yea, it looks that way.
Software will be generated on a need basis, once it's purpose ha been fulfilled it will be disposed. No need to store your excel app anymore, your devices will just generate the app when it's needed to view the file you need to view. On the fly!
It's next level.
nostr:nevent1qvzqqqqqqypzpdcg7uuj7kyyqcsjcwyzu7emcrvmprtzl906zuxsnyf8an38wrj7qqs06sz4cmm2ex6utxv3rjzf6qxjzpejz3ha7mae708clt5c9x664acdmgs4n
Sounds like FPGAs.
I know FPGAs are powerful but is that something they do now? I know they're used for mining but asics are better from what I understand. Didn't know they could do this kind of Ai stuff.
The original concept was a chip that re-programs its instruction set based on the tasks it's being asked to do most.
FPGAs are fascinating. I hear they aren't that great for LLM stuff though. CUDA is apparently the king.
General compute architecture is good enough for what I’m talking about. You don’t need efficiencies in hardware for most software tasks. You just need software that does exactly what you need at any given moment (and nothing else… all the extra features cause confusion and require learning/choices)
yup, that’s what I’m talkin’ bout! And if it’s made for you on-the-fly you won’t have to learn how to use it… it will just have the features you need right now
Yes. Interesting thought here. Once consumers are giving the option to remain ignorant of how the software they use actually works, where does that lead? Does it provide an profitable opportunity to those who do take the time to learn these things or does it make learning these thing obsolete compared to what will be created using generative tech? And what will be required to maintain these new technologies? Will that requirement create a whole new industry/skill set demand we can't quite fathom?
Personally I don't see a future where the need to understand how generative tech works will go away.
I don’t think it requires an ignorant approach. Could be open-source/transparent/auditable what the software is going. I use tons of open source software even though I haven’t personally read all the code. You have to figure out which layers you find most important to understand based on where you think problems might arise.
You don't think the tech will evolve fast enough to be able to patch these area where problems would arise?
The competency of the tech may become so intelligent that average people might not need to know such things as they wouldn't be pushing the tech to it's ever expanding limits.
I think it’s similar to other layers of computing technology. Everything is a new layer of abstraction.
I’d prefer a future where each layer is transparent to inspect even if you don’t choose to regularly audit everything.