At this point computation is in essence commodity. And commodities have demand cycles. If other economic factors slowdown or companies go out of business they stop using compute or start less new products that use compute. Thus it is entirely realistic to me that demand for compute might go down. Or that we are just now over provisioning compute in short or medium term.
I wonder, is the quality of AI answers going up over time or not? Last weekend I spent a lot of time with Preplexity trying to understand why my SeqTrack device didn't do what I wanted it to do and seems Perplexity had a wrong idea of how the buttons on the device are laid out, so it gave me wrong or confusing answers. I spent literally hours trying to feed it different prompts to get an answer that would solve my problem.
If it had given me the right easy to understand answer right away I would have spent 2 minutes of both MY time and ITS time. My point is if AI will improve we will need less of it, to get our questions answered. Or, perhaps AI usage goes up if it improves its answers?
With vision models (SOTA models like Gemini and ChatGPT can do this), you can take a picture/screenshot of the button layout, upload it, and have it work from that. Feeding it current documentation (eg a pdf of a user manual) helps too.
Referencing outdated documentation or straight up hallucinating answers is still an issue. It is getting better with each model release though
Always worth trying a different model, especially if you’re using a free one. I wouldn’t take one data point to seriously either.
The data is very strongly showing the quality of AI answers is rapidly improving. If you want a good example, check out the sixty symbols video by Brady Haran, where they revisited getting AI to answer a quantum physics exam after trying the same thing 3 years ago. The improvement is IMMENSE and unavoidable.
The problem is it's inability to say "I don't know". As soon as you reach the limits of the models knowledge it will readily start fabricating answers.
Both true. Perplexity knows a lot about SeqTrack, I assume it has read the UserGuide. But some things it gets wrong, seems especially things it should understand by looking at the pictures.
I'm just wondering if there's a clear path for it to improve and on what time-table. The fact that it does not tell you when it is "unsure" of course makes things worse for users. (It is never unsure).
More so I meant to think of oil, copper and now silver. All follow demand for the price. All have had varying prices at different times. Compute should not really be that different.
But yes. Cisco's value dropped when there was not same amount to spend on networking gear. Nvidia's value will drop as there is not same amount of spend on their gear.
Other impacted players in actual economic downturn could be Amazon with AWS, MS with Azure. And even more so those now betting on AI computing. At least general purpose computing can run web servers.
Probably some type of balanced trade. Holding each trade partners currency and balancing it at times. Or even between multiple holders and currency pairs.
If there is need you can now build very complicated systems as everything is digital anyway.
Maybe stable coins would be finally useful. Each currency has own stable coin and then they are automatically traded in massive market... /s
This comes to not "smartness" of LLMs. But reality that we do not even want anything novel in these exercises or exams. And same areas are repeated multiple times so naturally there is lot of these in training data.
This is one area where LLMs really should excel at. And that doesn't really mean that students should not also learn it and be able to solve same issues. Which is real dilemma for the school system...
On other side in western systems funded by taxes the incentive is still to give out as many degrees as possible as schools get funding based on produced degrees.
Mostly done to get more degree holders which are seen as "more productive". Or at least higher paid...
It fails the basic human behaviour. In general humans are not ready to admit fault. At least when there is no social pressure. They might apologize and admit mistake. Or they might ask for clarification. But very rarely "You are absolute right" and go on entirely new tangent...
It is IRC, but with modern features and no channel splits. It also adds voice chats and video sharing. Trade off is that privacy and commercial platform. On other hand it is very much simpler to use. IRC is a mess of usability really. Discord has much better user experience for new users.
> Discord has much better user experience for new users.
Until you join a server that gives you a whole essay of what you can and cannot do with extra verification. This then requiring you to post in some random channel waiting for the moderator to see your message.
You're then forced to assign roles to yourself to please a bot that will continue to spam you with notifications announcing to the community you've leveled up for every second sentence. Finally, everyone glaring at you in channel or leaving you on read because you're a newbie with a leaf above your username. Each to their own, I guess.
/server irc.someserver.net
/join #hello
/me says Hello
I think I'll stick with that.
At least Discord and IRC are interchangeable in the sake of idling.
This only happens if the server is BIG _or_ if the admin is a grifter who's 100% sure their server will hit it big and has 120 channels and 40 bots and 9 users total.
It works. But for most it is not sustainable. It in most cases collapses eventually. But ideas and words and now pictures and videos do sell as in get pre-orders or pre-payments.
If something even more drastic happens. China might even attempt unification with some reasoning like protecting Taiwan from USA or other nations.
reply