its.stonehippo.com
Sometimes I take pictures of birds. He/him.
https://stonehippo.com
1,534 posts
218 followers
377 following
Prolific Poster
Conversation Starter
comment in response to
post
That's intentional. It's a feature to make it easier to avoid being anywhere near someone wearing these things.
comment in response to
post
The railway mania a good parallel, socially, but perhaps not technologically. The utility of railways was more-or-less obvious, and they didn't typically decide in the middle of moving people and freight to make up stops that weren't on the line. 😉
comment in response to
post
Yes, saying "this tech will disappear" isn't exactly right. But it's a natural response to the polar opposite we've subjected to for the past several years, "this tech will disappear everything else".
comment in response to
post
Now, being cheap and maybe-kinda useful in narrow ways is not a bad place to be. But it also doesn't warrant massive investment or deserve a lot of hype.
comment in response to
post
Could there be a new S-curve change ahead in the near future? Maybe, but the current research is making it clearer that's not going to come from the current approach. So these systems aren't going to get much better, even if they get cheaper.
comment in response to
post
Now, you can go a long way with a 1 in 4 win rate; it's arguable that's about where IT has been for several decades. But "AI" not moving that needle much, despite $$$ spent, is not a predictor of big change.
comment in response to
post
The utility at scale is not there. Do some people find the tech valuable? Yes. Is it proving to be the transformative engine it's hyped to be. So far, that seems to be a no. IBM recently released research that 75% of "AI" projects are failing to return value.
comment in response to
post
Yes, the tech behind LLMs and foundation models is here to stay. But its future as "the thing that's going to change everyone's lives and make a trillion dollars" is just as doubtful as AR/VR, IoT, and blockchain turned out to be.
comment in response to
post
I have Hulk-level rage about the design choices on top of LLMs. Just the worst choices to help people understand what these systems are actually doing and why they're not doing what might think they are.
comment in response to
post
Yep. It's such a poor design that it can't be tested well enough to find all of the issues to fix the next iteration.
Doesn't help that today's explosion also seems to have fatally damaged the next iteration test booster, too.
comment in response to
post
I have fantasized about having a drone like this while stuck in a traffic jam, because I have this urge to know what the hold up is. I fully recognize that is a useless impulse and that this is a terrible idea in reality.
comment in response to
post
Booooooooo
comment in response to
post
In brief, this is the sort of support we need to keep digital junk in line.
comment in response to
post
YEP
comment in response to
post
It's also odd to justify your unsustainable thing by comparing it to another unsustainable practice, unless your goal is to get us to chuck both in the bin.
comment in response to
post
I'm with you on theory of the mind in general. It's the Grand Unified Theory level of things that I think is problematic. We've built a stunning amount of understanding, but assuming we've almost solved the puzzle is a persistent mistake.
comment in response to
post
The false equivalence is deeply annoying and the worst kind of "um, actually" BS.
comment in response to
post
Thanks for this. It's a good reminder that theories of the mind, particularly those that try to get at how we think through how we use language, are often limited and limiting.
Now I need to go read that Quanta piece, too.
comment in response to
post
It doesn't help that their definition of "smart" isn't particularly strong, either.