The Titans Paper: Google's Next Transformer Breakthrough

Google Research just dropped a paper that fundamentally changes how we think about AI architecture, now with AI memory.     While everyone’s obsessing over model size and test-time compute, Google Research just dropped a paper that fundamentally changes how we think about AI architecture. This isn’t just another research paper – it’s the next Transformer moment, meet Titans. The original Transformer paper solved a fundamental problem: how to process sequences without recurrence. DeepSeek showed us how to make the training of that architecture

2.12.2025

GenLogs: Building a New Freight Industry

An underdiscussed aspect of working in venture capital is responsibility. As an industry, we’re on the front lines of deciding what businesses should be built and why — of what the next economy looks like.  If you told me that I’d wait 12 months after joining Venrock to make my first investment, I wouldn’t have believed you. But after spending time with fellow investors, marrying things I learned about what makes a great company with my own ideas, I quickly learned discretion, in venture capital, is the better part of valor. The responsibility I had to not only Venrock, but to society,

2.3.2025

Why We Invested in Oumi: Open Source is How America Wins the AI Race

The next chapter in AI development won’t be written by a single company or country. It will be written by thousands of researchers, developers, and organizations working together on open platforms.     This weekend, the world of AI just experienced a fork in history—part swift kick in the pants, part multi-year acceleration of AI research. In one weekend, history changed. We were steady as a freight train on a pre-DeepSeek timeline, dominated by a few companies with access to vast computational resources, ready to invest $500B in a historic (and somewhat comical)

1.29.2025

Winners & Losers in the DeepSeek Era

In every breakthrough, when new facts emerge, there are winners and losers. In an era of DeepSeek, who wins and loses?     DeekSeep r1 represents a significant leap in AI – bringing an order of magnitude improvement in cost efficiency and accessibility at frontier AI performance. It’s a fork in history in AI. But, it likely could only be built now, not a year or two ago, given its extensive reliance on prior investments and distillations from existing leading frontier models.   Let’s remind ourselves what makes DeepSeek unique:   DeepSeek r1 Overview: Cost

1.28.2025

DeepSeek is the Fork in History the US Needed

In reality, this is the best thing that could have happened and reinforces why the US will dominate the AI race. Scrolling X this weekend, it was DeepSeek, DeepSeek, DeepSeek, and more DeepSeek. And for good reason. DeepSeek just shook the AI world. Making us all fundamentally rethink how we develop AI going forward. And of course the spiraling narratives start; calling for the end of the AI Bubble, the end of Nvidia’s ascent, the rise of China AI, and the slowdown in growth. Nasdaq futures already down Sunday night “DeepSeek Shakes Up Stocks as Traders Fear for US Tech

1.26.2025