I've watched @anneleeskates since she was working with me @floodgatefund. Even then she pushed me to be better, work harder, think smarter. I'm so excited to see her debut Parable and even more excited to have the chance to work with her again!👏👏👏

I learned so much from @johnolilly and how he described his observations of the early days of @figma. Became the basis for how we think of superthinkers

I was founding engineer, like my friend Prem. This is how I think about the founder-employee social contract, now that I’m a founder. Founders must recognize that financial scales are wildly different between ourselves and employees. For example, when I joined a company as first engineer, I had to pay five figures just to early exercise all my equity. I was a new grad, and any salary bump was game-changing. In contrast, a founder with a few years of career experience typically has more savings and doesn't have to pay to obtain their equity. In big acquisitions, liquidation for a founder can be the difference between being a billionaire and a very rich millionaire. For an employee it’s the difference in whether their student loans are paid out, whether they can rent a bigger apartment, or whether they can become a homeowner. Founders hold the cards in acquisitions and we should advocate for employees to be paid out. Employees are a privilege. They are betting their career and wealth on us. We have to honor that and look out for our people.

A year in the making. Thanks to @twilio and @Atlassian for their feedback as we built this with them. This is our take on the elusive problem of routing between LLMs: A preference-aligned approach that offers a practical mechanism to capture subjective evaluation criteria.

When we met @qasar and Peter Ludwig as they were getting started with @AppliedInt, @m2jr clearly articulated why he thought they would be legendary founders and 5 rounds later, we're still believers. Congrats!

What are people using for personal knowledge management these days given AI? I'm worried platforms like Roam have such great capabilities but are not right given all that's happening in AI? Curious to hear what people are doing...

How's everyone liking our new "Free Speech" era? Everyone enjoying the self-censorship? Feeling comfortable with the president threatening to stick the IRS on universities? How bout making people who speak truth on cybersecurity unemployable? That working out well for everyone?

The AI wave has dramatically altered the physics of startup 0->1. Previously, two smart builders would spin up an early product (or sometimes just a Figma prototype) and get it in front of potential customers, then iterate their way to satisfied users. The trouble now is that MVPs have become a commodity. Even I can hack together something half-decent in Replit or Lovable! And when that MVP does turn into a real product, satisfied customers alone don’t seem to protect startups from brutal competitive pressures, even in seemingly niche categories. Afterall, if your customer is willing to replace humans with an agent, they’re not going to hesitate to replace that agent with another. Some founders are now leading early customer conversations with something that can’t be vibe-coded. Something that feels 10x scarier than pulling up a product demo that might not load. And something that directly addresses the top hair-on-fire problem for an average exec in 2025. They’re leading with a future forecast - a clear, credible vision of how their customer’s world is about to change. Despite the conventional wisdom, they’re turning (part of) the fundraising deck into the sales deck. And it seems to be working! Our customers like many of us are consumed by deep uncertainty. How will AI impact our industries, companies and careers? And how quickly? If you have the subject-matter depth (and bravery) to credibly answer those questions, you have an unfair advantage. Your future will resonate with the early customers and design-partners that matter… the ones way more likely to build an org structure and team that will still be relevant in the era of GPT-6. Yes, you’ll still need a wedge. Yes, there will be copycats. And yes, your future may need to evolve. But if you’re co-creating the future with the right design partners, your early customers will be a lot harder to steal. Due to the fact that my new investing career depends on it, I plan to spend quite a few cycles exploring all the impacts AI is having on the 0->1 stage of company building. So far, I’m seeing a few things stand out: - Leading with a credible future forecast really matters - Picking the right design partners is more important than ever - LLM-powered customer simulation can speed this stuff up If you’ve seen other trends, smart experiments or stuff that is working at the intersection of AI and early-stage company building, I’d love to hear about them.

If Trump has the power to do this, then any president, of either party, can destroy any private institution in the country if it defies him… This would include nonprofits, churches, foundations, and charities all would exist only at the whim of the executive. Republicans should be careful what they wish for.

America is brain-draining itself. This guy is a computer vision PhD student. Some founders in our portfolio are nervous about their immigrant hires in the US on visas.