LLM defenders are always like, "Everyone uses it! There's loads of misinformation about AI!" Okay, then how come whenever I talk about how shit it is, I get almost universal agreement, and every time a company uses genAI to make something, the audience rebels against it en masse. - ThreadSky

smolrobots.bsky.social • 5 days ago

LLM defenders are always like, "Everyone uses it! There's loads of misinformation about AI!" Okay, then how come whenever I talk about how shit it is, I get almost universal agreement, and every time a company uses genAI to make something, the audience rebels against it en masse.

Comments

spite-fiend.bsky.social•5 days ago

My attention span is already destroyed enough by adhd, social media algorithms, and short form content. I don't need to get rid of what little capability I have left by using the slop machines

cursed.monster•5 days ago

But, but, but, the emperor’s raiment is so majestic!

smolrobots.bsky.social•5 days ago

The issue, as I've expressed in two threads today, isn't so much the evangelists eagerly lapping up their slop, it's the people - often young people - who are utilising this technology before they've even understood what it is and how it works.

smolrobots.bsky.social•5 days ago

It's people who are instinctively reaching for a text generator rather than maintaining their own communication skills. Like, if you can handle it, that's fine I guess. If you go into it with eyes open and gobble up the slop and only use it for slop-appropriate things, then I guess that's okay.

smolrobots.bsky.social•5 days ago

I still think you shouldn't go near it, even to see what it does or whatever, but at least you're a consenting adult. It's when people are just lazily and ignorantly using it for every text-based communication need they have in their lives that it scares the shit out of me.

smolrobots.bsky.social•5 days ago

We are going to have a massive, massive problem in a few years' time with an entire cohort who are simply incapable of thinking coherently about anything. This isn't being addicted to a smartphone or being trained by algorithms: it's a much deeper, more disturbing problem than that.

smolrobots.bsky.social•5 days ago

It is intensely horrifying to me that, whenever I open a PDF - even if it's one *I wrote* - Adobe is automatically prompting me to use its AI to summarise the content because it's a "long document". That doesn't feel like convenience - it feels like laziness.

smolrobots.bsky.social•5 days ago

I hate the idea that we (or they) are building in tools to software that, by default, go "Oh man, this looks *hard*. You wanna just read the CliffsNotes? Enough to get the gist? That'll do, right?"

kitanin.bsky.social•5 days ago

Plus, of course, they added a "helpful" widget in the corner of the window that pops up a bunch of AI assistant stuff that... drove me to switch PDF readers, actually.

I suppose that _could_ be called a benefit. Just not to them. 😏

crankyotter.bsky.social•5 days ago

More like more training morsels for the hive.

stealthisbook.bsky.social•5 days ago

The only use I've found for an LLM that isn't more trouble than it's worth is in generating a content free, plausibly constructed, prose shaped object to fill space.
But then back in the 90s Aldus Pagemaker would happily fill any text block you wanted with lorem ipsem, so whatever

Comments

Posting Rules

Reply