Patrick Bragg

Stripe is buying OpenRouter at a reported $7.5 billion. What a payments giant wants with a startup that routes prompts between models is a fair question, and the answer in the founders' leaked letter is the singularity, which apparently began on January 1. The real answer is less funny, and it has to do with where the price of a request is known. The application sends a prompt and gets text back. The gateway is the component that knows which model served it and what it cost. Per-seat pricing works because a seat is knowable in advance. A task run by an agent isn't, and the cost lands after it finishes. So metering ends up at the routing point, which is also where model choice gets made on price. A monthly bill is easy to produce. Cost per task is the number that moves as the loop gets longer, and it's the one teams tend to be missing.

1

1 Comment

Unpopular Opinion: The US isn't just richer than us. They are more patient. I have been analyzing why the US ecosystem churns out companies that define the decade, while we mostly churn out great consumer apps. The difference isn't talent. It’s the appetite for Infrastructure. In the US, capital is willing to fund the "boring" rails that take 5 years to build but 50 years to replace. In India, we want the "shiny" product that launches in 3 months. You can see the gap in two brutal comparisons: 1. Gen AI : US funds the Foundation Model (The Brain). India funds the Customer Support Wrapper once the revolution has happened. 2. Mental Health : US funds Clinical Intelligence & Outcome pipeline. India funds "Directories" and Booking Aggregators. We are funding access, but we aren't funding the actual care infrastructure. The Reality Check: "Infrastructure" is hard to sell in a slide deck. It doesn't have a viral loop on Day 1. I feel this : Consumer trends fade. Infrastructure sticks. The biggest companies of 2030 won't be the ones delivering chips in 10 minutes. They will be the ones building the Models and the Rails that the world runs on.

23

Most early-stage fintech teams don’t lack engineers. They lack product direction. Shipping faster won’t fix uncertainty. Clarity will. If you’re building, sprinting, releasing… but still questioning the direction — this is for you.

6

The hardest thing in venture is falling into the pattern matching trap based on prior experience. It’s super challenging to keep an open mind and to preserve first principles thinking while evaluating a new opportunity. This is why I maintain that the best venture deal ever is Mike Moritz’s investment into Instacart. Not just because the exit was amazing ($10b public company!) but because Moritz had invested $50m+ into Webvan, the online grocery delivery service that was one of the biggest casualties of the dot-com boom when it filed for bankruptcy in 2001. Twelve years later, Moritz led the series A of Instacart. To me, the Instacart investment shows the first principles thinking, the intellectual humility, and the emotional resilience that separates great investors. Most people who lost $50 million watching a grocery delivery company implode would have built a mental scar tissue around the category. “We tried this. It doesn’t work. Consumers don’t want it. The unit economics are broken.” Moritz could easily have justified the pass. But first principles thinking requires you to decompose the question “did online grocery delivery fail?” and ask why it failed — and whether anything has changed since. By 2012, the answer was obviously yes. Smartphones meant consumers always had the app in their pocket. GPS made real-time logistics tractable. The gig economy created a flexible labor pool that made asset-light delivery possible. Webvan failed because it tried to own the warehouses, the trucks, and the last mile in an era when the infrastructure didn’t exist to support any of it cheaply. Instacart was a fundamentally different business, though it might have looked superficially similar. In venture, pattern matching is seductive precisely because it’s usually right . The trick is knowing when your pattern is a useful heuristic and when it’s a cognitive trap. The best investors hold their theses loosely enough that new evidence can update them. Moritz’s Instacart investment is the clearest proof of this.

288

24 Comments

The Headline family is experiencing incredible momentum, and today our portfolio company DeepIP is announcing a $25M Series B, bringing total funding to $40M.  This is the kind of company that doesn’t just add AI to a workflow, but it sets the standard for an entire category. The momentum speaks for itself: 1️⃣ 10x ARR growth in < 18 months 2️⃣ Trusted by 400+ law firms and leading corporate IP teams globally 3️⃣ 2M+ chats on the DeepIP platform We’ve had the chance to back FX (Francois-Xavier) L. and Edouard D., leading their Seed round with Headline. Watching them take on a problem as complex and high-stakes as patent work has been genuinely impressive. And working with them has been an absolute pleasure: these 2 are coupling very high ambition and humility. Tech built in Europe, but opening the US office very early and FX starting a life in NYC with his family with 0 hesitation in order to build a global category leader in the IP industry. What convinced me early is still what stands out today: patent teams don’t need another standalone AI tool. They need something that fits the way the work actually happens: documents, long-running matters, multiple stakeholders, and zero room for sloppy context handoffs. Big congrats to the whole DeepIP team + Korelya Capital, Balderton & Serena Proud that we’re doubling down at Headline and continuing the support for what’s next!

124

10 Comments

New episode of Uncapped with Y Combinator's leadership, Garry Tan, Harj Taggar, and Jared Friedman. We talked about YC's core value prop, how AI is changing the process of finding product market fit and raising capital, San Francisco and California, the future of YC, and more. (0:00) Intro (0:18) YC's product (5:05) AI and the new builder (13:01) Pivots and upcoming trends (22:26) Making something people want (24:50) What’s in store for SaaS (33:02) Capital in the age of AI (36:28) The human capacity for desire (42:18) Building in America (44:29) Fixing San Francisco (47:58) Scaling YC https://lnkd.in/ggWMW9zK

105

2 Comments

It is a wild time to be building a company! Thankful that Where2Wheel has a real world physical component that can’t be easily replicated by AI. Uncaptured & overlooked niches + good tech paired with a physical real-world aspect seems to be the best moat nowadays!!

1

A good heuristic for whether your company is moving fast enough for AI: Append "engineering" to every function. Then rate it. GTM → GTM Engineering ✅ Product → Product Engineering ✅ Talent → Talent Engineering 🚫 Talent is where I see it most often breaking. The best teams are running very different playbooks: new sourcing signals, new tools, new targets, different interview loops & closing strategies. The talent leads are former founders or engineers. h/t Umesh Khanna for the inspiration

60

5 Comments