"It's our lord and savior" (mostly AI companies pushing this)
"It will kill us all" doomerism (it won't, it has no intelligence or sentience or agenda)
"It has some use cases!" centrism (problematic because it enables the lord and savior types & it's also wrong as work products are highly unreliable, even for basic things like summaries and tabulation)
"It's a bullshit tech hype, paired with theft and environmental waste" (correct)
@thomasfuchs I think some of the comments here are missing the distinction between "generative AI" and Neural Networks. Neural networks are great, useful tools with many applications. Generative AI is an unpredictable, expensive mess built on top of Neural Networks, and I can't see the benefits ever outweighing the harms.
I think an awful lot of BS jobs, like call centers, are at risk from BS AI. Yes, the error rate would go up, but the costs would plummet. Making call centers less useful might be part of the plan...
AI's main threat is to jobs. Automated driving isn't generative AI, but will, eventually, kill nearly all the 10 million or so professional driving jobs in the US and some huge number of jobs elsewhere.
Some sort of argument would go a long way to helping me understand why.
Again, take customer support.
This is an area where 90% of the queries are one of perhaps a dozen questions. If generative AI can answer those with 90% accuracy, and then defer to a human, you've saved 80% of your personnel.
Human customer support representatives have << 100% accuracy, too.
It is impossible to be completely sure of non-trivial events in the future. Your very certainty makes me doubt.
@TomSwirly Then doubt it and rather believe all the marketing hype AI companies did for the last 50 years do to get their stock price or valuation up, I'm not going to write a dissertation if you have super reliable sources like that.
@thomasfuchs I don't think it will kill "us", but it will indirectly lead to people dying because of how much the biggest players in tech seem to trust it with things that it clearly is not even close to ready for. Even if it doesn't kill us, the hypercapitalist hell we live in may as well kill the tens of thousands of people being canned because everyone thinks "AI" is more capable than the people who wrote it.
@thomasfuchs I admit to being a centrist only when it comes to my selfish reasons:As a journalist, I want working AI transcription. Plop an interview recording in it, out come words. Have been using Whisper Transcription for a couple of months. Bizarrely uneven results. Sometimes 90% swell, sometimes a string of unpunctuated words. And, I interviewed pianists for a piano festival, and twice it changed "pianist" for "penis." Amusing the first time, problematic the second. AI is very, very, stupid
@markwedel@thomasfuchs I’ve used Otter for years. Prior to that, I either had to spend hours transcribing myself (at roughly one third speed) or pay transcription services, which mostly took a day or so to respond. The big issue for me is that Otter was always more accurate than the paid services. So objectively better and significantly faster. I don’t get that elsewhere (eg for editing LLMs are a joke).
Add comment