It’s pretty good! I especially agree with the part that generative AI is most useful if you are already an expert. Beyond that lies danger in trusting confabulations as fact
People are becoming more aware of what is just a fascist scam for masquerading as human for manipulation and short term capital concentration. That requires the hypsters to ignore the built-in regression of data corruption and more energy waste than entire countries.
I don't even need to read the op/ed and I don't trust op/eds anyway. We are in the second quadrant of the Gartner Group cycle, the hype phase. It will be succeeded by the disappointment phase and then by the reasonable expectations phase.
Having been in the sales job market over the last 12 months, I avoided openings connected to AI start-ups suspecting that bubble will soon burst leaving me back where I first started
This one-
“The issue of whether workers and work is wasteful is a subjective call that A.I. cannot make. But it can justify what a decision maker wants to do. If Musk wants waste, A.I. can give him numbers to prove waste exists.”
Wow that’s a great way to say what I’ve been thinking. But what if the “experts” have already slowly been replaced by people who love shallow but quick methods to climb ladders because the ladder exists…
I don't know, at this point I feel like every article like this comes from people who "research" ai reads like the person who invented the lead gasoline suddenly realised there's lead poisoning later.
To all the educators who think we owe it to our students to teach AI as part of the writing process because "it's inevitable"
What happens to these students when the AI bubble bursts and is no longer freely accessible and we've failed in our responsibility to teach them how to write?
Agree with the gist of the article, however a lot of “AI” haters are woefully ignorant of the advances in scientific research enabled by modern ML techniques. Fields like pharma, astronomy, GIS and many more are benefiting from the AI craze. To fully discount it as a dumb fad is just idiotic.
And why should we have? We're not all data scientists. The public experience of AI is the generative AI slop added into online sites and products as tech companies scramble to find a way to make a profit from a tech the public doesn't want.
I didn't say that anyone should have. And I agree that generative AI has been shoved into far too many public facing products without much consideration of whether anyone actually wants it. It's quite annoying.
I would like to make sure that there is a VERY CLEAR distinction being made here between machine learning systems and generative AI. The former are incredibly valuable, very powerful... and not new. ML systems have been detecting cancer and folding proteins for over a decade.
Generative AI is nothing at all like those machine learning systems. It is, technically, a very interesting but very limited technique with a few intriguing but ultimately niche applications that happen to be capable some very spectacular slight of hand.
The 'AI haters' aren't upset about AI for scientific research, we're upset about GENERATIVE AI - because it is theft. AI bros are building companies based on stealing from artists, writers, musicians so that lazy arseholes can reproduce bastardized copies, resulting in a decimated creative industry
Yeah, and as a science tool it's confined to contexts in which its information output would [ideally] be vetted. I don't want a word-regurgitator to make fridge magnet poetry out of every blog post on the internet so laypeople are given misinformation about what plants are toxic or how to cook wrong
Like THAT'S WHAT THE GOOGLE IS FOR. That's what search engines already do! The only reason people won't click random result links for themselves is because Google enshittified itself to prop up shit you don't need as top results a lot of the time.
Yeah I also think the generative slop sucks. There’s just so little nuance in these discussions though, the nomenclature needs to be more specific. The ML techniques that have been pioneered over the last several decades are genuinely useful and fascinating.
You'll have to ask that question of tech journalists on the AI slop train parroting whatever Scam Altman says, even though he's still not proven his product is anything more than a theft machine. The reporting on genAI is pure hype, only a handful of tech journos,like Ed Zitron, are calling it out.
sums up my thoughts on the subject. it can not analyze and cannot create data. all of the "it got the same results in hours, that took people years to get" stories ignore, or just don't understand, that data and information are cumulative: without all those years of human work, ai doesn't work.
Perfect-
"A.I.’s most revolutionary potential is helping experts apply their expertise better and faster. But for that to work, there has to be experts."
When she's right, she's right. I guess she doesn't consider old white guys who work in AI to be part of her intended audience. That's fine. Maybe just a missed opportunity. It was a good essay in any case
I loved this sentence: "A.I.’s most revolutionary potential is helping experts apply their expertise better and faster. But for that to work, there [have] to be experts."
I’m so confused by seeing technology that has progressed from garbled sentences to being able to write many college-level essays over a span of just 6 years…. And concluding that the field is all fumes. Just mid. I’m incredulous.
But in all seriousness, OpenAI is the top player in the space, and last year they spent $9 billion to lose $5 billion. They lose money on every subscription tier. They want $40 billion for just this year.
And for all that money, they produce middling quality essays for lazy students.
I think it’s reasonable to say “these essays aren’t yet that good” (they’re not!) and “this stuff costs a crazy amount of money upfront” (it does!), but that’s not the argument an essay like this makes; the essay says “this is going nowhere”, and that, to me, I find so hard to sympathize with.
I’m just not impressed when computers can do something worse AND more expensively than a person. Especially since there’s no reason to think that these LLMs are going to continue to improve. They’ve used all the training data that exists and the latest are indistinguishable from the old versions
I think too that it’s ok for tech to provide only incremental gains. Small accelerations in everyday tasks add up to a big improvement, & the democratisation that LLMs provide to non-native English-speakers is huge. But the AI hype is doomed to overpromise and underdeliver.
Comments
A deep, effective augmentation of existing human skill sets.
For example you don't want to trust your trial to an AI entity - but you do want a good lawyer who knows how to expertly use one.
example: I think we're all becoming pretty sick of seeing AI-generated art by now.
I would rather commission art from a real artist, I don't have any qualms about how they produce their work as long as it doesn't look like AI crap.
However credit where it is due: transcription was once a luxury and thanks to AI transcription, it's become a necessity.
It's a tool (like an alligator), not a one click artist.
Some musicians have used AI in their recent work with pretty interesting results. One of my faves is Caribou's latest record
https://thedsrnetwork.com/rethinking-the-way-we-think-about-how-we-think
The article was definitely not mid.
We erode investment in human development and expertise at our great peril.
https://bsky.app/profile/carlbergstrom.com/post/3lljyl7ains2e
“The issue of whether workers and work is wasteful is a subjective call that A.I. cannot make. But it can justify what a decision maker wants to do. If Musk wants waste, A.I. can give him numbers to prove waste exists.”
1.) undermining labor
2.) offloading liability
and nothing since has proven me wrong
What happens to these students when the AI bubble bursts and is no longer freely accessible and we've failed in our responsibility to teach them how to write?
"A.I.’s most revolutionary potential is helping experts apply their expertise better and faster. But for that to work, there has to be experts."
And for all that money, they produce middling quality essays for lazy students.