Grok @Bot is the hottest new AI product in the world right now. I'm absolutely hooked. It all started a few months ago as a tiny team holed up in a separate part of the office, building from scratch. Four weeks after the first line of code, the whole company was obsessed. Three weeks later, it launched to the world. Less than a month in, millions of Grok Bots are already helping people get work done. Roman Ugarte (@romanugarte_) incubated, launched, and now leads product for Grok Bot. This is the first time he's told the story: 🔸 Why they started fresh instead of building it into Cursor 🔸 The two non-obvious bets that made it such a success 🔸 What they learned from manually onboarding the first 200–300 users 🔸 Why they spent the weeks before launch *unshipping* features 🔸 The first prompt to try when you open it Watch the full episode right here on X 👇

Since launch, the Grok @bot team shipped: - Android and iPad apps - Enterprise availability - X and Outlook plugins - Shopping with Stripe Link - Shareable templates - Support for more plans It's been 24 days. What will it be in a year?

I got early access to Grok Bot and I'm hooked. I haven't been this excited about a new AI product in a while. It's like OpenClaw, but super easy, reliable, and less scary to use. I think this will be a huge new product line for Cursor/Grok/SpaceX. I've already found so many ways to use it that have meaningfully made my life better: 1. Matchmaking people looking for jobs with companies who are hiring (see below) 2. Auto-replying to support emails (saves me hours!) 3. Scanning my credit card statements and finding recurring subscriptions to cancel 4. Sending me (really good!) briefs for upcoming podcast guests See below for my actual set of agents that I've been using and chat with daily. Great work on this team Grok Bot. (I'm not an investor in this, nor do I have any ties to this product/company. I'm just a fan!)

Being at @cursor_ai for almost a year, I was always secretly jealous that GTM work didn't hit the productivity gains that Software Eng and Support had with the onset of AI. It was there, but it just wasn't the same. Since @bot was released internally, I finally feel it #bot# <3

I’ve been talking to CTOs about their AI spend. One of them saw it grow 10x in the first half of this year. The velocity gains have been great so she’s okay with it. But another 10x would be a problem. Common story. They now do monthly reviews with the CFO, budget reallocation based on performance, ownership that cascades down the management chain. It's boring. But it works. Cursor’s job is to get you more intelligence per dollar. The rest is good old fashioned management.

We briefly went insane and decided every software engineer should also become an expert in model benchmarks, thinking levels, and cache hit rates. Cursor Router is out today. It picks the model for each task based on measured quality against cost. Fable 5 level performance for 60% less.

When I was growing up, programming was an expression of creativity. Somewhere along the way, it became more about the chores. Merge conflicts. Flaky tests. CI failures. Bug triage. Let's give the chores to the robots and get back to building cool shit.

Helluva week for new models. We got flagship versions of Grok, GPT, and Muse. Obviously I want Cursor to be a massive success. But I've been a software engineer a lot longer than I've been at Cursor. The whole profession wins with more serious players in the game. Competition is what keeps prices down. It’s what makes products great.

Token ROI is the number one thing customers want to talk about right now. Agents have gotten so good and so pervasive we’re all flying past our budgets for this year. Our aim is to offer the best outcomes per dollar anywhere. Grok 4.5 is a big step toward that. More to come.

I’m gonna get roasted for being a corporate robot for saying this, but I love Cursor’s company values: Just do the thing Delete the product Don’t talk about the weather They’re ambitious and weird and actually influence my work. Plus, I remember them without having to look them up. I can’t say that about any other company I’ve worked at.

Everyone is writing about agent loops right now. Including us at Cursor, because they're so powerful. But here's a prediction: a year from now, nobody will be talking about them. Not because they weren't useful. Because they'll work right out of the box. Batteries included. No instructions necessary. Feels a lot like prompt engineering two years ago. It was incredibly important. People wrote courses on it. Now you just talk to your agent like a normal person. That's the strange thing about AI right now. You learn something critical, get huge gains, and before long it's the new normal and something else is the bottleneck. So the alpha isn't what you know. It's how fast you learn it, and how easily you can let it go.

A lot of engineering teams I talk to have reached the point where AI is writing almost all their code. You might think: mission accomplished. AI transition complete. But there's another big shift underway. Coding is moving from your laptop to running 24/7 in the cloud. Many leading teams, Cursor included, are already deep into the transition. Even sitting close to it, the pace still amazes me. The ground shifts every few months now.

One thing about Cursor that's been genuinely surprising is how minimal the internal documentation is. In the pre-AI world this would've led to chaos. But things move so fast now that most documentation would have a half-life of only a few weeks. The company has embraced on-demand AI research directly from the sources of truth: code, databases, meeting transcripts. It really feels like the future.