A product discussed on Y Combinator.

Boris Cherny: We Cut 80% of Claude Code’s Prompt
Jul 27, 2026 · 35:52
Boris Cherny, creator of Claude Code, joins Diana Hu at Y Combinator's Startup School 2026 to discuss how Opus 5 eliminates prompt injection, why Claude Code deleted 80% of its system prompt for the new model, and how builders can 'unhobble' AI by giving Claude harder problems. He reveals that Opus 5 is not prompt-injectable thanks to three safety layers, and that deleting most instructions actually makes the model smarter. Cherny explains the concept of product overhang—the gap between model capability and existing products—and shows how a single prompt using dynamic workflows rewrote Bun's entire codebase from Zig to Rust in 11 days. He also describes a two-week prompt that spawned thousands of agents to rewrite Claude's Electron app in Swift, and shares how Claude now maintains its own codebases with daily routines removing dead code and unifying abstractions. Cherny argues coding is solved for most tasks but advises CS students to learn practical application and build things people want.

What Top AI Labs Are Really Doing With Observability
Jul 24, 2026 · 30:30
Olivier Pomel, CEO of Datadog, recounts how two French engineers in New York built the company that monitors the entire cloud, starting with a Y Combinator rejection that became a driving force. He attributes Datadog’s success to staying close to raw customer feedback and support tickets, building culture by example rather than writing it down, and betting early on the cloud explosion. Datadog now employs 8,000 people and offers 20–25 products. Pomel reveals that the company went public on COVID lockdown day, saw its stock drop 65% in two weeks, and recovered through consistent storytelling to public investors. On AI, his co-founder told engineers that in two quarters they would not be writing any code anymore; Datadog is investing in agents and smarter observability. His key advice for founders is to move faster on both hiring and firing, since hiring quickly and fixing mistakes beats waiting for the perfect hire.

He Built the World's #1 Open-Source Coding Agent
Jul 24, 2026 · 44:27
Jay V, founder and CEO of Opencode, explains how his open-source coding agent grew from 650,000 to 13 million monthly active users in 2025—a 20x increase—while processing 7 trillion tokens per day and reaching $40 million annualized revenue. Anthropic's attempt to block Opencode users from using Claude Code subscriptions inadvertently fueled the growth by equating the two products in users' minds. Opencode's success hinges on offering choice across 70+ models, including open-source alternatives like DeepSeek and GLM, which are cheaper and faster than frontier models, making coding agents accessible globally—especially in developing countries where a $200/month Claude Code subscription is prohibitively expensive. Jay shares that 12 Fortune 500 companies use Opencode, often discovered via bottom-up adoption. The product's terminal-first design and open positioning were deliberate choices. Jay also recounts his 16-year entrepreneurial journey: the same legal entity (incorporated 2010) applied to Y Combinator nine times before acceptance in 2021, and each prior failure—from a serverless platform to a terminal coffee shop—contributed to the product and market instincts that made…
Powered by PodHood