dogs of Braintrust

going to track progress for my new co here — no more building in the shadows 😤 one of the most demanding use-cases for AI Agents we’ve seen has been on social listening and engagement. companies want to know who is talking about them and what they’re saying.

New cookbook: Evaluating a prompt chaining agent To produce production-ready agents, you need to understand what's going on under the hood. In this cookbook, we demonstrate how to evaluate and measure the performance of these types of agents. braintrust.dev/docs/cookbook/…

o3-mini-high with the same prompt 👀

Taken from Elad Gil.

Sam Altman on Self-Belief vs. Self-Delusion

as ai continues to get better agency will matter more than intelligence

New cookbook: Prompt versioning and deployment Small changes to a prompt can dramatically change the quality and tone of LLM responses. Here’s a workflow for refining prompts, measuring effectiveness, and deploying the right one in production. braintrust.dev/docs/cookbook/…

Puentes is starting this Monday! I'm flying 8 engineers to a mansion in SF. Hosting dinners every night with founders like Guillermo Rauch, Finn Mallery, and Ted Nyman. My goal is to expose them to the city and create opportunities for them to stay. DM if you want to help! Meet the engineers👇🏼

It was a beautiful night. Thanks for coming Guillermo Rauch ❤️

Full circle moment. It was an honour meeting you Guillermo Rauch

Puentes just finished - it was sick. I took this city for granted, and these guys reminded me that we’re all so damn lucky to live here. Thank you everyone that helped! We’ll be back.

Today, we're officially launching Origins. The podcast discovering the early, unheard stories of icons from all industries.

I wrote an essay about Gmail’s useless email-writing AI assistant: “AI Horseless Carriages”. link and TLDR in thread

Would you hire a salesperson that works 24/7/365 following up with customers? We built a custom agent for a customer that uncovered over $2.7M in revenue. Meet an autonomous closer that reads every CRM record in real time, predicts the exact next touch, schedules it, and

you gotta be a special type of delusional to build consumer

Um Y Combinator does this count as a launch? We’re building Doe. dm for info

"Balance" is a long term game, that too many people treat as a short term game

Peter Thiel: The most common mistake businesses make is to go after markets that are too big.

I’m excited to announce doe coming out of stealth. Richard Ou and I are building vertical-specific AI tuned for industries still stuck with legacy software and nuanced workflows. Dental is just the start. Proudly backed by Y Combinator and an exceptional team.