In 1997, the usefulness of the internet vs the phonebook and newspaper was already readily apparent - at least to me. You could find what you need faster, and the results were just as accurate. In some cases, you got the information much faster - for instance, finding a phone number for someone across the country or world, or reading news from that location. And the information you got was pretty accurate. You didn't really have to worry that "The Times of London" website was showing you made-up news that wasn't in the print addition available that day in London, or that the Online Yellow Pages of Schenectady, NY was showing you hallucinated phone numbers for Al's Garage. But with LLM AI, you just never know. It seems to me that the whole point of AI is to take some of the cognitive load off of you, the user. If you can't trust what it tells you, and you have to laboriously verify everything it tells you, that defeats the point entirely.Yes to all? Anything I know meaningfully know about, say personal investing, I started at a beginner level. You’ve got to start somewhere. A 4th grade level seems perfect.Is this for entertainment or curiosity purposes, or are there decisions you want to make based on a 4th grade level explanation of a topic?
Part of the problem might be people seeing a new technology and trying to cram it into their current work flow. Sort of like 1997 when people saw the internet and could only imagine shoving current tools and processes onto it. They, rightly, didn’t see much use in logging onto the internet to get the same info they could get through the phone book and newspaper.
They just could not imagine the high speed internet of 2010 with things like uber, the rise of social media, Wikipedia (remember all the hate it used to get?), Airbnb, etc.
As a tutor it’s already starting to show its chops
If Tyler Cowen, smarter than me by a factor of 10, can find uses for it I probably can too
https://marginalrevolution.com/marginal ... ng-o1.html
Maybe eventually AI will get there, but for now, from what I can tell, it's used to do things I find it very easy and quick to do on my own: read a technical document in my field and summarize it accurately, draft an email, manipulate data and calculations in an Excel sheet, etc. This is totally different from the situation in 1997 because I can do whatever LLM AI is supposed to help you with more efficiently without using AI than with using it. Maybe when it gets to the point where it can explain a technical document outside my field and be trustworthy at it, I'd start to find it useful.
Statistics: Posted by snic — Sun Jan 19, 2025 5:48 pm