How we built Grok Bot in a month | Roman Ugarte (SpaceXAI)
- Source
- lennysnewsletter.com
- Author
- Lenny Rachitsky
- Date
Lenny Rachitsky interviews Roman Ugarte, the engineer at xAI who took Grok Bot from a blank repo to public launch in roughly seven weeks. The conversation covers a decision that a lot of teams building coding agents face right now: whether to build a standalone product or ship inside an existing editor like Cursor. Ugarte explains why the team chose to go standalone, along with the early product bets that ended up driving adoption once the tool was live. He also talks through personally onboarding the first 300 users, a detail that says as much about early-stage product work as any of the technical choices: at that stage, someone on the founding team is still doing support and watching usage by hand. For engineers evaluating how fast a small team can realistically take an agent product from zero to a public launch, this is a specific, recent data point rather than a general playbook, and the standalone-versus-plugin tradeoff is one worth reading in full.

Context
Grok Bot is a knowledge-work from SpaceXAI, and its "built in a month" label refers to an internal product, not a public one. Lenny Rachitsky's episode description of a conversation with Roman Ugarte, who helped build it, says a small, isolated team went from the first line of code to a working internal product in four weeks, with a public launch three weeks later. It also says the team personally onboarded nearly 300 of its first users. This is the publisher's summary of the guest's account, not a transcript or an independent measure, so it supports the timeline and the hands-on onboarding but not any claim about why the launch succeeded.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
podcastOpenAI’s Head of ChatGPT: We’re entering a new era of AI (again) | Tibo Sottiaux
podcastAI’s third era: the rise of persistent AI coworkers | Tara Seshan (OpenAI’s product lead)
podcastOpenAI’s Head of Design: This is the best time in history to be a designer | Ian Silber
podcastFull Stack Builder (LinkedIn)
podcastOpenAI Codex lead on the new shape of product work | Andrew AmbrosinoLenny Rachitsky
podcastHead of Claude Code: What happens after coding is solved | Boris ChernyLenny Rachitsky
podcastWhy humans are AI’s biggest bottleneck (and what’s coming in 2026) | Alexander Embiricos (OpenAI Codex Product Lead)Lenny Rachitsky
Checking sign-in…
Loading comments…