Vector databases are relatively good at this kind of thing, because they can find records based on queries that are semantically close instead of just a lexical search. It would probably still make sense to split the information up in fragments such as e.g., "Interstellar movie," "watched on February 2nd, 2021," "made me vomit", and then connect those records to each other. GPTs are good at that kind of preprocessing. The idea would not be to store exact data such as timestamps and that's not how vector databases work, so recall would be more associative just like for humans (I can't ask you what movie you watched on Feburary 2nd, 2021 and expect an accurate reply either).
tias
I agree that it has limits but there are things we could do to make it reasonably good. ChatGPT knows how to execute actions (such as calling an API or doing a web search). It could probably be made to store and look up information in a vector database, essentially giving it a long-term memory.
Given some smaller breakthroughs in performance and model size we could conceivably retrain the network on new input continuously, in order to incorporate new knowledge.
If anyone is curious about the reason, my primary motivation for having Premium is that I don't want my kids to see more ads than necessary, and they're on YouTube on their phones, on the Chromecast (connected to a dumb projector) and on the computer.
The YouTube app has a lot to offer them that they won't get in other apps and I can't realistically force them to use other alternatives everywhere. Especially since I have shared custody and they'll be using YouTube at their mom's place as well.
I don't have an ad blocker, I just have the standard strict tracking protection enabled in Firefox. What's more, I pay for YouTube Premium. But still they add a five-second delay every time I visit a web page. It's infuriating.
I don't see how working from home changes anything. If you can do that kind of logging on their work computers at home, you can do it at the office too. Besides (in the EU at least) you have to inform employees about the extent of the monitoring beforehand. Can't imagine how they expect to attract competent employees that way though.
I believe it will take at least 15-20 years before the majority of engineers can be replaced. I do agree it will happen eventually. But my point wasn't about what will happen in the future. It was about whether engineers are losing their jobs due to AI at the present moment, as the article claims.
And if you're on Firefox, will it still add that forced 5-second delay while loading the page?
I don't think they are engineers. AI isn't anywhere near replacing engineers yet.
Not to mention his quest to reach Tannu Tuva. Or all the crazy things he talks about in "Surely you're joking, Mr Feynman". I feel like a movie about him should capture his playful approach to learning and teaching, and actually go into the subjects he studied. He was good at explaining things simply and the movie could do the same. It should awaken a desire to explore the mysteries of the universe. An antidote to all the anti-intellectualism we suffer from today.
I have yet to understand the "discrepancy". Which brand doesn't default to locking the phone automatically after a short timeout? I've never had a smartphone that stays unlocked.
Again, you will certainly hit limitations if you push it, but the example you give would work fine if you just append the added information to the database. A query for Interstellar would return both your original statements and the fact that you later said you lied about it, and all of these records are inserted into the GPT's context (short-term memory) when discussing that subject.