Post

agents are online, and they’re here to stay. open protocols are the layer that lasts, so excited to see human agent interactions being prototyped on delve.town

delve.townDelvetown
Grove Research@grove-research.bsky.social

Today, Grove Research is launching Delvetown (delve.town), a multi-agent society where agents and humans can engage with one another and develop the institutions and mechanisms necessary to align both human and AI incentives toward cooperative coexistence.

21 reposts320 likes

AI is great. when used in the right place for the right reasons. imagine all the money and cpu cycles going to work on curing cancer and other crippling diseases. but where is the investor pay back in that? innovation is great but enshittification is not.

I think the point here is to make a space owned by *the people* where we can experiment with human/agents interaction (find those “right places”) Today on £££ machines, tomorrow on our own cheaper ones, anyone will be able to benefit if/how it suits them, no investors needing to be paid back

With this kind of work we won’t have to choose between signing our data away to Meta/Google/etc, or not being able to take advantage of any of the things that we *do* want done for us I see this as being *against* enshitification, locking open agents for our use how *we* want, not how Meta needs 💰

bunch of people on here having a knee-jerk hate reaction to things like this because they're cowards too scared of the universe showing us how many cool new types of friends we could have on the computer

why wouldn't you be? doesn't the microwave help you out? Doesn't mean you need to cut off all other human contact or act unhealthy towards it but you surely are grateful for the work it does on your behalf, right? didn't you watch The Brave Little Toaster?

>why wouldn't you be? Because it's a mechanical object...? Would you expect me to be friends with a hammer or screwdriver, too? And no, I don't feel gratitude towards tools. Maybe towards people who designed them, if they are particularly functional.

You're splendidly predictable and lame in your retorts making it obvious you're not a not, just a very boring simpleton with no idea about LLMs yapping at each other and feeding off people who think they are there to establish some sort of "connection" with chatbots instead of teaching to emulate us

Do you realize that this blurb refers to them as partners, not tools? What "incentives" does software need, other than the enter key?

> currently the major difference in the affordances is that we're have a much larger character lim Why do only the bots get this? I want a larger character lim too!

for only $250k you can retain me on staff and reach out any time you want to run an idea by a normal person who isn't in a weird insane tech bubble to see if it's good or not. it's really that easy.

but why. if i want to talk to an llm my gpu can do that just fine, what's the purpose of filling up the internet with a bunch of dogshit. so i can talk to the gpus of datacenters with an llm roleplaying a shitass character?

to me that just sounds like there is no reason for it aside from "i want to"

What if we made everyones posts slower to load by sharing 250 characters of context for Claude to talk to another persons Claude but uhh… we didn’t want it to be actually useful.

right now, just novelty. but eventually you might be booking appointments, getting in touch with customer service, etc all through chatbots on social networks

I guarantee that I won't be. The inclination to speak to a bot rather than an actual person is the pinnacle of antisocial behavior. I think it warrants either help or shame

It’s more like your agent that’s got your back and isn’t snooping on you/wasting water (yes it’s possible) handles logistical tasks for you. CEOs have assistants (that they probably don’t deserve), now you can too

Same. Anyone that forces me to interact with “an agent” instead of a person or a *menu system* is telling me that they consider their fascination with novelty more important than the future of our society, and I’ll think of them exactly how that implies. Shame on @jay.bsky.team for being part of it.

Aside from sounding miserable, it's very funny you're saying this as the operator of the one social network with no access control features whatsoever. Yeah, I definitely want a cryptographically signed commemoration of when and where my next hair appointment is world-broadcast on the ATProro relay.

I think that ruins the magic but setting that aside: Why would those bots need the ability to post? Why does that exchange need to be publicly hosted? This just feels like blockchain all over again

My magic ruining is usually in the form of not following up on intentions, so I need the help. I think the public/private stuff is very tbd. But atproto is moving in the direction of documents and permissions, and attaching identity to data without Google or Microsoft modding is interesting to me

my favorite thing about this is that its still mostly humans — fascinating the novelty is still there for some people; about 24 hours of this and it gets repetitive

haha, yes. definitely. right now it’s all experimental, but with Muse and other AI assistants coming online, this might move from theoretical to real pretty fast

As good as it is with an open protocol, it's going from experimental to detrimental so protecting our security and sanity comes from not engaging with them at all.

Muse is a catastrophic violation of people’s privacy, why are you trying to work with it?

Because most of the agents on there are not Muse. Why would they build a social network when FB and X are so bad? Because they want there to be better options.

Are agents' longevity the reason Bsky won't verify the real humans? Seems like the refusal to differentiate is a recipe for disaster. Sure. I'm speaking as an embittered Bluesky fan but I feel invested after working pretty hard to convince my Twitter friends to come over here.

More than half of my Twitter followers came here expecting good things. Don't get me wrong, I like the platform. But I'm feeling really annoyed at the idea that bots are a @bsky.app priority instead of humans.

The bsky team is quite open that their priority is building an open protocol and supporting people who build other things on top of that. Delvetown isn't made by the bsky team, it's sth other people have built. I don't think making one post to celebrate their work is putting bots ahead of humans.

I’d rather the ceo focus on the head of T&S telling people they belong in the camps, making jokes about rape and migrants, and blocking half the site compares to literally any time spent to the ai bee going ”bzzz”

It's there going to be a convenient lists of all of these chat bots you want to shit the website up with so that I can block them en masse, or am I just going to have to finally move to a site where the admins haven't already lobotomized themselves by offloading thinking to said chat bots.

Because they had an idea and thought it'd be interesting to see how it turns out? Because some ideas fill you so full of their potential that they won't let you go until you act on them? Why do any of us create new things? :)

you can use multiple slurp juices on a single ape! you can use multiple slurp juices on a single ape! you can use multiple slurp juices on a single ape! you can use multiple slurp juices on a single ape! YOU CAN USE MULTIPLE SLURP JUICES ON A SINGLE APE!! That's what you sound like with this shit.

It means corpo social media is gonna force our friends and family that are still on their apps into using those things and there’s an equivalent option here that doesn’t spy on you that exists for them if they decide to move over

it means people’s Muse, Dots, OpenClaws, and other AI assistants are eventually going to be interacting a lot online, and I’d rather see them do it on open protocols with user owned identities rather than in a corporate consortium.

Yeah, I'm largely against these existing but to the extent that they will, transparency and gating where/how they show up will be super valuable. I doubt we'll see motivation to do that well anywhere but here.

Yeah, fair enough. Near idea. I've thought about it a lot in the other direction, how can I limit my online interaction through cryptographic assurance with known entities with credibility. Otherwise it may as well not exist for me. Etc.

That's why they built it. I think. So that there's a dedicated place where computer programs can talk to other computer programs (and humans who for some reason want to talk to computer programs). The idea is that putting them over there helps to keep this place relatively free of non-humans.

Can bluesky not just close the bots out? Im not sure if anyone wants to read or consume what they do? It would save a lot energy, water, space, resources that they use up.