3 ms·
Because we've seen the hype cycle run before, With AI, with VR, where unrealistic promises are made and widely believed for no real reason, and no one cares whe
by h2odragon 2y ago
Because we've seen the hype cycle run before, With AI, with VR, where unrealistic promises are made and widely believed for no real reason, and no one cares when they're never realized.
"AI can do X" for all X is an unrealistic hope.
- red-iron-pine 2y agoonce it hits the "For Dogs" phase you know it's done
- arisAlexis 2y agoAi passes the Turing test and mensa admission test and aces lawyer,coding and medical tests. That's my point, traumatized opinions are biased.
- h2odragon 2y agoIs it possible the people administering these tests are less than honest about the outcome? Perhaps they have a vested interest in the results they've shown?
- DamonHD 2y agoIt's also not the case that all (or even most) AIs ace all (or even most) meaningful tests that aren't pattern or memory based. Adding multi-digit uncommon integers is too hard for many LLMs for example. My fellow uni students could pass tests on one of my courses by regurgitating the (wrong) answers that had been published and reused for years. I was not popular with faculty for pointing that out, and had most of a year's papers cancelled after the fact...
- wryoak 2y agoAll opinions are biased. It is a tautology. I guess my biased question is to do with cost. I’ve not been able to run any models of the caliber that could pass a medical test on my local machine. What kind of hardware do I need to own or to lease in the cloud to write code that doesn’t arbitrarily include comments without comment initializing tokens, and how much will the whole pipeline run me? Is that already cheaper than my salary? Some 25% of code at Google is now written by AI, but all of that code is still reviewed by humans. Is it worth it if you still have to pay for both? Quite possibly, for some of the huge companies, but most devs are working in smaller shops: is it still a reasonable expense for them?
- Ekaros 2y agoIs passing a formulaic test a high bar? Lot of people pass these tests with training. Taking the lawyer example, does model that passed test also be able to correctly write all documentation for a case from start to end? Without hallucinating any non existing cases or references?
- arisAlexis 2y agoUmm yes actually just google the latest news on this. They are better than paralegals too
- abenga 2y agoFor legal stuff, I have heard of lawyers being excoriated in court because of presenting AI-hallucinated case law; not sure when it started passing this bar.