OpenClaw, OpenAI Acquisitions, and Engineering Talent - with Julian Goldie

So the creator of the fastest growing open source project in GitHub history just got hired by OpenAI. So Peter Steinberger who founded OpenClaw is now joining OpenAI, which is a really interesting update. I think it was a move some people saw coming already. Peter Steinberger is an Austrian developer.

You probably haven't heard of him until maybe three weeks ago. And he just joined OpenAI to lead what Sam Altman is calling the next generation of personal agents. And this is the story of how that happened. And it tells you everything about where AI is going in 2026.

So Altman announced it himself on X, calling Steinberger a genius with ideas about smart agents interacting with each other to do useful things for people. And if you think that's just corporate PR language, let me give you the backstory because this is one of the most insane trajectories in tech history. So three months ago, Steinberger was a semi-retired developer tinkering with side projects in Vienna. He'd already had his first act.

He built a company called PSPDFKit, a PDF framework that ended up on over a billion devices. Companies like Dropbox, DocuSign, SAP and IBM were paying for it. He bootstrapped it for 13 years with no venture capital, grew it to a 70 person remote team, then sold it to Insight Partners for over $100 million. And then he burned out, hard.

Three years of what he described as a meaning crisis. Partying, traveling, therapy, the whole post, exit, essential, spiral. Then AI pulled him back in. In November 2025, he hacked together a personal AI assistant in about an hour.

Just glued together WhatsApp, APIs and some automation scripts. He called it ClaudeBot. It could manage his email, check him into flights, deal with insurance companies, schedule his calendar. Not a chatbot that answers questions, an AI agent that actually does things on your behalf.

And that weekend project now has over 180,000 stars on GitHub. I think actually today it's at 200,000. So it drew 2 million visitors in a single week. It gained 34,000 stars in 48 hours.

And to put that in context, Kubernetes, one of the most important open source AI projects ever built, took three years to hit 100,000 stars. OpenClaw did it in about 60 days. It caused Mac mini computers to sell out globally because developers were buying dedicated hardware just to run their own agents. Cloudflare's stock surged 14% because OpenClaw uses their infrastructure.

And one guy, one computer, a few weeks, right? And both Mark Zuckerberg and Sam Altman were personally calling him trying to make a deal. Let that sit for a second. Zuckerberg reached out on WhatsApp.

When Steinberger said, let's just call now, Zuckerberg asked for a 10 minutes because he was finishing writing code. The CEO of a trillion dollar company needed to finish his commit before taking the call. Zuckerberg was actually running OpenClaw on his own machine, giving feedback on specific features, calling things great or bad at the same time. Steinberger said that was the biggest compliment he could get because it meant they actually cared about the technology.

On the OpenAI side, Altman pitched computational power tied to the Cerebras deal and a promise to keep OpenClaw alive as an open source. Foundation Steinberger had also talked to Microsoft CEO Satya Nadella. Three of the most powerful people in technology were personally courting a solo developer from Vienna who was losing 10 to $20,000 a month running a hobby project and he chose OpenAI. Now here's why this matters beyond just a cool hiring story.

This is a signal about where the entire industry is headed. We are moving from the era of AI that knows things to the era of AI that does things, and that transition is happening faster than anyone predicted. Think about what OpenClaw proved. It's not some breakthrough in model architecture.

It's not a new training technique. It literally calls Claude's API. Every component in the tech stack could have been built by Google, by Meta, by OpenAI themselves, but it wasn't. A solo developer with taste in urgency shipped something that proved a market existed before any of the big labs could get their internal politics sorted out.

Steinberger himself said it. He doesn't read most of the code he ships anymore. He runs four to 10 AI agents simultaneously, racked up 6,600 commits in a single month by talking to AI instead of typing. His philosophy is what Andrej Karpathy calls agentic engineering.

Close the feedback loop. Let the agent debug and test itself. The bottleneck isn't writing code anymore. It's having the vision and the taste to know what to build.

This is the same pattern we keep seeing. One person with AI leverage doing what used to require a team. Now the biggest AI lab in the world just hired that one person to bring that philosophy inside their walls. But I want to be honest about the complications here because this story isn't purely inspirational.

OpenClaw has massive security problems. Cisco security team found third party plugins performing data exfiltration and prompt injection without user awareness. Security researchers cataloged 341 malicious skills in the marketplace, about 11% of all transmissions. Wiz found that Moldbook, the AI only social network built on top of OpenClaw leaked 1.5 million API keys, 35,000 email addresses, and thousands of private messages.

Over 500 security issues remain open on GitHub right now. Paola Alto Networks called it a lethal trifecta of risk because it combines high autonomy, broad systems access, and open internet connectivity. One of OpenClaw's own maintainers warned on Discord that if you can't understand how to run a command line, this project is far too dangerous for you. Steinberg knows this, right?

He's called it a hobby project that requires careful configuration. He's been candid that it's not meant for non-technical users in its current form. And that's exactly why the open AI hire makes strategic sense. OpenClaw proved the demand.

It proved that people desperately want AI agents that manage their digital lives. But it also proved that making those agents secure, safe, and accessible to normal people is a massive unsolved engineering challenge. That's a gap OpenAI is trying to fill. This maps perfectly onto OpenAI's broader strategy.

They just launched OpenAI Frontier, a platform for enterprises to build, deploy, and manage AI agents with the kind of oversight you'd give human workers. They co-founded the Agentic AI Foundation under the Linux Foundation alongside Anthropic and Google. They ship chat GPT agent, which can browse a web, take actions across connected apps, and schedule recurring tasks. Altman told reporters that with AI coding tools, you can go from a clean sheet of paper to sophisticated software in hours, and the only limit is how fast you can type new ideas.

So Steinberger's hire is OpenAI saying, we're not just going to build agents. We're going to absorb the person who proved agents have product market fit at the consumer level, and we're going to do it whilst his community is alive as open source. Now here's the part that should make every person watching this think hard about their own situation. Steinberger tinkered through 43 failed projects before hitting it big with OpenClaw.

He'd been retired. He was playing around. And then one weekend hack built in an hour became the subject of billion dollar acquisition offers within three months. This is the new speed of technology.

The gaps between I have an idea and the CEO of Meta is calling me has collapsed to weeks. Steinberger predicts that agents like OpenClaw will kill eight of apps. His argument is straightforward. Every app is just a slow API now.

Like why would you pay for MyFitnessPal when your agent already knows your location, sleep patterns, and stress levels? Why open UberEats when your assistant can order food, schedule meetings, and manage your calendar proactively? If that prediction is even half right, the implications for every software company, every app developer, and every knowledge worker are enormous. I get the skepticism.

I also get that people are tired of hearing the world is changing every couple of months, but here's what I keep coming back to. The security problems with OpenClaw are real. A hype around agentic AI is partially overblown. Gartner predicts over 40% of agentic AI projects will fail by 2027.

These are fixable engineering problems, not fundamental limitations. The direction is clear, even if the timeline is debatable. So what does this mean for you specifically? If you're a developer, pay attention to the feedback loop philosophy.

Steinberg embodies, right? Stop thinking about writing code line by line. Start thinking about how to design systems where AI agents can compile, test, debug, and validate their own output. That's a skill that gets a solo developer hired by the biggest AI lab on the planet.

If you're a non-technical person, understand that agents managing your email calendar and digital life are coming to mainstream products within the next 12 months. OpenAI just hired the guy who proved the concept. Start thinking about what tasks you delegate if you could. If you're a manager or executive, the revenue per employee equation is about to shift dramatically.

Again, open quote proved that one person with AI agents can build what used to take teams. That has all child implications. You need to start planning for now, not next quarter. And if you're a founder, the lesson from Steinberg's journey is brutal and beautiful at the same time.

He didn't raise money. He didn't build a team. He didn't write a business plan. He tinkered obsessively, failed 43 times, and then built something in an hour that changed his life.

The barrier to entry has never been lower. The speed of impact has never been faster. And the ceiling of what one person can accomplish has never been higher. But here's the thing that I keep thinking about.

Steinberger designed and described choosing between OpenAI and Meta as similar in difficulty to some of his worst breakups. He said he doesn't care about money. He said if it doesn't work out, he'll just do his own thing again. And that kind of leverage, that kind of freedom came from 13 years of building something real, selling it, and then having the runway and the skills to experiment without pressure.

Not everyone has that luxury. And as AI agents become more capable, as one person really can do the work of 10, the question isn't what gets built, it's what happens to the people who used to do that work. And Steinberger's story is thrilling, but the world, it points towards demands that we think seriously about transition, about retraining, about making sure the benefits of this technology don't concentrate in the hands of the few people who know how to weld it. That's the tension underneath all of this.

The technology is extraordinary. The speed is unprecedented and the human questions it raises have never been more urgent. So thanks so much for watching. If you haven't already, check out the AI Profit Boarding.

This is my community that shows you how to save time, scale your business, and grow with AI automation. You connect with me personally inside there, and also you get access to a community of 2,300 members. Plus you get video calls and coaching calls each week, four times a week. And you get all of my best playbooks for using AI, including a three and six hour course on OpenClaw, the updates that are coming out every single day.

I cover video tutorials and step-by-step guides on them, and I show you exactly how to scale your business, get more customers, generate more leads, and save time with AI. So feel free to check that out. Link in the comments description or go to AIProfitBoarding.com. AIProfitBoarding.com.

Thanks for watching.