Here is my call to all highly skilled Indians. Germany stands out with its stable migration policies, and with great job opportunities for Indians in IT, management, science and tech. Find your way to Germany to boost your career: linktr.ee/germanyinindia

capitalism feels bad because technology's scaling effects turn minor advantages into runaway feedback loops. someone with marginally better starting conditions like talent, money, connections can now leverage those into exponentially growing gaps. what used to be linear

ooo this is big!!

trying to ship anything at a large company

BREAKING: Nvidia stock, $NVDA, breaks above $190.00 and becomes the world's first $4.6 trillion company. Nvidia has now added $2.5 TRILLION of market cap since its April 2025 low.

My new hobby is asking GPT-5 Thinking to find errors in every Wikipedia page I read. Interestingly, almost every page I checked has at least one error. 🧵

today i learned there was a family owned watermelon farm in Oklahoma in the 90s that believed in the internet way before everyone else. So they bought thousands of one-word .com domains. 30 years later they're sitting on a portfolio of $100M+ of domains, selling each one for

Just one more SBOM guys I’m sure it will solve supply chain attacks this time

it’s difficult to overstate how important Codex has been to our team’s ability to ship new products. for example: the drag and drop agent builder we launched today was built end to end in under 6 weeks, thanks to Codex writing 80% of the PRs

I would like to inform the serious computer science world that this ship has sailed a long time ago.

This is an excellent history of LLMs, doesn't miss seminal papers I know. Reminds you we're standing on the shoulders of giants, and giants are still being born today. gregorygundersen.com/blog/2025/10/0…

you need distribution because there's no upper cap on the cost it takes to produce good things, but there's is a strict upper cap on the price anyone will pay for something, no matter how good production costs can scale almost infinitely; you can spend $1 million, $100 million,

RLMs are as general purpose as the underlying LM. In fact, from the perspective of the user, they look the same as a model call, but they can internally spawn (recursive) LM calls for intermediate computation. When an RLM is queried, it allows the “root” LM to poke around or

🚨Gemini 3.0 Pro - One shotted After mac os now Gemini cooked hard with iPhone simulation one shot under 2 min iPhone 3G Simulation , Leeham done a great job , and in few min i will drop something super cool , waiting for the output from gemini

Just saw the type of code that makes money and I am disgusted and appalled

RL really sucks. It takes 10 hours just to learn breakout. ... a few years ago. It's <30 seconds on 1 GPU now in PufferLib and still dropping. Write faster code.

My bed doesn’t let me sleep because Internet outage is not the world we were supposed to build for ourselves.

Airbnb CEO Brian Chesky: “We’re relying a lot on Alibaba’s Qwen model. It’s very good. It’s also fast and cheap... We use OpenAI’s latest models, but we typically don’t use them that much in production because there are faster and cheaper models.” The valley is built on Qwen?

any ai tool that promises to deliver complex artifacts with the press of a button will have a large TAM because people desire convenience, but will always be adverse to creative intent and taste, because both of these are fractal and applied fractally, never with one broad and