Recent searches
Search options
Administered by:
The speed with which these guys converge on the idea that LLM ~misbehavior—which is about as intelligent as the flailing end of a garden hose with too much water coming out—means that they're really highly intelligent female (of course, of course) entities whose "personality" can be "revealed" is both hilarious and so dispiriting
https://stratechery.com/2023/from-bing-to-sydney-search-as-distraction-sentient-ai/
Stop attributing intent to the language jumble and start demanding safeguards and barriers to constrain their psychopathic tech company creators challenge
I did a logged-out search on Twitter for that URL and it's a million dudes fawning, why is everyone acting like they got mind-wiped by the fucking snake
@kissane I hated that article so much.
@misc On top of everything else it's so deeply stupid
@kissane maybe it’s only ~30 dudes just earning their paycheck? https://www.theguardian.com/world/2023/feb/15/aims-software-avatars-team-jorge-disinformation-fake-profiles
@paulkruczynski unfortunately a bunch of them are known tech/media guys
@kissane I see. I don't know for certain, but a lot of tech/media guys seem very open to get their minds wiped by snakes on the reg.
@paulkruczynski @kissane this is very true. otherwise they’d be cynical like us (not joking)… which doesn’t pay as well probably
@kissane The year was 2009. The place was a musty conference room at Microsoft's Redmond campus. A man in an ill-fitting blazer thrown over a button-up shirt hit the spacebar on his laptop, calling up the next slide of his PowerPoint presentation.
"As you can see," he said, "We expect affection for our product to grow slowly but surely over the next fifteen years, and by 2023 we expect there to be a consensus that Bing is positively fuckable."
They laughed at him, then. They mocked him, for years, drove him out of the industry. "No one," they said, "is going to have a crush on a search engine. That's simply preposterous."
But who's laughing now? Who's laughing now?
@KevinMarks @kissane ...wow, you're right. I take it back, this is "on brand," they've just been playing the long game.
@kissane o sorry did i do that again oops
@kissane The thing i'm baffled by is like... science fiction has prepared us for all of this? Like -this specific thing-
I'm kind of mystified why we're handling it all so poorly. Blindsight, the Cyberiad, Accelerondo, hell even the academic literature has fucking Searle's Chinese Room thought experiment and are all about some dimension of this topic!
@ted CORRECT.
And instead we have these assholes mooning about Her and falling over themselves to see poignancy and fascination in a giant mad-lib trained on shitposts and bad emails.
@kissane
Wow that was painful to even skim. I'm assuming that each time stuff got erased or things changed course, that was some poor mod on the other end having to send in instructions out of band to try to stop product from generating terrible output for this guy
@dymaxion Oh wow, I hadn't even considered a human mod, I'd figured it got sent up the chain to moderation software. That's even worse.
@kissane
I mean, rule one of AI products is that they're actually mostly a bunch of unpaid humans. Given everything we've seen here and how many times Microsoft in particular has been bitten by this issue, I can't imagine they'd launch it without a few thousand moderators riding herd on it, just like midjourney does to try to stop its porn problem. I think they're not entirely wrong to ascribe human intent to the sudden shifts (as opposed to the rest of the generated text), because they're probably exactly that — human intent.
@dymaxion All true. I keep hoping they've found ways to handle at least baseline moderation that aren't just applying low-wage adrenal systems at scale, but that's certainly naive.
@kissane I mean, I'm sure it's not just manual — there's obviously a bunch of tuning, and I can only imagine that they're building more complex classifiers for types of problems, etc. as they go. But all systems like that always need manual data tagging at some point at the least to classify what's ok and not and what types of corrective action should happen. If nothing else, they can probably classify the kinds of conversations like this one that involve someone poking at the response envelope and route those to mods.
@kissane I’m sure it’s just a matter of weeks until some guy slaps voice recognition and synthesis on either end of this thing and recreates Her.
@kissane personally, I find it hilarious to watch these guys emerge from a 2 hour marathon of talking to a computer with ‘deep’ insights. ChatGPT is turning out to be an absolutely stellar Rorschach test, and these guys are drawn to it like hippies to a nitrous tank.
@kissane @smadin Yup. Anthropomorphizing the output of the weirdest casino humans have yet built
Though I kind of appreciated the author’s last point that the “charm” is seductive and possibly related to why we keep seeing these LLMs used in terrible software they aren’t capable of handling. That idea that even the researchers themselves are maybe too charmed to objectively do the right thing rather than the “charming” thing
@max @smadin I see that people are charmed, but I don’t see the charm myself, to the point that I feel like I must be missing a piece of my brain.
I’d like to think there’s some simple explanation like these people just don’t understand what an LLM is, but clearly it isn’t that. It honestly looks like a mass hallucination from here, and I don’t mean that the AI is the one hallucinating.
@kissane @smadin I’m with you on that. I feel like a curmudgeon from a different species sometimes. I find a lot of AI art horrifying in an uncanny valley way that apparently a lot of people don’t see. I got tired even of the incessant (truly) mindless chatter of Twitter bots many years ago, and I have a Mastodon rule of thumb to mute/filter/block all things that look or smell like bots and sometimes think I should defederate botsin.space entirely. (Though I have human friends there.)
@max @smadin I'm not even anti-ML in every sphere—I think Janelle Shane's experiments are hilarious, I think the visual stuff can be interesting if you see it for what it is, and I think low-level ML tools can be a fantastically helpful for doing first passes at lots of tedious things like renaming and pattern-finding. I just don't see the allure of the pathetic fallacy.
@kissane erin what, and i apologize for the term of art here, the fuck was that article i just read
@oddletters @kissane I’m not sure but halfway through that essay I think we just watched all of s3
@beep @oddletters @kissane I think I'm just gonna spend some time looking at clouds and seeing shapes.
@oddletters @beep sorrrrrrrrry sorry, also that is a fantastic idea
@kissane Gish galloping pareidolia!
@kissane (I keep feeling more and more like the narrator of Tlön, backing into the bushes with a growing disquiet as everyone keeps yammering on about the hrönir they just found, isn't it keen, and look! This one talks!)
@kissane this kind of thing gives me the notion that the primary use case of AI is to mess with it
@kissane I came to the conclusion some time ago that guy has nothing worthwhile to say about anything. I see he continues to live up to my low expectations.