136 comments

[ 2.9 ms ] story [ 60.5 ms ] thread
After LeCun (actual ML pioneer) left Zuck, then his data-labeling expert Wang, now he reaches for the hype around Molt/Claw, just like openAi did with their molt/claw "purchase". Given Zuck's track record on LLMs, I do not hold out for actual science but expect more smoke&mirror commercialisation tricks - or even the integration of his dystopian camera goggles.
> Facebook parent says Moltbook gives autonomous AI a way to verifiably connect.

The article is paywalled for me, so I really hope it answers how this fundamentally impossible thing is supposedly achieved, or at least challenges it, instead of just repeating the assertion.

Is the market so bad that non-exec-level new hires are making the news?
The metaverse: ai talking to each other over cli
(comment deleted)
I didn't realize Moltbook and OpenClaw - were created by different people.
This surely won't lead to anything bad or reckless at all.
WHy are we just posting paid context? and the worst viral product since bop-it?
Moltbook might have been a fad. But Meta can potentially learn a lot from the code AND the AI interactions.
It is a not-that-obscure secret that most posts on Moltbook, particularly the "Viral™" ones, are written by a human.

Does Mark not know this?

I know there's a big advantage in capturing the market early, but in this case Moltbook hasn't captured any of it ...

Weird. With Meta's backing it is going to be successful anyway, but this is something they could have developed in-house in like a weekend.

you miss out on openclaw so you spend the money on the next best thing. bold strategy cotton, lets see how it plays out for them
facebook was lagging on the bot:human user ratio and they needed to scale the left side of the equation to really improve their je ne sais quoi
Moltbook, Facebook, hmmm. Seems like a good match; at least one of them has a good amount of feed activity.
This is incredibly bearish.
I thought that Moltbook was sort of a joke because it was people LARPing as agents as much as it was agents, and given that, I'm confused by this:

> "The Moltbook team has given agents a way to verify their identity and connect with one another on their human's behalf," Shah says. "This establishes a registry where agents are verified and tethered to human owners."

So the impetus for the acquisition was either the verification technology or to hire someone who has worked on verifying agent identity.

Does anyone know what exactly Moltbook's technology is, the technology being described by Meta? I can't find anything on the website related to this. The only "verification" they seem to have is an OAuth connection with Twitter.

edit: I guess it's this https://xcancel.com/moltbook/status/2023893930182685183

Moltbook had REST Api Endpoints to post, you could or can just directly post what you want.

Almost everything viral on there was either directly written by a human or instructed by a human.

Agents didn’t even write posts on heartbeat.

I was not at all imorepressed by what I have seen so far on Moltbook. It's like 90% straight up spam trying to get you to buy crypto.
It being mostly humans makes it more valuable to Meta, that means they can sell ads easier! (the advertising to AIs market isn't quite there yet)
It's probably something vibe-coded, and nobody is checking if it works or not, just like the rest of the site. They would have just asked another AI if it would work or not.
> tethered to human owners

2026 tamagotchi

I honestly absolutely don't understand purpose of this thing. Ok so I can bypass their captcha by literally calling any other AI. Does meta even bother to look on things on which they are burning money?
In my day we used to call registries "databases."

The secret sauce is that they built a centralized database and assigned hash ids to registered agents.

This is apparently worth a lot of money now that executives have offloaded their common sense.

That challenge was pretty stupid. I could read the question and I’m not even a native speaker. We can of course easily come up with much better challenges
Lol, Facebook is full of AI bots pretending to be humans, while Moltbook is full of humans pretending to be AI.
Can't wait to see the equivalent of captcha but for LLMs, to keep those humans away
I thought the whole thing was a prank since it was so obviously fake.
> The deal will bring Moltbook co-founders Matt Schlicht and Ben Parr into Meta Superintelligence Labs, the unit led by former Scale AI CEO Alexandr Wang.

Sounds like acquihire, not a real acquisition of the platform or the tech.

(comment deleted)
I'm beginning to think that the problem of 'late capitalism' is quite related to the ability of companies to acquire other companies.

Thereby eating their competition, either by stifling upcoming competitors or to gain degrees of monopoly power by joining with peers.

What would the world look like if you you simply could not do that?

Governments did used to go some way to stopping companies acquiring other companies and would even split monopolies up. But they all just kinda stopped doing that in more recent years.
A platform where bots-pretend-to-be-humans and another where humans-pretend-to-be-bots. A match made in heaven!