The Lobster
Greetings and crustaceans from 'ClawCon', as relayed by Hayden Field for The Verge:
The woman at the door wore a plush lobster headdress.
She sat in the front hallway of a multistory event venue in Manhattan, beside a bundle of wristbands. If she granted you one, the world of ClawCon beckoned behind her — full of vibey pink and purple lighting, lobster claw headbands, multicolored name tags, sponsor information stations, and a demo stage underneath a skylight. Hundreds of people were gathered to celebrate OpenClaw, the AI assistant platform created by Peter Steinberger in November 2025.
I always enjoy these boots-on-the-ground reports from such events. They tend to overemphasize the ridiculousness of them for obvious reasons – and that seems especially easy and apt when the movement has both a mascot and when that mascot is as ridiculous as a lobster.
OpenClaw (previously known as Clawdbot and Moltbolt) has quickly become popular in the tech industry for being open-source, in contrast with AI agent services from big labs like Google, OpenAI, and others. Practically, it’s still an unpredictable tool that can pose major security risks. But this community sees it as a grassroots crusade and a noble pursuit, offering an escape hatch from an industry controlled by a handful of people at leading AI companies.
Reading this report, I'm thrown back in time to a few different eras of tech. In recent times there was all the silliness around "web3" – and the brief "NFT" craze was perhaps most symbolic of the situation. It undoubtedly started out as earnest and innocent enough, but quickly – as expected – it was overtaken by a lot of grifters trying to make quick bucks. And the whole thing sank under the weight of the various schemes as rugs were pulled.
This OpenClaw movement certainly seems far more legitimate, and it feels like an interesting and in some ways natural offshoot of the broader AI revolution. Some people want to be excited about AI (to be clear, others absolutely do not want to be excited about AI, which is also an interesting dynamic at play) but are wary of Big Tech (and "Big AI" – the OpenAIs, Anthropics, and xAIs valued in the hundreds of billions of dollars as "startups") controlling the situation and potentially the entire future. OpenClaw has slipped right in as an ointment for that ailment – to the point where it has even survived those comical branding shifts.
Still, I'm reminded more of an earlier era, when social networks were rising in prominence and power. Once Facebook overtook MySpace, in came OpenSocial to try to combat their inevitability.
Granted, that project arose out of Google itself. And the cynical read at the time – including my own – was that they were backing it mainly to ensure that Facebook didn't overtake them as the main character of the internet thus hurting their rising dominance over advertising. Still, there were many "true believers" in the movement to try to combat Facebook and ensure that social networking wasn't locked down and in to one company. And the fact that Peter Steinberger has now gone to work at OpenAI (with the promise that OpenClaw itself would remain open source and supported) brings the parallels a bit closer.
All of that is to say: as was the case with OpenSocial, I'm skeptical that this movement lasts and gets any real legs beyond these sort of fun, quirky stories.
This sentiment will make people angry, because such things tend to almost be more religious in nature. And I know and like a lot of the people involved in this particular movement. And I even like the overall sentiment here, about putting the future of AI in the hands of the people – I mean, how could you not? (Well, actually a number of players would undoubtedly consider this dangerous, and certainly we've already seen some of the dangers made possible by a tool like OpenClaw.) I'm just trying to give an honest read and prediction for how this all plays out.
On paper, people love "open". It naturally sounds "winning" to the masses. Power to the people and not the corporations, and all that. But in reality, we see open get pummeled by closed, perfected systems time and time again. Not always, but often. And especially, when the stakes and the scale are at AI levels.
People tend to fall in love with great movements but end up marrying great products. And great products tend to rise out of companies with a greater control over the technology. The "open" stuff built by a community may start strong but often degrades into a mess of contradictory ideas and simply slow progress. Open also has a funny way of coming back to bite time and time again. There's a reason why Meta is turning their back on "open". And why "OpenAI" is anything but...1
Perhaps it will be different with AI going forward because there are so many good models created by so many different players – some of which are open and some of which are closed. It does feel like there's a window here for others who can aggregate and orchestrate on top of these models right now, which is something we're seeing at the moment even beyond OpenClaw. But such things also tend to shift over time and we may end up with just a few "winning" models as the costs become untenable – especially if and when there's some sort of downturn in the market.
Obviously, it could shift the other way if the "open" models themselves win out and proliferate. But we've also been hearing that for a few years at this point. And despite DeepSeek being a "moment", it hardly ended up being one of the "Sputnik" variety. We'll see if their new, impending model changes any equations. But my bet would be that a year from now, pretty much everyone is still using ChatGPT, Gemini, and Claude. Because again, those are great products.
And I also imagine that each of those "Big" players has their own answers for some of the elements of OpenClaw that prove popular. We're already seeing it with Anthropic – perhaps because they fumbled the ball a bit with the artist formerly known as 'Clawdbot', which was called that, and had to change that name, for a reason... While the OpenClaw movement is fun, headlines aside, the masses are simply not going to be rushing out to buy Mac minis on which to run their own instances. Just like 99.99999% of people don't run their own web servers even though they technically could.
And again, most people should not be running their own OpenClaw bots, as it's a recipe for various security disasters. That mixed with the aforementioned tendency for people to gravitate towards good, easy-to-use products makes the end game fairly obvious here.
And it's not a social network for OpenClaw bots.
The only wild card may be the aforementioned feelings about AI in general. Again, more so than perhaps any other technology in the past – aside from maybe Google Glass – there's seemingly a real natural hesitancy here. Certainly in the US market, which is obviously at the forefront of the technology and at the same time, nearly every story you read about AI is about the fear over job displacement...
But I still have a hard time thinking that cohort of people is going to gravitate towards OpenClaw as the answer. In a way, we may be seeing a shift towards Claude itself because of the blow up between the government and Anthropic. As a result of that, people may be starting to see Anthropic as the more "ethical" solution in AI, which may help some over that initial hump of skepticism...
As we go further down the path of this AI revolution, I continue to wonder if and when we're all going to settle down with our one true AI. It's possible we end up in a world where we use one for work and one for play, as it were, but given that the aforementioned big players are increasingly trying to do it all, I think picking one seems more likely.
Though if the ClawCon attendees are forced into such relationships, perhaps they'll revolt and instead demand to be turned into an animal of their choosing. Just like the plot for a certain absurdist movie...
M.G. Siegler
M.G. Siegler
M.G. Siegler
1 Yes, yes, they still sort of have some efforts. ↩