🚀 Ling-1T — Trillion-Scale Efficient Reasoner Introducing Ling-1T, the first flagship non-thinking model in the Ling 2.0 series — 1 Trillion total parameters with ≈ 50 B active per token, trained on 20 T+ reasoning-dense tokens. Highlights → Evo-CoT curriculum +

No dev team? No problem. Caffeine turns your ideas into launch-ready apps — just by chatting. Start for free →

Just came back from China, some early thoughts: Life observations: - the great firewall is not that great, so easy to get a vpn or an e-sim that has one - hongkong is surprisingly really run-down, locals have said it’s way past its prime - everyone of all ages in china

Yup

Why Chinese are richer than people think. "Millionaire" is defined by *liquid investable wealth*, which excludes houses. But Chinese mainly invest in houses due to comparatively low returns of other assets. In fact, about 30% of families own more than one house. How does it

A solid paper from Stanford, Princeton, Harvard, University of Washington, and many other top univ. Says that almost all advanced AI agent systems can be understood as using just 4 basic ways to adapt, either by updating the agent itself or by updating its tools. It also

Generative AI is not magic, it is math that creates. From generating art and code to simulating human creativity, every GenAI model works differently, but with one shared goal: creation from data. Here is a simple breakdown of the five major types of Generative AI models

I’m halfway through this book and the clarity around training LLMs is unreal. If you’re building or researching with LLMs, this one’s hard to skip.

What we learned about memory in 2025 8 comprehensive resources: ▪️ Memory in the Age of AI Agents ▪️ When Will We Give AI True Memory? (interview with Edo Liberty, founder & CEO Pinecone) ▪️ Why AI Intelligence is Nothing Without Visual Memory (interview with Shawn Shen,

Anthropic just released a massive new Economic Index Report, and it’s one of the most data-rich looks yet into how Claude is actually used in the real world and what that means for jobs, productivity, and inequality. 🔹 Most Claude tasks are high-skill, often requiring

A few random notes from claude coding quite a bit last few weeks. Coding workflow. Given the latest lift in LLM coding capability, like many others I rapidly went from about 80% manual+autocomplete coding and 20% agents in November to 80% agent coding and 20% edits+touchups in

Karpathy Guidelines for coding agents github.com/forrestchang/a…

10 AI projects that will get you hired in 2026: 1. RAG App with Real Evaluations Build a production-grade RAG system with evals, monitoring, and feedback loops. Stack: OpenAI or Claude + vector DB + evals Guide: pinecone.io/learn/retrieva… 2. Autonomous Research Agent An agent

The most underrated line in Andrej Karpathy post: "I can approach code that I couldn't work on before because of knowledge/skill issue." I was an engineer. Then I became a PM. For years, I didn't touch code. I refused when being offered access to repos. Now I'm building again. Not

Latest benchmarks for Kimi 2.5 and Qwen-Max-Thinking. I really don't see a US lead vs open source Chinese models. The Chinese models are free and much cheaper to run. DeepSeek V4 might leapfrog all of these models...

Difference Explained Well 📚📘 Save it for later 💯 #llm #rag #ai #agent #agenticai

Eileen Gu said after the competition: "That was the best slopestyle run of my life today. Under the greatest pressure, I’m just happy I was able to perform my best run. I'm proud to have represented this sport and to have given it my all."#MilanoCortinaOlympic2026 #谷爱凌

This paper unlocks every algorithm used by hedge funds. 151 trading strategies. Get it here (361 page PDF):