Re: ActivityPub and AI bots?

On Mon, Feb 3, 2025 at 9:38 PM Aaron Gray <aaronngray@gmail.com> wrote:

>
>
Again for architectural level programming it really requires some level of
> experience or experiments over time to come to solve real world problems in
> a manner where they will test the tests of time in a real world
> environment, rather than just applying oversimplified patterns as it
> tends to do.
>
> For non programming tasks such as sorting through toot's for sense, or
> SPAM, or whether it is another machine's output, that is quite an
> interesting open question.
>

Here too, I tend to agree that context-rich and semantic, rather than
one-size-fits-all, token-based, approaches to identifying spam, inauthentic
activity, bot farms, etc would be table stakes for applying AI to
pre-moderation processing. One takeaway I've seen about the DeepSeek
release is that it might really be way less labor-intensive and
capital-intensive than people realized to build in coarse-grained
context-sensitivity and expose it to the user to create the kind of
last-mile human intervention/monitoring that you describe. (Even just the
UX of DeepSeek is instructive in this regard! A self-doubting, multi-answer
chatbot is way more useful than an overconfident, pizza-gluing one). It's
early days, though, and what DeepSeek actually spent in human hours is
probably still orders of magniture beyond what the Fediverse could cobble
together quickly, but it's at least directionally promising?  I'm not
opposed to language models in general (and I'll note there was an amazing
talk about algorithmic search/network analysis at the SocialWeb track at
FOSDEM[^1]), but I can definitely sympathize with a position I heard at
FOSDEM in many AI contexts, that it's healthy to wait for economics (and
ergonomics and economies of scale) to proceed a lot further before anyone
starts bolting on today's talking paperclips to critical infrastructure...

thanks for the MCP spec, that's pretty interesting! Any idea what people
are using for the "robust consent and authorization flows" mentioned in the
Implementation Guidance section?

Thanks,
__bf

>
>>>
>>> On 2025/02/02 2:56, Bob Wyman wrote:
>>> > Has anyone yet used the API's of Llama, or something else, to link
>>> > ActivityPub to AI? Is there a bot that can receive and answer user
>>> > queries via ActivityPub?
>>>
>>
> If this is going to happen as it will I think its worth doing properly and
> efficiently, and we should attempt to get ahead of the curve so we know
> what we are facing.
>
> Yes this does deserve our attention !
>
> Aaron
>
> >
>>> > bob wyman
>>> >
>>>
>>> --
>>> Daiki "tesaguri" Mizukami
>>> <mailto:tesaguriguma@gmail.com>
>>> OpenPGP: <openpgp4fpr:f706f9bbb8f63cde5f40298fdf5eb02be5c2542e>
>>> Fediverse: @tesaguri@fedibird.com <https://fedibird.com/@tesaguri>
>>> Matrix: @tesaguri:matrix.org <https://matrix.to/#/@tesaguri:matrix.org>
>>> GitHub: <https://github.com/tesaguri>
>>> Keybase: <https://keybase.io/tesaguri>
>>>
>>>
>>>
>
> --
> Aaron Gray - @AaronNGray@fosstodon.org | @aaronngray@threads.net |
> @AaronNGray@Twitter.com
>
> Independent Open Source Software Engineer, Computer Language Researcher
> and Designer, Amateur Type Theorist, Amateur Computer Scientist,
> Environmentalist and Climate Science Researcher and Disseminator.
>
>