Gregory's ServerThe problem with predicting the imminent pop of the AI Bubble by pointing out how incredibly bad its business fundamentals are, is that tech investors—the crowd that needs to panic for the bubble to pop—have an extremely high tolerance for falsehoods and unrealistic promises. Otherwise they wouldn't be investing in tech.
 9 comments
@baldur well, they try to rach a "too big to fail" status by destroying existing structures and then filling them in with their stuff, as it is tradition for their caste. but they have a fawlty product, so hmm. @tomtrottel Worked for finance in 2008. If you’re a black hole devoid of ethics and morality, it’d be logical to try the same playbook. Because people asked, no it’s not Ed Zitron’s posts specifically I’m reacting to. He usually holds off from making timeline predictions. HOWEVER, his pointing out that it’s all a house of cards causes dozens of other people to write posts of their own predicting that it’s all about to fall apart  @baldur important lesson from the crypto bubble here as well. its been multiple years since the dominant narrative because that its largely useless/a scam, and the price hit all time highs just a few months ago, and is still stable also fairly safe to call generative ai a sector that has more practical usecases currently (even tho much lowers than the boosters would like you to believe) than crypto has, so this effect will likely be even stronger  @laurenshof @baldur I disagree about "more practical usecases". As much as I hate everything about cryptocurrencies, they have had practical use cases for people denied access to receiving money through conventional channels, and to people who need to buy illegal goods or services. "AI" OTOH doesn't actually meet any need. @dalias @laurenshof @baldur Well, for flooding the web with disinformation, scams, and spam, it’s unparalleled at extruding vast quantities of information-shaped content product. @KronoGarrett @laurenshof @baldur Oh yes, it has harm-committing uses. I should have clarified that I meant beneficial uses. Look at open ai; they clearly stole more material for this latest round of training which is why they are covering it up |
The other problem is that the tempatation to commit fraud—fake the numbers to make AI look successful—is going to be incredibly strong, because those who believe in "AI" also believe it will happen eventually. For many of them, a little bit of criminal falsifying in order to buy time for the revolution to truly kick in won't even cause them to blink.