Tech keeps pushing AI into use well before it's ready. It's not just the privacy issue, **the chatbot doesn't work**, and it's a not a replacement for the thousands of staff that Musk and his tech bros have pushed out of the Social Security Administration. - ThreadSky

pamherd.bsky.social • 1 day ago

Tech keeps pushing AI into use well before it's ready. It's not just the privacy issue, **the chatbot doesn't work**, and it's a not a replacement for the thousands of staff that Musk and his tech bros have pushed out of the Social Security Administration.

Comments

equineducklings.bsky.social•1 day ago

They want to break Social Security. It’s part of their plan to undo the New Deal

dlitt.bsky.social•1 day ago

There is probably some limited use case for AI. Still, I'm mostly convinced it's just high-tech tulipomania; an excuse to commodify pointless products in a fever dream of valorizing capital. Walking on the clouds like Aristophenes or Wile E. Coyote before he looks down.

raininberkeley.bsky.social•1 day ago

I absolutely agree with you, but having said that, ChatGPT gives me more thorough info than my doctor or my nutritionist!

pamherd.bsky.social•1 day ago

The question, though, is whether the information it's giving you is correct! In the SSA case, it's expert staff who are finding that the chatbot is providing inaccurate information.

raininberkeley.bsky.social•1 day ago

Yeah I absolutely hear that! Definitely the info I get is just for personal use, and I take it with a grain of salt. Like I said, I agree with your original statement.

pamherd.bsky.social•1 day ago

Yeah, and I can see places where LLM's work, but I'm coming more into line with the view that we've rolled it out well before it's ready for prime-time--and the costs are huge, both environmental and the chaos it can create when it's wrong.

emilytav.bsky.social•1 day ago

💯💯💯 I wrote a piece about this last year… why is this so hard to understand 🤦🏻‍♀️🫠

https://techpolicy.press/sure-no-one-knows-what-happens-next-but-past-is-prologue-when-it-comes-to-ai

pamherd.bsky.social•1 day ago

This plays out in medicine too. The use of polygenic risk scores for embryo selection is analogous.

calixtus.bsky.social•1 day ago

The really weird part is that in most large organizations, the information has already been uploaded to internal file servers, wikis, sharepoints, employee blogs, etc.

A tool that slurps in all of that data and spits out transmogrified phrases is much less useful than a good internal search engine.

lowerkey.bsky.social•1 day ago

So, if AI is easily manipulated and constantly fails now, why is the government 🤡💩 pushing it on primary education?🤔

Oh, I get it! 🤨

It's easier to brainwash people and give them "alternate" facts. 🤬

jpaulemersonjr.bsky.social•1 day ago

Ask your AI if there are mountains of gold and silver on Mars.
This is why AI can't generate credibility.

Comments

Posting Rules

Reply