If AI has a future (a big if), it will have to be economically viable. An industry can't spend 1,700% more on Nvidia chips than it earns indefinitely - not even with Nvidia being a principle investor in its largest customers:
If you'd like an essay-formatted version of this thread to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
This is the fatal flaw in robotaxi schemes. The "human in the loop" meant to keep the murderbot from smashing into other cars, steering into oncoming traffic, or running down pedestrians isn't a driver, they're a driving instructor. That's a much harder job than drivingr, even when the student you're monitoring is a human, making human mistakes at human speed. It's even harder when the student driver is a robot, making errors at computer speed:
Another great piece about AI from @pluralistichttps://pluralistic.net/2024/04/01/human-in-the-loop/#monkey-in-the-middle
Cory pinpoints in a few succinct paragraphs how robot assistance has been and can be great and helpful and turns us into powerful "centaurs" but that the way to profits lies in "reverse centaurs" where AI is the pilot and humans are tasked to correct bot productions at superhuman speed, degrading us to robots. In other words, AI driven Automation is only profitable if it enshittifes our work and life.
This one focuses on LLMs and the idea of the "reverse centaur", where a robot does the fun stuff while a human does the tedious, error-prone work.
I'll note from the periphery that, despite the current hype, AI is more than LLMs. There are other AI systems (e.g., Chess and Go players, VLSI design tools) that do have an internal model of the domain about which they are reasoning. Unfortunately, there's a slippery continuum:
solves the problem perfectly and deterministically
significantly outperforms any human
about as good as an expert human, but makes different, weird mistakes
Le toujours formidable Cory Doctorow aka @pluralistic sur l'usage des AI (et le fait que leur business modèle ne peut reposer que sur l'exploitation des humains, à la manière des entrepôts Amazon, au lieu de leur libération)
Here's a fun AI story: a security researcher noticed that large companies' AI-authored source-code repeatedly referenced a nonexistent library (an AI "hallucination"), so he created a (defanged) malicious library with that name and uploaded it, and thousands of developers automatically downloaded and incorporated it as they compiled the code:
If you'd like an essay-formatted version of this thread to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog: