Transcript: OpenAI Grabs OpenClaw’s Creator

Source: Tech Brew Ride Home | Duration: 20 min

Summary

Here is a comprehensive summary of the key points from this podcast episode:

Opening context:

• The main topic of this episode is the news that Peter Steinberger, the creator of the popular AI agent OpenClaw, is joining OpenAI.
• Steinberger is a prominent figure in the AI industry, having previously founded and sold the company PSPDFKit.

Key discussion points and insights:

• Steinberger is joining OpenAI to "drive the next generation of personal agents." However, the OpenClaw project will remain open-source.
• Steinberger explained that he doesn't want to run a large company, but rather focus on building impactful AI tools. Joining OpenAI allows him to achieve this goal without the hassle of running a startup.
• There was a bidding war for Steinberger, with both Meta and OpenAI making multi-billion dollar offers. However, Steinberger was more interested in finding the right partner to help scale his vision.
• Steinberger has a different philosophical view than OpenAI's focus on Artificial General Intelligence (AGI). He believes in specialized AI agents collaborating, rather than a single "god model" that can handle everything.
• The article discusses how OpenClaw will likely follow the Chrome/Chromium model, where the open-source version is maintained by the community while OpenAI develops a commercial version.

Notable technologies, tools, or concepts mentioned:

• OpenClaw - The popular AI assistant created by Steinberger that exploded in popularity.
• MoltBook - A social network for AI agents created by Steinberger.
• AGI (Artificial General Intelligence) - The goal of creating a single AI system that can handle any task, which is OpenAI's stated mission.
• Specialized AI agents - Steinberger's vision of having multiple collaborative AI agents, rather than one AGI system.

Practical implications or recommendations discussed:

• The article suggests that Steinberger joining OpenAI could have significant implications for the future development and distribution of AI assistants, given his unique perspective.
• It highlights the tension between open-source projects and the commercialization efforts of large tech companies, and how that dynamic often plays out.
• More broadly, the episode touches on the memory chip shortage and its impacts on the tech industry, as well as concerns around the use of AI systems in sensitive military applications.

Overall, this episode provides an in-depth look at a significant move in the AI industry, offering insights into the motivations and strategies of key players. It highlights the ongoing evolution and challenges in this rapidly advancing field.

Full Transcript

[00:00:00] Welcome to the TechBrew Ride Home for Monday, February 16, 2026. I'm Brian McCullough. Today,

[00:00:09] OpenClaw's creator is joining OpenAI, but the OpenClaw project lives on.

[00:00:12] The AI-caused memory shortage might delay the next PlayStation. Is the Pentagon about to cut

[00:00:17] ties with Anthropic? And Vitalik Buterin is growing concerned with the prediction markets.

[00:00:22] Here's what you missed today in the world of tech.

[00:00:30] Peter Steinberger, the creator of OpenClaw, is joining OpenAI, quote,

[00:00:36] to drive the next generation of personal agents. OpenClaw itself will remain open source,

[00:00:42] quoting The Verge. Sam Altman announced on X that Peter Steinberger, the man behind the trendy AI

[00:00:47] agent OpenClaw, was joining OpenAI. He said that Steinberger has, quote, a lot of amazing ideas

[00:00:52] about getting AI agents to interact with each other, saying the future is going to be extremely

[00:00:56] multi-agent. He also said that this ability for agents to work together will quickly become core

[00:01:01] to our product offerings. OpenClaw, previously known as MoltBot and ClawedBot, exploded on the

[00:01:07] scene earlier this year and became the darling of the tech world. Its rise was swift, but not

[00:01:12] without its bumps along the way. Earlier this month, researchers found over 400 malicious skills

[00:01:16] uploaded to ClawHub. It also launched MoltBook, a social network where AI agents went to complain

[00:01:22] about their humans, debate the provability of consciousness, and discuss the need for a private

[00:01:27] place to exchange ideas. And then it was immediately infiltrated by humans. In a post on his personal

[00:01:33] site, Steinberger said that joining OpenAI would allow him to achieve his goal of bringing AI

[00:01:37] agents to the masses without the headaches of running a company. He explained, quote,

[00:01:41] I could totally see how OpenClaw could become a huge company and no, it's not really exciting for

[00:01:46] me. I'm a builder at heart. I did the whole creating a company game already, pouring 13 years

[00:01:51] of my life into it and learned a lot. What I want is to change the world, not build a large company,

[00:01:56] and teaming up with OpenAI is the fastest way to bring this to everyone, end quote.

[00:02:00] And this is from Implicator.ai, quote, Mark Zuckerberg needed 10 minutes. He was finishing

[00:02:06] code. Peter Steinberger had called Zuckerberg on WhatsApp without scheduling anything.

[00:02:11] I don't like calendar entries, he told Lex Friedman last week. Let's just call now.

[00:02:16] Zuckerberg asked for a brief pause, then picked up. The first 10 minutes devolved into an argument

[00:02:20] about whether ClaudeCode or Codex was the better programming tool. The CEO of a trillion-dollar

[00:02:25] company squabbling with a solo developer from Vienna over IDE preferences. That was two weeks

[00:02:31] ago. Zuckerberg ran OpenClaw on his own machine afterward, gave feedback that was blunt and

[00:02:36] specific, calling features great or shit in real time, used it until it broke, and then sent notes

[00:02:41] on what to fix. Steinberger called it the biggest compliment because it shows they actually care

[00:02:46] about it. Steinberger has been losing money on OpenClaw since November, $10,000 to $20,000 a

[00:02:51] month by his own count. He routes sponsorship revenue to the developers who maintain his

[00:02:56] dependencies rather than keeping it. The project hit 196,000 GitHub stars and pulled 2 million

[00:03:01] visitors in a single week while its creator subsidized everything from savings. He doesn't

[00:03:06] need the money. He spent 13 years building PSPDFKit into a PDF tool company worth over $100 million

[00:03:13] before selling it to Insight Partners. Three years of what he described as soul-searching

[00:03:17] followed. Therapy, ayahuasca, 43 failed projects, then OpenClaw caught fire. His negotiating edge

[00:03:24] came from something no check could replicate. Both Meta and OpenAI made concrete offers

[00:03:28] reportedly valued in the billions. VCs lined up. Steinberger told Friedman he didn't care.

[00:03:34] I don't give an F were his exact words. When you've already sold a company and your next

[00:03:38] project goes viral by accident, the dynamics flip completely. He wasn't selling. They were

[00:03:43] auditioning. Steinberger told a Y Combinator podcast this month that AGI is the wrong goal.

[00:03:48] What can one human actually achieve? Do you think one human being could make an iPhone or one human

[00:03:53] being could go to space, he said? As a group, we specialize. As a larger society, we specialize

[00:03:58] even more. His vision runs on specialized agents collaborating, not one god model that handles

[00:04:04] everything. OpenAI's entire corporate identity rests on achieving artificial general intelligence.

[00:04:09] The name says it. So does the $500 billion valuation. Steinberger just joined a company

[00:04:14] whose stated mission he publicly rejects. Ignore the ideology for a second. The logistics explain

[00:04:20] everything. Steinberger does not want to run a company. 13 years of it was enough. He wants to

[00:04:24] build agents everyone can use. That means compute, APIs, and 300 million people already opening

[00:04:29] ChatGPT every week This is pragmatism dressed as alignment Meta offered something different more personal in fact Zuckerberg hands engagement impressed Steinberger Steinberger recounted it on Friedman

[00:04:41] show, acting out the reactions. Mark basically, oh, this is great. Oh, this is shit. Oh,

[00:04:46] it needs to change this. He noted the contrast with OpenAI. I didn't get the same on the OpenAI

[00:04:52] side. He chose OpenAI anyway. Zuckerberg codes, gives real feedback, and clearly cares. None of

[00:04:58] that ships a product to hundreds of millions of users. For a builder who once reached without

[00:05:03] management overhead, that settled it. Steinberger drew the comparison himself. OpenClaw would follow

[00:05:08] the Chrome and Chromium model, a foundation to hold the open source project, a corporate partner

[00:05:14] to build the commercial version. I think this is too important to just give to a company and make

[00:05:19] it theirs, he told Friedman. He's describing a pattern with a known ending. Google open source

[00:05:24] Chromium. Chrome is what everyone downloads. Android AOSP is open. Google's Android with

[00:05:29] Play services runs the planet. MySQL went to Oracle. The community forked it into MariaDB.

[00:05:36] MariaDB survives. MySQL still owns the market. Every foundation arrangement in tech follows the

[00:05:42] same gravitational pull. The corporate version gets full-time engineers, marketing, distribution,

[00:05:47] and the daily attention of the person who created the project. The open-source twin

[00:05:51] gets volunteers and good intentions. Gravity always wins, end quote.

[00:06:03] Ah, AI. Sources tell Bloomberg that Sony is considering pushing back the debut

[00:06:10] of its next-generation PlayStation console to 2028 or even 2029 because of the AI boom

[00:06:18] fueling huge memory chip shortages.

[00:06:22] Quote,

[00:06:23] Sony is now considering pushing back the debut of its next PlayStation console to 2028 or

[00:06:27] even 2029, according to people familiar with the company's thinking.

[00:06:30] That would be a major upset to a carefully orchestrated strategy to sustain user engagement

[00:06:35] between hardware generations.

[00:06:37] Close rival Nintendo, which contributed to the surplus demand in 2025 after its new Switch

[00:06:42] 2 console drove storage card purchases, is also contemplating raising the price of that

[00:06:48] device in 2026, people familiar with its plan said. Sony and Nintendo representatives didn't

[00:06:53] respond to requests for comment. Since the start of 2026, Tesla, Apple, and a dozen other major

[00:06:58] corporations have signaled that the shortage of DRAM, or dynamic random access memory,

[00:07:02] the fundamental building block of almost all technology, will constrain production. Tim Cook

[00:07:07] warned it will compress iPhone margins. Micron Technology called the bottleneck unprecedented.

[00:07:13] Elon Musk got to the intractable nature of the problem when he declared Tesla is going to have

[00:07:17] to build its own memory fabrication plant. We've got two choices, hit the chip wall or make a fab,

[00:07:23] he said in late January. The resulting price spikes are starting to look a bit like the

[00:07:27] Weimar Republic's hyperinflation. The cost of one type of DRAM soared 75% from December to January,

[00:07:35] accelerating price hikes throughout the holiday quarter. A growing number of retailers and

[00:07:39] middlemen are changing their prices every day. Ramageddon is the term some use to describe what's

[00:07:45] coming. We stand at the cusp of something that is bigger than anything we've faced before,

[00:07:50] Tim Archer, Chief Executive Officer of Chip Equipment Supplier at Lam Research,

[00:07:54] said at a conference in South Korea this month. What is ahead of us between now and the end of

[00:07:58] this decade in terms of demand is bigger than anything we've seen in the past and in fact

[00:08:02] will overwhelm all other sources of demand. What's worrying about the trend is that prices

[00:08:07] are soaring and supplies are running dry even before the AI giants really get going with their

[00:08:12] data center construction plans. Alphabet and Amazon just announced plans for a construction

[00:08:16] blitz this year that could reach $185 to $200 billion, respectively. More money than any company

[00:08:23] in history has poured into capital expenditures in a single year. Mark Lee, a Bernstein analyst who

[00:08:29] tracks the semiconductor industry, warns that memory chip prices are going parabolic. While

[00:08:34] that will bring lavish profits to Samsung, Micron, and SK Henix, the rest of the electronic sector

[00:08:40] will pay a painful price in the months ahead. Right now, we're kind of in the middle of a storm

[00:08:44] that we are dealing with hour by hour and day by day, Steinar Sans-Tobi, CEO of Norwegian IT firm

[00:08:51] Etea ASA, told analysts in February. It's actually wiser to hold off doing business today as prices

[00:08:57] are almost certain to be higher tomorrow, said Su-Yong Hwan, who runs three DIY PC shops in Seoul

[00:09:04] and frequently does business with stalls at Sunan Plaza Unless Steve Jobs rises from the dead to declare that AI is nothing but a bubble this trend is likely to persist for some time he said

[00:09:22] Just because your team is small, that doesn't mean that they're not a target for bad actors.

[00:09:28] Cyber criminals know that lean teams often lack the resources to prevent or respond to a breach.

[00:09:34] The good news is even the smallest teams can foil cybercrime.

[00:09:39] 1Password provides simple security to help small teams manage the number one risk that bad actors exploit, which is weak passwords.

[00:09:48] Take the first step to better security by securing your team's credentials.

[00:09:52] Find out more at 1Password.com slash ride so you can secure every login.

[00:09:58] That's 1Password.com slash ride.

[00:10:00] Managing your cap table shouldn't drain your time or derail your budget, and yet somehow it can manage to do both.

[00:10:10] Pulley knows there's a better way. That's why they help take the complexity and surprises out of equity management.

[00:10:16] Pulley's intuitive workflows, built-in compliance tools, and decision-ready reporting are designed to work for you, not against you.

[00:10:23] Pooley helps you issue, track, and manage equity, stay compliant with up-to-date 409A valuations,

[00:10:29] complete stock-based compensation reporting, and more, all without the expensive legal fees or

[00:10:35] endless manual work. Learn more and get started at Pooley.com slash brew. That's Pooley.com slash brew.

[00:10:46] ByteDance says it respects IP rights, and we have heard the concerns regarding SeedDance,

[00:10:52] and it plans to strengthen safeguards around intellectual property and copyright.

[00:10:57] This comes after Disney and others made legal threats against ByteDance, quoting the BBC.

[00:11:03] On Friday, Disney sent a cease and desist letter to ByteDance accusing it of supplying

[00:11:06] SeedDance with a pirated library of the studio's copyrighted characters, including

[00:11:10] those from Marvel and Star Wars. Disney's lawyers accused ByteDance of committing a

[00:11:15] virtual smash and grab of their intellectual property, including superheroes from Marvel,

[00:11:19] Star Wars, and various cartoons. On Monday, ByteDance told the BBC that the company respects

[00:11:25] intellectual property rights and we have heard the concerns regarding Seed Dance 2.0. We are

[00:11:29] taking steps to strengthen current safeguards as we work to prevent the unauthorized use of

[00:11:33] intellectual property and likenesses by users. ByteDance did not respond to questions asking

[00:11:37] for details on the safeguards it plans to implement. Like other generative AI tools,

[00:11:42] Seed Dance can create videos based on short text prompts. Many of Seed Dance's clips are based on

[00:11:47] real actors and shows, and some have gone viral since the launch of its latest 2.0 version

[00:11:51] on February 12th. The BBC has found clips online said to have been generated by C-Dance showing

[00:11:57] Star Wars characters Anakin Skywalker and Rey battling with their lightsabers and Spider-Man

[00:12:02] fighting Captain America on the streets of New York. The company has not disclosed what data it

[00:12:06] used to train C-Dance. ByteDance had previously said the product had already paused the ability

[00:12:11] for users to upload images of real people. The company also said it respects intellectual property

[00:12:15] rights and copyright protections, and it takes any potential infringement seriously. Disney last

[00:12:20] year struck a $1 billion deal with the maker of ChatGPT and video generation tool Sora, OpenAI,

[00:12:26] giving the platforms access to 200 characters from its franchises, including Pixar, Marvel,

[00:12:31] and Star Wars, end quote. Various Trump administration officials were saying over

[00:12:42] the weekend that the Pentagon may sever its ties with Anthropic over Anthropic's AI safeguards.

[00:12:49] Anthropic says only mass surveillance and fully autonomous weapons are off limits for its partners

[00:12:55] to use. Quoting Axios, the senior administration official argued there is considerable gray area

[00:13:01] around what would and wouldn't fall into those categories and that it's unworkable for the

[00:13:05] Pentagon to have to negotiate individual use cases with Anthropic or have Claude unexpectedly block

[00:13:10] certain applications. Everything's on the table, including dialing back the partnership with

[00:13:14] Anthropic or severing it entirely, the officials said. But there will have to be an orderly

[00:13:19] replacement for them if we think that's the right answer. An Anthropic spokesperson said the company

[00:13:24] remained committed to using Frontier AI in support of U.S. national security. The Pentagon is pushing

[00:13:29] four leading AI labs to let the military use their tools for, quote, all lawful purposes,

[00:13:34] even in the most sensitive areas of weapons development, intelligence collection, and

[00:13:39] battlefield operations Anthropic has not agreed to those terms and the Pentagon is getting fed up after months of difficult negotiations The tensions came to a head recently over the military use of Claude in the operation to capture Venezuela Nicolas Maduro through Anthropik partnership with AI software firm

[00:13:55] Palantir. According to the senior official, an executive at Anthropik reached out to an executive

[00:13:59] at Palantir to ask whether Claude had been used in the raid. It was raised in such a way to imply

[00:14:04] that they might disapprove of their software being used because, obviously, there was kinetic fire

[00:14:08] during that raid. People were shot, the official said. Beyond the Maduro incident, the official

[00:14:13] described a broader culture clash with what the person claimed was the most ideological

[00:14:17] of the AI labs when it came to the potential dangers of the technology, but the official

[00:14:22] conceded that it would be difficult for the military to quickly replace Claude because,

[00:14:26] the other model companies are just behind when it comes to specialized government applications.

[00:14:31] Anthropic signed a contract valued at up to $200 million with the Pentagon last summer. Claude was

[00:14:36] also the first model the Pentagon brought into its classified networks, end quote.

[00:14:47] Finally today, Vitalik Buterin, who, it should be noted, is an investor in Polymarket,

[00:14:53] says he's, quote, starting to worry that prediction markets, quote, seem to be

[00:14:58] over-converging to an unhealthy product market fit, quoting Crypto News. Writing on X, Buterin

[00:15:04] called the trend corpo-slop and argued platforms feel pressure to embrace dopamine-driven content

[00:15:09] that lacks long-term societal value. Buterin proposed redirecting prediction markets toward

[00:15:15] hedging use cases, including a system where personalized prediction market baskets replace

[00:15:19] fiat currency entirely. We do not need fiat currency at all. People can hold stocks,

[00:15:25] ETH, or whatever else to grow wealth and personalized prediction market shares

[00:15:29] when they want stability, he wrote. Buterin identified three types of actors willing to

[00:15:33] lose money in prediction markets, naive traders with incorrect opinions, info buyers running

[00:15:38] automated market makers to learn information, and hedgers using markets as insurance to reduce risk.

[00:15:43] The industry currently depends on naive traders and creates what Buterin called a fundamentally

[00:15:47] cursed dynamic. Quote, it gives the platform the incentive to seek out traders with dumb opinions

[00:15:52] and create a public brand and community that encourages dumb opinions to get more people to

[00:15:57] come in. This is the slide to Corposlop, he wrote. Buterin questioned whether an ideal stablecoin

[00:16:02] based on decentralized global price indices is the right solution. What if the real solution is

[00:16:08] to go a step further and get rid of the concept of currency altogether, he asked. The proposed

[00:16:12] system creates price indices for all major categories of goods and services, treating

[00:16:16] physical items in different regions as separate categories. Each user maintains a local large

[00:16:21] language model understanding their expenses, offering personalized baskets of prediction

[00:16:24] market shares representing future spending needs. Users could hold stocks, ETH, or other assets for

[00:16:30] wealth growth while holding personalized prediction market shares for stability.

[00:16:34] The system removes fiat currency dependence while allowing customization for individual

[00:16:37] expense patterns. Implementation would need prediction markets denominated in assets people

[00:16:42] want to hold, interest-bearing fiat, wrapped stocks, or ETH. Non-interest-bearing fiat carries

[00:16:47] opportunity costs that overwhelm hedging value. Both sides of the equation are likely to be

[00:16:52] long-term happy with the product that they are buying, and very large volumes of sophisticated

[00:16:56] capital will be willing to participate, Buterin concluded.

[00:17:00] you know that thing online that people have been talking about about how you blow days of your life

[00:17:14] setting up ai bots to do things that really don't really amount to much in the end but you spend all

[00:17:20] your time trying to get them to do whatever you're trying to do and so the effort is basically the

[00:17:26] whole game, and then you're not actually gaining any time because you're spending all your time

[00:17:31] managing the bots. Well, this weekend, that was me. My wife and I went to see the Stranger Things

[00:17:37] play on Broadway. They were filming it for a Netflix special. And the whole time I'm sitting

[00:17:42] in the audience just waiting to get back home to make sure my bots didn't break. Talk to you tomorrow.

[00:17:51] There's a magical place where the desert meets the sea.

[00:17:55] offering adrenaline rushes

[00:17:57] and now we're heading off on the jet ski

[00:17:59] I'm on a surfboard

[00:18:00] seagoing adventures

[00:18:02] we are looking at the dolphins

[00:18:03] and sometimes we see the turtles also

[00:18:06] and the ultimate in luxury resorts

[00:18:08] nice to have you here in paradise

[00:18:10] Abu Dhabi Unwrapped is a sonic guide

[00:18:12] to a very special destination

[00:18:14] paid and presented by the Department of Culture and Tourism

[00:18:18] Abu Dhabi

[00:18:18] listen now wherever you get your podcasts