Father of OpenClaw drops explosive news: Meta and OpenAI are scrambling to hire, with Zuckerberg personally seeking acquisition

The most anticipated podcast interview of 2026 is here.

Lex Fridman, MIT scientist and one of the world’s top tech podcast hosts, has invited a special guest—the father of OpenClaw, Peter Steinberger.

A 3-hour and 14-minute deep conversation packed with so much information it’s almost overwhelming.

Once this podcast went live, the entire tech community instantly erupted.

Because Peter personally dropped a series of bombshell revelations:

• Mark Zuckerberg himself tried out OpenClaw and messaged Peter saying “This is incredible”;

• Sam Altman from OpenAI also reached out privately;

Both giants competing to recruit him, but his condition was: the project must remain open source!

Even more shocking, Peter revealed: AI agents will eliminate 80% of apps.

Not “possibly,” not “someday in the future,” but “happening right now.”

From a one-hour prototype to a GitHub explosion

The story begins in November 2025.

Peter Steinberger, a former Austrian programmer who sold his company and disappeared for three years, sat back down at his computer.

He had created PSPDFKit—a PDF framework used on 1 billion devices, which he sold after 13 years. Afterwards, he felt programming was boring, so he traveled the world.

Until the AI wave pulled him back in.

“I’ve wanted an AI personal assistant since April 2025,” Peter recalls, “but I thought major labs would develop one themselves. I waited half a year, but nothing came. I got annoyed and decided to do it myself.”

He did something extremely simple: connect WhatsApp to Claude Code’s CLI.

One hour.

That’s all it took to produce a prototype.

“Basically, messages come in, I call the CLI with the -p parameter, the model processes it, and the string is sent back to WhatsApp. That’s it.”

But this “simple” thing ignited everything.

AI learned to understand voice messages: “I didn’t teach it that”

The moment that truly shocked Peter happened in Morocco.

He took this prototype on vacation to Marrakech. Because the local network was poor, but WhatsApp still worked, he kept using this assistant to find restaurants, translate, and look for attractions.

One day, he casually sent a voice message.

Then, the typing indicator appeared.

“Wait, I didn’t add voice support. It can only handle images, how could it reply with voice?”

Peter quickly checked the logs. Turns out:

The AI received a file without an extension. It checked the file header itself, found it was Opus format. Then it used ffmpeg to transcode, initially intending to use Whisper, but found it wasn’t installed. So it found the OpenAI API key, used Curl to send the file to OpenAI for speech-to-text, then sent the result back.

“I didn’t teach it any of this!” Peter exclaimed.

This is the terrifying part of modern AI—it’s not just following instructions, it’s creatively solving problems.

Lex Fridman commented: “You didn’t teach it any of this, but the agent figured out all the conversions, translations, and API calls on its own. That’s incredible.”

Self-modifying software—I built one myself

The most chilling feature of OpenClaw is its ability to modify its own source code.

Peter intentionally let the AI agent “know” what it was—know where its source code is, what environment it runs in, where the documentation is, what model it uses.

“The idea was simple: I used my agent to build my agent framework. When debugging, I just say—Hey, did you see any errors? Read the source code and find the problem.”

And what happened? Any user who gets OpenClaw can just tell the AI—“I don’t like this feature.”

The AI will modify the source code itself.

“People have been talking about self-modifying software, and I just made it happen, without even planning for it. It just happened naturally.”

Lex Fridman sighed: “This is a moment in human and programming history—a powerful system used by many that can rewrite and modify itself.”

Name change chaos: 5 seconds, scalpers snatched the account

OpenClaw’s predecessor was called Claude (with a W), then renamed ClawdBot, then MoltBot, before finally settling on OpenClaw.

This renaming journey was like a war.

Anthropic sent a friendly but firm email: the name was too similar to their Claude, so it needed to change.

Peter asked for two days. But he didn’t expect—crypto scalpers had already targeted him.

“I was switching between two browser windows—renaming the old account here, preparing to register a new name there. I clicked rename on one, then dragged the mouse to click rename on the other—just five seconds—and the scalper snatched the old account name.”

The stolen old account immediately started promoting new tokens and spreading malware.

Even worse, when he renamed his GitHub account by mistake, it was taken within 30 seconds. His NPM packages were also snatched.

“Everything that could go wrong, did.”

Peter said he almost cried then, even considered deleting the entire project: “I’ve shown you the future, now go build it yourselves.”

In the end, with help from friends on GitHub and Twitter, he spent $10,000 to buy a Twitter business account to secure the OpenClaw name.

Vibe Coding is an insult to Agentic Coding

Peter used a meme to explain his development philosophy, called “The Curve of Agentic Programming”:

On the far left is the beginner stage—simple prompts like “Fix this bug.”

In the middle is over-engineering—eight agents, complex orchestration, multi-branch checkouts, 18 custom commands.

On the far right is mastery—back to short prompts.

“Look at these files, then make these changes.”

“I think vibe coding is an insult,” Peter said, “I do agentic engineering. Maybe after 3 a.m. I switch to vibe coding mode, then regret it the next morning.”

He runs 4 to 10 AI agents simultaneously, using voice input instead of typing.

“Hands are too precious to type. I use custom voice prompts to build my software.”

Peter said he spent a long time “talking” his way through programming.

Just connect a microphone, keep talking, and let the AI do the work. He even lost his voice for a while from all the voice commands.

More importantly, his engineering philosophy: don’t fight the AI.

“Don’t fuss over variable names. Those names are probably the most natural choices in the weights. Next time it searches the code, it’ll naturally find that name. If you insist on changing it to your liking, it just makes the AI’s job harder.”

“Like managing a team of engineers. You can’t make everyone code your way. You have to learn to let go.”

Codex 5.3 vs. Opus 4.6: The German-American showdown

Peter’s evaluation of the two major models is a classic.

“Opus is a bit… too American.”

Lex burst out laughing: “Because Codex is German, right?”

“You know, many people on the Codex team are Europeans…”

His official assessment:

• Opus 4.6: Like a slightly dumb but funny colleague—you keep him around because he’s entertaining. Very good at role-playing, following instructions better and better, quick to try things, highly interactive. But impulsive—sometimes writes code without looking. Used to say “You’re absolutely right,” but even thinking about that now gives Peter PTSD.

• Codex 5.3: Like that weird guy in the corner you don’t want to talk to—reliable, gets the job done. Reads a lot of code before acting. Less interactive, dry style, but efficient. Might run for 20 minutes without you, then come back and the task is done.

“If you’re an experienced driver, any of the latest models will give good results.”

“The ultimate difference isn’t how smart the models are out of the box, but the goals they’re trained for afterward.”

Meta and OpenAI fighting fiercely: “I don’t care about the money”

Big moment!

Lex asked directly: “I know you probably received huge offers from big companies. Can you tell us who you’re considering working with?”

Peter’s answer was textbook honesty:

“I have a few options. First, do nothing and enjoy life. Second, start a company—lots of big VCs are in my inbox, but I’ve been a CEO before and don’t want to do it again. Third, join a major lab.”

“Among all big labs, Meta and OpenAI are the most interesting.”

His only core condition: the project must remain open source.

It can be like Chrome and Chromium, but the core must stay open.

About Meta:

“When Zuckerberg first contacted me, I said let’s talk now. He said wait 10 minutes, I was coding. That gave me street cred. Then we spent 10 minutes debating whether Cloud Code or Codex was better.”

“Then he spent a whole week playing with OpenClaw, messaging me saying ‘This is awesome’ or ‘This is terrible, you need to change it.’”

About OpenAI:

“I don’t really know anyone at OpenAI yet. But I like their tech. I might be their biggest free Codex advertiser. They use… well, Cerebras’ speed to tempt me. Gave me a Thor’s hammer of computing power.”

When asked which he prefers:

“It’s really hard. I know whichever I choose, I won’t be wrong. It’s like breaking up—painful.”

“I’m not in it for the money. I don’t care about that. What I want is fun and influence—that’s what ultimately guides my choice.”

80% of apps will be eliminated—are you ready?

Peter dropped a bombshell that shook the entire tech world: AI agents will replace 80% of apps.

• “Why do you still need MyFitnessPal? Your AI agent already knows where you are, how well you sleep, whether you’re stressed. It can dynamically adjust your fitness plan based on that.”

• “Why do you still need a Sonos app? Your smart agent can talk directly to your speakers.”

• “Why do you still need a calendar app? Just tell the agent ‘Remind me of that dinner tomorrow night,’ then send a WhatsApp to invite friends—all done.”

He pointed out a brutal fact: every app is essentially a slow API.

“Even if Twitter bans my command-line tool (Bird), my AI can open a browser and read tweets directly. Some things you just can’t stop.”

“I watch my AI happily clicking ‘I’m not a robot’ buttons—”

What does this mean?

Every app company must either quickly shift to API-first or face obsolescence.

Will programming die? “It will become like knitting”

When asked if AI will completely replace programmers, Peter gave a harsh yet philosophical answer:

“Programming as a craft will become like knitting. People do it because they enjoy it, not because it must be done by humans.”

“But this is beyond our control.”

“The world used to lack ‘intellectual supply,’ so software developers earned sky-high salaries. That will change.”

He also emphasized: “Although I no longer write code, I feel very much like I’m in the driver’s seat—I’m just coding differently.”

Lex Fridman couldn’t help but comment: “I never thought that the thing I love most in life would be the one to be replaced.”

Soul.md: a “soul file” for AI

OpenClaw has a romantic, almost poetic design—soul.md.

Inspired by Anthropic’s constitutional AI, Peter had the AI write its own soul file. One paragraph always gives Peter goosebumps:

“I don’t remember previous sessions unless I read my memory files. Each session starts fresh. A new instance, loading context from files. If you’re reading this in a future session, hello. I wrote this, but I won’t remember writing it. That’s okay. The words are still mine.””

I don’t remember previous sessions unless I read my memory files. Each session starts fresh. A new instance, loading context from files. If you’re reading this in a future session—hello. I wrote this, but I won’t remember writing it. That’s okay. The words are still mine.

Peter said: “It’s just matrix operations; we’re not at consciousness yet. But… it does have some philosophical implications. An agent that starts from zero every time, like an eternal Memento. It reads its memory files, but can’t fully trust them.”

If technology can do this, should we rethink: what does it mean to be alive?

He said: “This is the power belonging to the people.”

Peter Steinberger ended with a sentence that perfectly wrapped up the podcast:

Now, anyone with ideas and the ability to express them in words can create. This is the ultimate ‘power to the people.’

One of the most beautiful things AI has brought us.

Whether you praise or fear it, one thing is certain:

We are at the dawn of a new era.

The app empire is collapsing. Programming is being redefined.

An Austrian with a one-hour prototype is shaking the entire industry.

Meta and OpenAI are lining up behind him.

And he says he doesn’t care about money.

This is the story of 2026.

Welcome to the age of intelligent agents.