Mercor connects experts in their field with leading AI labs to test, challenge, and improve frontier AI models. Whether you’re a software engineer, PhD, consultant, or doctor, you can earn up to $200/hr while working on your own terms. Apply today at work.mercor.com.

US export controls banning Nvidia in China are huge mistake - handing monopoly profits, data & developers in massive China mkt to Huawei will accelerate their AI stack around the world. Let Nvidia win, pin down Huawei, keep developers in Cuda & plow profits back into US. 🇺🇸🚀

George is always insightful!

Welcome “Chicago Mafia” to the Tech Week in SF!

I was talking to a data vendor this week (not Scale) about how insanely cool it is that we can spin up tiny, experimental, expert-level data collections in like.. a matter of days I remember just a few yrs ago this was so hard and we basically told customers this was impossible

Kafka’s architecture is showing its age. Streaming needs its Iceberg moment: 📉 Cut costs by up to 90% ⚙️ Evolve each layer independently 🔓 Enable open, multi-tool access How? A 3-layer model: 🔹 Data – Durable, scalable object storage 🔹 Metadata – Decoupled and

The $15B acquisition of Scale AI by Meta signals a big shift in AI: The competitive edge is migrating away from who has the biggest model and most GPUs to who has the best pipeline for generating proprietary, high-impact data. As such, the ability to recruit experts matters!

Well said, Ed Sim !

Now, students and hobbyists can learn data + AI skills using the same tools trusted by millions of enterprise professionals at no cost. With Databricks Free Edition, users can: 🛠️ Build AI agents and apps with Mosaic AI 📘 Collaborate on data science and ML projects 📈 Create AI

So stoked that we are doubling down on Vellum!

Keep building! Vellum

Right on the money! It would turbo charge your AI centric career even more, if you get a graduate degree in CS or related field, especially for those who were Liberal Art majors. You can do that part-time (e.g. at Stanford). But it is optional, b/c you learn way more at jobs..

It's mid-2025 and people are still searching for the basics on building a RAG Chatbot?! Watch me build one in 3 minutes 👇 1: Create a Document Index (15 seconds) 2: Upload Documents (5 seconds) 3: Build your Chatbot Workflow (2 minutes) 4: Start testing!

I’m soo psyched to announce this one. We have raised again! Raised the bar for AI inference. By snapshotting GPU memory, we can cold start containers running vLLM or many other things extremely fast. This will be a game changer for many customers.

Power law at work=>every VC’s dream!

OpenAI has released its first open weights language model since GPT-2 over five years ago. gpt-oss has - efficient mxfp4 MoEs - native tool-calling & reasoning - attention sinks for long context This looks like a great model for self-hosted agents. Try it on Modal!

To say Ion Stoica is a Force of Nature in AI + Data Infra would be an understatement! Mad respect and ❤️ for Ion and his group at Berkeley!

I first met Ion Stoica in 2015, and have developed a high conviction in Databricks since, despite it had a minimal revenue and had not launched its enterprise product yet. Back then, I thought Databricks would be be 10x bigger than Cloudera but had no idea it could become so big.

Congrats, team Databricks and Ali Ghodsi! Keep building! ❤️

I agree with Jason ✨👾SaaStr.Ai✨ Lemkin that Databricks could become a $200B market cap Co. if it goes IPO next year. Because, other than net margin, Databricks beats Palantir in almost every operating matrics, such as growth, NDR and AI native readiness..