I continue to be saddened by the lack of a good #data, #machinelearning or #AI community on the #Fediverse ... the conversation just really seems to be all on the birdsite! Or am I missing something obvious? I follow all the hashtags etc, but really feels a little thin on the ground.

@AndreasThinks Keep posting (with hashtags). Keep following others with similar interests. #DataScience #MachineLearning #AI

@AndreasThinks it's a bit of a constant that people on the #fediverse innately distrust #BigTech. Unfortunately, because Big Tech monopolise the #AI market right now, that means they also tend to shun anything AI or #MachineLearning, or at least tar it with the same brush. Hopefully we'll see an upsurge of #EthicalAI in #FOSS spaces that will lead to legitimately good and open conversion about how we can use AI in ways less predatory.

@AndreasThinks my husband @Demios is a data scientist (and PhD student too!), I suspect that unfortunately the bird site is where the hype and moneyed networks are so they tend to be over there.

@AndreasThinks Sadly, people aren’t any more willing to think on mastodon than they are anywhere else.

@Darohan @AndreasThinks You can talk to me about breaking up monopolies or the incredible ressource, energy and water consumption that "AI" is going to bring. I am also interested in "AI" use cases that justify this cost.

I think, the community here is a bit more aware of all the downsides of the current hype.

@kaffeeringe @AndreasThinks those concerns are absolutely valid, and I share them, but I feel like they are symptoms of Big Tech's incredibly exploitative "tech for tech's sake" approach, rather than issues innate to AI itself. There are legitimate uses for AI and machine learning being found throughout the medical, resources, and advanced research fields that I'd say are pretty valid?

@AndreasThinks I suspect this is a reflection of the specialized nature of those topics–they’re a bit unapproachable to amateurs–which means the population of potential people to participate in those conversations is a bit small.

And it just happens to not have much overlap with the Fediverse userbase, as workers in the field also just might not have that much free time to dip their toes here and sort through all of the other topics that do get more airtime here.

Maybe ironically, better AI would help make Fediverse feeds more useful to people like them, as the chronological sort is unwieldy.

@Darohan @AndreasThinks Yes. It produced the prototype of the Covid vaccine within a day! That is amazing.

But that isn't what Big Tech intends to do with all these ressources.

Sam Altman just admitted that "we" need way more energy than "we" planned. He says, we need nuclear power and fusion for that. Fusion! The first experimental reactor is supposed to produce the first kilowatts in 2035. Than they build a demo reactor...

And in the meantime?

https://www.bloomberg.com/news/videos/2024-01-16/openai-s-atlman-and-makanju-on-global-implications-of-ai

@Darohan @AndreasThinks And for what? So people can generate mails with CoPilot that the recipients summarize via CoPilot. Or you can easier find something on your smartphone.

@AndreasThinks I used to post about new machine learning papers daily but as you mention, there isn’t a lot of conversation over here these days.

Plus, the growing tar and feather everything remotely connected to “AI” (and I use the term loosely since I don’t like using that term since what we have now isn’t AI) sentiment doesn’t really lead to open conversation …

So I’ve just stopped talking about the stuff I’m working on. But there are at least some people on here who still talk about ML related stuff on and off ….

Depending on what you are interested in, whether it’s practical implementations, new advances, a particular field etc. I might be able to point you to some …