Founders take note of this!

Absolutely x.com/i/status/2066184052575…

Databricks just launched the industry's first Context Engineer Associate cert! 💻 It assesses your ability to design, assemble, and govern the information AI agent systems receive at inference time: retrieval, memory, tool integration, context-window management, and governance. The exam is in beta and debuts at @databricks Data AI Summit where you can be one of the first to take it. It's walk-in at the cert room with one free attempt on-site. The learning material itself is free, the hands-on labs may vary, and the other Databricks certs will be 50% off on-site as well. 
I'm excited to attempt it myself! 🙏 💻You can also tune in to the Summit virtually for free ✨ dataaisummit.databricks.com/… 
📚Learn more about the Context Eng Cert here:  databricks.com/learn/certifi… Can’t wait to attend this year! LMK if you'll be there in person - Let's meet up! 🎉 #DatabricksPartner

Congrats to our friends at SpaceX/X.ai, @sequoia and many other VCs! Happy for you all!

Btw, Modal runs Sandboxes too!

Seats are the receipt. Tokens are the workout log. We are seeing the most AI-forward enterprises (95th percentile) now generate ~3.5x more tokens per employee than the median firm, up from ~2x last April, thanks to @RonnieChatterji and team. The AI gap is no longer about access. It's about absorption.

Frontier model performance on an open model, post-trained in under 24 hours. @trajectorylabs is showing what's possible when great open models meet the right training infrastructure. Proud to power the compute behind this work alongside @nvidia .

Excited to be speaking at the AI B2B Founder & VC Dinner in San Francisco. Looking forward to discussing what it takes to build and back the next generation of AI companies alongside an incredible group of founders and investors. See you tmrrw! RSVP: luma.com/founders-28sz

Right on the money! Developers are voting with their feet in choosing open source model at ~1/30 of the costs over closed sourced frontier models in ~80% of their workflow. Thanks, @brian_armstrong and @Shaughnessy119 for sharing!

AI progress requires (1) compute, (2) algorithms, and (3) data. - The leading compute company is worth $5 trillion. - The leading model company is worth $1 trillion. - @mercor_ai is the leading data company and is currently valued orders of magnitude lower. There's an opportunity in how the market is mispricing the value of data. Data is the oil of the AI revolution. It is the primary way that models and enterprises build competitive advantages.

Databricks CEO interview insights: Significant acceleration beyond 65% YoY growth in Q1. Over 81% of databases launched on Databricks are now from agents. Cybersecurity market revolution - existing cybersecurity vendors can't keep up with malicious agents. Therefore, Databricks entered this market. Additionally, as AI makes software development much faster, it is much easier to enter this market. Reaffirmed no 2026 IPO plans - bad year for IPO due to big dislocations: election year, energy uncertainty and mega IPOs. "Chinese open-source models are absolutely dominating" given their significant cost advantage. Source: Bloomberg

Interesting observation but the rational is more than just the size, in MHO. It also depends on the nature of the biz, or whether the growth drivers relate to AI/data or not.

5 out of 26 PortCos of ours are on this list. Keep up the good work, folks! 🚀🎉🎈

With today's launch of Nemotron 3 Ultra, @nvidia continues to expand its investment in open-source AI. Their flagship frontier-reasoning model, built for long-running autonomous agents, is available Day 0 on Modal. - 550B with 55B active parameters - Hybrid Transformer-Mamba MoE architecture - 1M context - Up to 5x faster inference - Up to 30% lower cost

This move by Benchmark signals a shift in emphasizing returns vs. stage. Why spray and prey while one can double or triple down on iconic founders and Cos.?

Using open models and inference clouds (which serve open models) is a leading indicator of what is to come. The advantage of open weights is that you can train, serve, and continually improve your own model. You get control over cost, quality, latency, deployment options, etc. It's just a matter of time before every company does this for their most important AI workloads.

Superintelligence will be built on Self Improvement. Today @hexoai, we’re excited to release ‘SIA’ - an open-source Self-Improving AI, to achieve any goal through recursive self improvement. While trying to solve a problem, SIA doesn't just improve it's abilities by updating it's harness, it updates it's own weights as well.

After some mathematical rewrite, turns out all of transformer is a series of gemm epilogue. Given a few optimized primitives, LLMs (and novice humans) can write speed-of-light kernels for all transformer ops!