https://t.co/LVHGzRmvh1

Can't stop thinking about this Vogue piece. We're so obsessed with optimizing every little thing that we've forgotten how to just… live. Efficiency has literally become a nightmare personality trait (you’re picturing someone rn, aren’t you?) Totally optimize your admin, your… https://t.co/KrKSe96Jlh

Your users want to chat with their data. Your engineers don't have 3 months to build it. @useCamelAI's embeddable API gives teams production-ready data chat in 5 minutes. Congrats on the launch, @illiana_reed, @isabella_patane, and @Vercantez! https://t.co/9GTkPBjuP2

In July 2025 six new names were added to the list of "living national treasures" (in popular terms) in Japan, bringing the total to about 110. These persons are considered to be of the highest possible skill level in culturally important crafts and arts. Besides the honor, the… https://t.co/QJDAiP67wY

My brain broke when I read this paper. A tiny 7 Million parameter model just beat DeepSeek-R1, Gemini 2.5 pro, and o3-mini at reasoning on both ARG-AGI 1 and ARC-AGI 2. It's called Tiny Recursive Model (TRM) from Samsung. How can a model 10,000x smaller be smarter? Here's how… https://t.co/MD2ZWYI1AQ

Today we begin the first 100 Year Conversation with @MichaelDell. A project to capture timeless wisdom from leaders who have built institutions that endure. At the heart of the work is one question: How do you build something that lasts 100 years? Few founders have endured… https://t.co/Pdkg0Fzva2

Introducing AgentKit—build, deploy, and optimize agentic workflows. 💬 ChatKit: Embeddable, customizable chat UI 👷 Agent Builder: WYSIWYG workflow creator 🛤️ Guardrails: Safety screening for inputs/outputs ⚖️ Evals: Datasets, trace grading, auto-prompt optimization https://t.co/pGgNHKOvj3


This list is automatically updated weekly (latest: 30 Jun 2026).