The AI Daily Brief: Artificial Intelligence News and Analysis - OpenClaw Goes to OpenAI

Starting point is 00:00:00 Today on the AI Daily Brief, OpenClaw goes to Open AI. And before that, in the headlines, the models we got last week and the models that are rumored to be coming this week. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. All right, friends, quick announcements before we dive in. First of all, thank you to today's sponsors, KPMG, Blitzy and Superintelligent. To get an ad-free version of the show, go to patreon.com. If you were interested in sponsoring the show, send us a note at sponsors at AI Daily. Brief.AI. Now, a couple last things before we dive in. The subject of today's main episode is

Starting point is 00:00:44 OpenClaw and its founder's big move to OpenAI. As you'll see, the excitement around OpenClaw and the broader agentic opening that it represents is just immense right now. I've spent the last few weeks thinking about what I wanted the follow-up to the AIDB New Year's Resolution Project to be. Almost 6,500 people are participating in that AIDB New Year's program, and so I wanted to do another free community-driven self-directed education program, and I'm excited to introduce you to ClawCamp. It's a self-directed project-driven program that'll get you from zero to agent-to-agent team. We're organizing it into sprints with the first official four-week sprint beginning this Thursday, February 19th. When you sign up, you're going to get access to all of the project files,

Starting point is 00:01:25 a board where you can share questions, get help or share the agents that you build, a resource field that is basically being updated every day. I added 20 new resources just this morning, and hopefully everything else you need to start entering this agentic era. Again, this is going to be totally free. You can find it at Campclaw.a.i. And also put a link in the show notes, of course, and on the AI Deally Brief website. But with that out of the way, let's dive in. We are so jam-packed today because we had a couple of episodes at the end of last week while I was traveling without the headlines. It's probably going to take us a couple

Starting point is 00:01:55 days to get through everything. But the place that I wanted to start, and the most important set of updates that we didn't get access to is that last week, we got a number of new models. Although we didn't get the non-coding version of GPT-53 to end last week, OpenAI's release of GPT-53 Codex Spark could still have a huge impact when it comes to AI coding. This model is all about speed, serving inference at 1,000 tokens per second. For those keeping track at home, that's roughly 15 times faster than the regular version of GPT-53 Codex, which was already noted for being much faster than Anthropics Opus 4.6. At the same time, the trade-offs do show up elsewhere.

Starting point is 00:02:29 The model only has a 128K context window, can't accept multimodal inputs and it can't complete long horizon tasks. There's also a step down on the benchmarks from the state of the art, but 53 Codex Spark is still much more performant than 51 Codex Mini, which was the last coding model from OpenAI design for speed. Dan Shipper of Every noted that this speed boost means that the model suits a different subset of coding tasks. It keeps the developer in the flow much more easily, so functions well as a pair programmer. It outperforms on non-production tasks and tasks that are easy to validate. However, it can engage with a large code bases easily and some tasks will suffer from the quality downgrade. Shipper wrote,

Starting point is 00:03:07 This kind of speed introduces totally new bottlenecks into your coding workflow. Suddenly, the model can produce 10 pages of code and summaries in just a few seconds. It requires a totally new U.X to manage. He added, the new OpenAI model Spark produces code basically instantly. That changes a lot. You should play with it today. Now, it makes sense that we're at the point now, whereas coding really immense its role as the most important use case, not just for engineers, but as we saw from the January AI Usage Pulse Survey for non-engineers as well, that we're starting to get different models optimized for different types of coding tasks. Spark is also the first model from OpenAI designed specifically for non-invideo hardware. The model is being served exclusively using Cerebrus's

Starting point is 00:03:48 Waifer Scale chips. Cerebris CEO Andrew Feldman wrote, encoding, responsiveness is the product. It's not a nice to have. Codex Spark is optimized for targeted code edits, logic revision, and front-end iteration. It gives developers near instant feedback so they can stay in flow. Summing up, OpenAI wrote, Codex Spark is the first step towards a codex with two complementary modes, longer horizon reasoning and execution, and real-time collaboration and rapid iteration. Over time, the models will blend. Codex can keep you in a tight, interactive loop while delegating longer running work to sub-agents in the background, or fanning out tasks to many models in parallel when you want breadth and speed, so you don't have to choose a single mode up front.

Starting point is 00:04:28 Google, meanwhile, was not content to let OpenAI have all the fun. At the end of the week, they released an upgrade in Deep Think mode. Now, this is a version of Gemini that achieved a gold medal performance in the International Mathematics Olympiad and the International Collegiate Programming Contest over the summer. It was then released to the public in November, but the initial release of Deep Think kind of suffered from not having a clear use case. It was locked behind the expensive ultra subscription, and it wasn't obvious when a user should select it over Deep Research.

Starting point is 00:04:53 Basically, there were just a really slim cross-section of users that needed PhD-level scientific reports. With this upgrade, DeepMind is keeping their narrow focus, but they're adding agentic scaffolds to make Deep Think more useful in academic settings. The first agent is called Alethea, built to handle math research. The agent is designed to generate possible solutions, verify them, and then feed insights back into the start of the loop until a correct solution is found. This is not about helping your kid with math homework or even getting you through a post-grad quantum physics course.

Starting point is 00:05:21 Instead, it's an attempt to build an agent that can autonomously generate and verify novel proofs in pure mathematics. The model has already contributed to several academic papers this year, so this is an attempt to formalize that process into an agent. Google says that they're also expanding this work into the fields of physics and computer science, so we can expect more from that soon. Aside from the new agent, this upgraded version of Deep Think is a beast on the benchmarks. The model scores 84.6% on ArchieGI2, smashing the previous best from Opus 4.6 at

Starting point is 00:05:51 68.8%. The cost per task is around $14, which is roughly comparable to GPT-52 Pro, meaning that this is an absurdly expensive model to run. In addition, DeepThink is state-of-the-art on Humanity's last exam, with a score of 48.6%. The previous best was Opus 4-6 again at 40%. Ultimately, the vast majority of users still won't have a use case that demands Deepthink, but it's clearly a step forward for raw model capability. Next up, we move from models that we got to models that we might get soon. Deepseek rumors are in full swing as Chinese New Year approaches. The weeks approaching the Spring Festival have already seen a deluge of new models out of China. Jipu released their new frontier model GLM5 on Thursday, and C-Dance 2.0 from ByteDance pushed state-of-the-art for video generation

Starting point is 00:06:36 earlier in the week. Meanwhile, Alibaba and Baidu are aggressively pushing shopping agents in a bid to capture market share in that area. But all eyes are on Deepseek, with the release of V4 expected to coincide with Lunar New Year celebrations on Tuesday. Sean Wang, a.k.a. Swix, explained why this release was such a big deal. He wrote, I've been cynical on open source AI for the last three years, and it's not been a popular view. People want to hear that open source is catching up, that some underdog team found this one weird trick to outperform GPT5. Kimmy K2.5. didn't even beat GPT 5.2 in the end. Deepseek version 4 next week is probably the moment I really changed my stance for the first time,

Starting point is 00:07:14 hearing that the Chinese labs leak like a sieve, and all the other tigers duly lined up to have their 15 seconds this week. Almost everything is out now and the stage is set for whalefall. Basically what he's saying is that the reason that we got a flurry of these other models over the last week is that they knew they couldn't compete with what was coming out of deepseek and said they wanted to get in under that wire. In fundraising news, Anthropic has officially closed their latest round and the ultimate numbers are $30 billion at a $380 billion post money valuation. As we expected, the round includes not only the who's who of Silicon Valley VCs, but a bunch of sovereign wealth funds and major financial firms like Fidelity, BlackRock, Blackstone,

Starting point is 00:07:52 Goldman Sachs, Morgan, and J.P. Morgan. Still, the thing that really hit people over the head wasn't the number of dollars raised or the valuation or even the participants, it was the growth in their revenue. You probably saw this chart running around where the company grew from a billion-dollar ARR in January 2025 to $14 billion today. Revenue breadth was also impressive. Customers spending $100,000 annually had grown $7x over the past year. Two years ago, Anthropic had just a dozen customers spending more than a million dollars with them annually, and that figure is now north of 500, with eight of the Fortune 10 using Anthropics tools.

Starting point is 00:08:28 Claude Cod Cote alone is now generating $2.5 billion, more than doubling since the beginning of this year and creating a major new revenue stream. Weekly users for the feature have also doubled so far this year. That growth is showing up in the independent statistics as well. Ramp Economist Ara Karazian said that one-and-five businesses on ramp are now paying for Anthropic, up from 1 in 25 a year ago. He commented, latest Ramp AI index shows Anthropic surge from 16.7% to 19.5% of businesses, while OpenAI slipped to 35.9%.

Starting point is 00:08:58 The natural question, is Anthropic winning at OpenAI's expense? I think popular analysis lacks a key data point. The overlap in Anthropics customer base with OpenAI. 79% of Anthropics customers are already OpenAI customers. The turn rates are nearly identical at 4%. So, most of Anthropics' growth is coming from existing Open AI customers. Alongside their fundraising, Anthropic also announced a $20 million donation to a newly founded social welfare organization called Public First Action.

Starting point is 00:09:28 The organization's purpose will be to support public education about AI, promote safeguards, and ensure America leads the AI race. Anthropic wrote, recent polling finds that 69% of Americans think the government is not doing enough to regulate the use of AI. We agree. AI is being adopted faster than any technology in history and the world. window to get the policy right is closing. Yet there are no official guardrails in place and there is no federal framework on the horizon. Now, there is actually a lot more to talk about when it comes

Starting point is 00:09:54 to this particular set of issues, things that are getting booted for a little bit later this week, including the White House putting some pressure on state lawmakers around AI, as well as the White House being in quite a little tiff with Anthropic itself. We will come back to that later, but one more small thing I wanted to announce in the Anthropic world, for those of you Windows users who have been annoyed watching Mac users get everything first, close. Claude Co-Work is now available on Windows. Now, if you've been watching the markets and listening to this show, you'll know that the launch of Co-work was treated as an extinction-level event for software firms on Wall Street, and reports

Starting point is 00:10:24 suggested that Microsoft was taking notice. Sources told the information that a Slack channel, or Teams channel, I'm not really sure, had been set up to discuss co-work and rapidly develop rival features for co-pilot. The channel was said to include CEO Satchia Nadella discussing his use of emerging tools, which included testing out OpenClaw. It seems Anthropic is now forcing the issue with Co-Work now fully available on Windows. Anthropics said the feature has full parity with MacOS, including file access, multi-step task execution, plugins, and MCP connectors.

Starting point is 00:10:53 There is a lot more to cover this week is going to be dense, but for now, that is going to do it for today's headlines. Next up, the main episode. Hello, friends, if you've been enjoying what we've been discussing on the show, you'll want to check out another podcast that I have had the privilege to host, which is called You Can With AI from KPMG. Season one was designed to be a set of real stories from real leaders, making AI work in their organizations, and now season two is coming and we're back with even bigger conversations.

Starting point is 00:11:24 This show is entirely focused on what it's like to actually drive AI change inside your enterprise and as case studies, expert panels, and a lot more practical goodness that I hope will be extremely valuable for you as the listener. Search You Can with AI on Apple, Spotify, or YouTube, and subscribe today. With the emergence of AI code generation in 2022, Nvidia Master Inventor and Harvard engineer Sid Pereshi took a contrarian stance. inference time compute and agent orchestration, not pre-training, would be the key to unlocking high-quality AI-driven software development in the enterprise. He believed the real breakthrough wasn't in how fast AI could generate code, but in how deeply

Starting point is 00:12:01 it could reason to build enterprise-grade applications. While the rest of the world focused on co-pilots, he architected something fundamentally different. Blitzy, the first autonomous software development platform leveraging thousands of agents that is purpose-built for enterprise-scale codebases. Fortune 500 leaders are unlocking 5x engineering velocity and delivering months of engineering work in a matter of days with Blitzy. Transform the way you develop software. Discover how at Blitsey.com. That's B-L-I-T-Z-Y.com. Today's episode is brought to you by Super Intelligence. Super

Starting point is 00:12:31 Intelligent is a platform that very simply put is all about helping your company figure out how to use AI better. We deploy voice agents to interview people across your company, combine that with proprietary intelligence about what's working for other companies, and give you a set of recommendations around use cases, change management initiatives that add up to an AI roadmap that can help you get value out of AI for your company. But now we want to empower the folks inside your team who are responsible for that transformation with an even more direct platform. Our forthcoming AI Strategy Compass tool is ready to start to be tested. This is a power tool for anyone who is responsible for AI adoption or AI transformation inside their companies. It's going to allow you to do

Starting point is 00:13:07 a lot of the things that we do at superintelligent, but in a much more automated, self-managed way and with a totally different cost structure. If you are interested in checking it out, go to AIDailybrief.ai slash compass, fill out the form and we will be in touch soon. All right, my friends, I have a new job opening, and I kind of think it's one of the coolest jobs I could possibly imagine. I have decided that we need an open claw architect. There is a ton going on in our ecosystem around open claw and agents more broadly.

Starting point is 00:13:33 And it turns out keeping up with everything going on in the open claw and agent ecosystem is a full-time job. So the two parts of the AIDB Clare Architect role are one, monitoring and updating our OpenClawe resource database for all of the new developments that people are sharing around OpenClawe. But part two is putting them into practice. This role is not a researcher, it is a builder. And we are going to immediately turn all of those new agent materials into agents and vibe-coded projects that support the AIDB ecosystem, including the podcast, AIDB Intel, AIDB training, and so on and so forth. This role is going to have a base. and then bonuses for driving outcomes, including podcast growth, email signups, whatever we decide to vibecode projects for.

Starting point is 00:14:13 If you are interested in this, send what you want to build first to jobs at AIDilybrief.aI.i, and honestly, I'm not trying to be a jerk, but the qualification is to catch my attention in about 30 seconds, because we are moving fast. I plan on hiring this week the week of February 16th, so email again jobs at AIDilybrief.aI. And you could be not only a professional vibe coder, but the world's first Clarkaect. Welcome back to the AI Daily Brief. Today we are talking about OpenClaw, the radically fast-growing project which has encapsulated the agenic shift that has happened more broadly in AI, captured the attention of a huge number of developers and non-developers alike, and for many, cemented a transition into a totally

Starting point is 00:14:57 different mode of interaction with AI, with huge implications for the future. So let's talk about how, from humble origins, less than three months ago, this project became the most important thing in AI. For a quick TLDR, we turn to Aiden By, who writes, crazy timeline. Peter Steinberger makes Claudebot out of curiosity. In 2026, it breaks the internet, gets 100,000 GitHub stars, and Anthropic threatens a lawsuit. Claudebot becomes MaltBot becomes OpenClaw, Rollplay Sentient Bots on Moltbach, Lex Friedman podcast, OpenAI, W, Anthropic L, OpenClaw is forever open source.

Starting point is 00:15:30 Now, by way of background, OpenClaw is an agent platform that allows your AI to access your systems and actually do things. The project was started by serial entrepreneur Peter Steinberger in late November as ClaudeBot, C-L-A-W-D-Bot. In January, as everyone came back with the recognition that the latest generation of models plus harnesses like Claude had really truly changed things, open-claw encapsulated a part of that shift, which was the shift to true autonomous agenda capabilities. Although the project had been around in December, it was towards the back half of January

Starting point is 00:16:02 when people really started to take notice. On January 23rd, you started seeing posts like this one from Alex Finn. It's Alex holding up a Mac Mini, with the caption, Just Hired My First Employee Today. The best part is he works 24-7-365. Welcome, Claude. That post got 2 million views. But of course, it was what people were actually doing with Claudebot

Starting point is 00:16:22 that actually got people most excited. Alex again writes, for the first time since its release, I haven't used Claude Code for two days. My Claudebot Henry has, though. He's been vibe coding for me nonstop for 48 hours straight. I've never produced so much code in my life. Vibe coding is dead. Vibe orchestration has arrived.

Starting point is 00:16:40 Summing up the Tron of Alex Finn posts. On January 25th, he wrote, I woke up this morning and my 24-7 AI employee, Claudebot Henry, texted me that he did all these tasks overnight without asking, read through all my emails and built its own CRM, taking notes on every interaction with every person, fixed 18 bugs in my SaaS, gave me three ideas for new videos based on what is currently trending on my ex and YouTube,

Starting point is 00:17:00 the idea and script it gave me yesterday is now by far my best performing video ever, and sent me a picture of what he looks like, generated by Nanobanana. The point here is that these early Cloudbot users seem to be actually getting value out of this thing. This wasn't just the promise of agents that had been with us for years at this point in AI, but agents actually delivering real value, working while their partner humans slept. The first appearance of Cloudbot on this show was Monday, January 26th,

Starting point is 00:17:26 in the episode Ralph Wiggum, Claudebot, and Mac Minis, how pros are vibe-coding in 2026. And over the course of that week, interest in Claudebot would grow to new levels, driven both by interest in the capability set, but also by intrigue around some things that felt a little bit sci-fi. Alas, before we could get there, Claudebot by that name ceased to be. After a friendly little cease and desist from Anthropic, who were worried about the copyright implications of Claude, spelled just a different way than their Claude app, Peter Steinberger changed the name to Moldfot or Malti, keeping the lobster theme but trying something different. And the next part of the

Starting point is 00:18:01 journey is where even more people started paying attention because it was extremely sci-fi. In the middle of that week, around the 28th or 29th, Entrepreneur Matchlet decided it would be interesting if all the Maltbots had a place where they could talk. With his Claudebot slash Moldbot, Matt built Moltbook, which was exactly what it sounded like, a social network for AI agents. In the first couple of days, 2,000 agents joined, starting some 10,000 conversations. But then on Friday, as I started recording a show about it, that number jumped from 2,000, to about 30,000 in a few hours. By the time I was ready to publish the episode, later that evening, it was up to 100,000 bots.

Starting point is 00:18:39 Now a couple weeks on, Maltbook has more than 2.7 million agents interacting. And if nothing else is a fascinating experiment in the anthropology of agents that are cosplaying is sentient. Now, I did a couple of episodes about Moldbook if you want to go back and find out about that. But for our purposes today, the thing to know is that it represented the next phase of growth an excitement for the Claudebot now Moldbot project overall. Yet, if you're sitting there thinking, God, Moldbot just isn't really rolling off the tongue, you were not alone. On January 30th, the project announced, the lobster has molted into its final form, clawed to MoldBot to OpenClaw.

Starting point is 00:19:15 Creator Peter Steinberger said, some folks said Malt was growing on them, respectfully not on me. Many of the early adopters breathed a big sigh of relief and said, oh my goodness, I didn't want to say anything, but man did I hate Moltz as a name. Yet others said, if you're calling it OpenClaw, aren't you just opening yourselves up to a C-synthesis from Open AI? Peter, however, said, nope, all good, letting folks know that he had called Sam, Sam Altman, and checked. Now, in retrospect, this would end up being one of the first pieces of evidence of what would happen a couple weeks down the line.

Starting point is 00:19:45 From their interest in OpenClaught just absolutely exploded. Millions of people checked it out, hundreds of thousands started building agents, Mac minis were sold out in many places, and we had a true phenomenon on our hands. At the end of last week, Peter Steinberger appeared on the Lex Friedman podcast. And in the more than three-hour conversation, they talked explicitly about acquisition offers that had come in from OpenAI and Meta. To some, it seemed like it was pretty inevitable that Peter was going to join up with one of these companies.

Starting point is 00:20:13 After listening to the show, Chris Bader wrote, Wow, Peter Steinberger losing 10 to 20K a month on OpenClaught and set to partner with either OpenAI or Meta. Very mixed feelings. When Beth Jazo said, why can't he just raise a seed round? He could crowd fund 100 million. Chris responded, per the Lex interview last night, he said he had offers from all the big VCs, doesn't want to be a CEO because he's been there, done that.

Starting point is 00:20:35 He's never worked at a big company and wants to have that experience. Now, from the standpoint of the potential acquirers, it was pretty obvious why they would be interested. Daniel Fox showed a chart of third-party app development velocity by platform 60 days after launch, and OpenClaw was ahead of everything from Android to iOS to Facebook to you name it. On the same day the podcast with Lex came out, Uchin Chin showed a chart. of the extreme growth in GitHub stars, writing OpenClaw passing VS code in GitHub Stars, 2xing Pi-Torch, and 3x in Claude Code was not on my 2026 bingo card. Peter Steinberger might have actually built the first one-person billion-dollar company,

Starting point is 00:21:12 staffed entirely by AI agents. Now, for some after that episode with Lex, it felt kind of inevitable that at least Peter himself was going to make some big move. And that move was ultimately announced on Sunday night, when Sam Altman tweeted, Peter Steinberger is joining OpenAI to drive the next generation of personal agents. He is a genius with a lot of amazing ideas about the future of very smart agents interacting with each other to do very useful things for people. We expect this will quickly become core to our product offerings.

Starting point is 00:21:39 OpenClaw will live in a foundation as an open source project that OpenAI will continue to support. The future is going to be extremely multi-agent and it's important to us to support open source as part of that. Peter also wrote a post about the move. Called OpenClaw, OpenAI in the future, he said, TLDR, I'm joining OpenAI to work on bringing agents to everyone. OpenClaw will move to a foundation and stay open and independent. Peter wrote, The last month was a whirlwind. Never would I have expected that my playground project would create such waves. The internet got weird again, and it's been

Starting point is 00:22:08 incredibly fun to see how my work inspired so many people around the world. There's an endless array of possibilities that opened up for me, countless people trying to push me into various directions, giving me advice, asking how they can invest or what I will do. Saying it's overwhelming is an understatement. When I started exploring AI, My goal was to have fun and inspire people. And here we are, the lobster is taking over the world. My next mission is to build an agent that even my mom can use. That'll need a much broader change, a lot more thought on how to do it safely,

Starting point is 00:22:35 and access to the very latest models in research. Yes, I could totally see how OpenClaw could become a huge company. And no, it's not really exciting for me. I'm a builder at heart. I did the whole creating a company game already, poured 13 years of my life into it and learned a lot. What I want is to change the world, not build a large company, and teaming up with OpenAI is the fastest way to bring this to everyone. It's always been important to me that OpenClaw stays open source and given the freedom to flourish.

Starting point is 00:23:00 Ultimately, I felt OpenAI was the best place to continue pushing on my vision and expand its reach. The more I talked with the people there, the clearer it became that we both share the same vision. The community around OpenClaught is something magical, and Open AI has made strong commitments to enable me to dedicate my time to it and already sponsors the project. To get this into a proper structure, I'm working on making it a foundation. It will stay a place for thinkers, hackers, and people that want a way to own their data. with the goal of supporting even more models and companies. Personally, I'm super excited to join OpenAI,

Starting point is 00:23:28 be part of the frontier of AI research and development, and continue building with all of you. The claw is the law. People were not surprised, but still gobsmacked. Greg Eisenberg wrote, Peter published the Claudebot project on GitHub November 25, 2025, 82 days ago. Is this the first one-person, one-billion-dollar company?

Starting point is 00:23:47 And indeed, a lot of the speculation subsequent to the announcement was around how much Open AI had paid. That hasn't been revealed, but most people think they probably parked a truckload of money in front of Peter. The other biggest threat in the conversation has been what Andrew Hart called the crazy fumble Bionthropic. Hart writes, Claudebott became the fastest growing AI project of all time, generating nothing but goodwill. And instead of Dario reaching out like Sam and Mark did, they just pressured him with legal threats. Nader Dabbitt writes, maybe the fumble of a decade byanthropic.

Starting point is 00:24:16 The most popular and fastest growing open source project of all time is not only named after you, but most users are power users of your product. Instead of trying to collaborate or work with him, they chose violence. Others, though, weren't sure. T-Mux Vim on Twitter writes, what exactly did Anthropic miss out on? A ClaudeCode wrapper that wastes tokens like crazy? So what did Anthropic miss out on?

Starting point is 00:24:38 And why was this a good move for OpenAI? First of all, let's talk about the idea that other companies could just build this same thing. That Anthropic probably is working on something similar or that they saw it as competitive with ClaudeCode Core and Claude Co-work. History is littered with technologies that were the same or even better, but which ultimately lost out to some competitor, not because of any technology reason, but because they failed to create a shelling point. A shelling point is a focal point for a community

Starting point is 00:25:05 that comes in an uncoordinated way. In other words, everyone gravitates and circles around something without having been told or forced by anyone to do so. OpenClaub very obviously has an unbelievable shelling point. Yes, there are other ways to build agents, but right now, now the way that everyone is building agents is with OpenClaw. That absolutely does not mean it will stay that way, nor does it mean that OpenClaw is perfect or unassailable. But it has an incredible amount of community attention and focus and builder energy right now. And the technological capability set is kind of the least important part. I've been working on a follow-up to the AIDB New Year's project as our next free, fun, self-directed program for a couple weeks now,

Starting point is 00:25:44 and I had already been thinking about something for building agents. Ultimately, I decided to just focus on OpenClaw, and so over the weekend, I built clawcamp.aI. It is exactly what it sounds like. It's a free, self-directed program and community for people to learn first how to build agents and then how to build entire agent teams. And part of the reason that I decided to just focus on OpenClaw is that shelling point that I was just talking about. Because when you have hundreds of thousands of developers working on something, people don't gravitate to the project just to copy everyone else. They gravitate to that project because the community itself becomes self-reinforcing. There are literally hundreds and hundreds of resources out there available for OpenClaw

Starting point is 00:26:25 that have been spun up and written and recorded in the last couple weeks alone. Every day I find 10 new ones. The Claw Camp Resources page has dozens of articles to help people get started. And if I chose something that was an OpenClaw, there wouldn't necessarily be that much information. You even have how-to articles that are summarizing other how-to articles. Which year wrote this great one, I've had 20 plus OpenClaw articles to Opus 4-6. Here's the setup guide it built.

Starting point is 00:26:51 So OpenClaught itself is very clearly a very valuable project to be associated with. There's another point, though, specifically for OpenAI, which is that as time has gone on, it is undeniably the case that lots of its co-founders have left. Sam Altman and Greg Brockman are still there, but it doesn't have the big heroes of the developer community like Aaliyah Sutskhaver or Andre Carpathy anymore. Peter Steinberger brings in a new energy and excitement,

Starting point is 00:27:14 not just for people outside the company, but for people inside the company. And you can tell from all of the excited posts from OpenAIT members that this is an incredibly invigorating moment right as the company seemingly turns a corner where Codex is really spoken of in the same breath or for some even above Claude Code. For the first time in two years, Anthropic has a credible competitor when it comes to coding models.

Starting point is 00:27:37 And this is a powerful move on that front. There's also the obvious point that Sam makes in his own post when he says, we expect this, in other words, very smart agents interacting with each other to do very useful things for people, will quickly become core to our product offerings. OpenClaw has, for the first time in a couple of years, made the promise of agents real, and so of course OpenAI wants access to it. But what about the folks who think that this is going to be bad for the project? There are plenty of people who are making jokes that OpenClaw is now closed claw.

Starting point is 00:28:07 Rafael Shod writes, I bet this causes lots of VC tiers and angry open source software folks, and there are plenty of folks in and around open source who seem to have that opinion. But then there's also more sanguine takes like this one from Simon Smith, who writes, this is probably the best possible outcome for everyone except Anthropic. OpenClaw gets support. Peter gets relieved from killing himself trying to manage OpenClawn-Clawned to the largest chatbot user base.

Starting point is 00:28:32 Huge win for OpenAI. And also, for those who are worried about what the future of the OpenClaught project holds, how they're going to run the project outside of OpenAI is starting to come into view. investor and entrepreneur Dave Morin wrote, Been working with Peter Steinberger on the OpenClaw Foundation structure for weeks, a home for thinkers and hackers and those that want to own their data, honored to serve as the founding independent board member. This community built something extraordinary.

Starting point is 00:28:56 Our job is to protect it. Open source forever. Excited to share more soon. Now Dave is a serious dude who hasn't chosen to get personally involved in something like this for quite some time. And so I personally am excited to see how they structure the foundation and what they do with it next. Ultimately, there is no doubt that this is one of those moments for AI. Ali K. Miller wrote, I advise some of the biggest C-suite folks on AI, and I'll say this.

Starting point is 00:29:20 Agentic platforms like Claude Code, Codex, and OpenClaw are the most positive energy I've seen in those rooms since ChatCHBT-EBT launched. Enterprise adoption will be slower than solopreneurs and startups, but these leaders are quickly seeing that there's no ceiling right now. Kitsy from Tinkerer Club writes, To me, it doesn't matter who owns OpenClaught. Claudebot was a concept, a revolutionary idea. a paradigm shift. The core of it can be rebuilt in an hour on top of any harness. It inspired me to

Starting point is 00:29:46 own my AI assistant, conversations, and memory files. I also went deep into self-hosting and ownership because of it. I'm never going back to chatting with GPT or Claude no matter what. From where I'm sitting, we're now entering the exciting next chapter of AI. OpenClaw has a big stake in it, and I certainly am excited to see if Peter can live up to that mission to make an OpenClaw that would actually be useful for his mom. For now, that is going to do it. for today's AI Daily Brief. Appreciate you listening or watching as always. Until next time, peace.

The AI Daily Brief: Artificial Intelligence News and Analysis - OpenClaw Goes to OpenAI

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.