You know, when I think of supercomputer applications, I think of deeply analytical problems based on solid math and well understood algorithms that can be highly parallelised to take the maximum theoretical advantage of the hardware at hand.
You know, the opposite of what the “AI” crowd is doing. Throwing vast amounts of crunching power at a barely understood hypothetical black-box problem in hopes that it potentially yields some interesting results. Maybe.
AI is a buzzword with no real ability more than the Machine Learning approaches that have existed for over 20 years now.
AI isn’t real, we have not achieved artificial intelligence. Its a marketing term made to hype up pump and dumps.
They made LLMs make ChatGPT sound like a person but it’s just as dumb as the programs that came before that didn’t sound like a person.
The current Machine Learning models have reached the limitations of their usefulness, and creating more powerful approaches means someone inventing a more efficient algorithmic approach to Machine Learning, which we still haven’t done after we reached this model 20 years ago.
Huh.
You know, when I think of supercomputer applications, I think of deeply analytical problems based on solid math and well understood algorithms that can be highly parallelised to take the maximum theoretical advantage of the hardware at hand.
You know, the opposite of what the “AI” crowd is doing. Throwing vast amounts of crunching power at a barely understood hypothetical black-box problem in hopes that it potentially yields some interesting results. Maybe.
AI is a buzzword with no real ability more than the Machine Learning approaches that have existed for over 20 years now.
AI isn’t real, we have not achieved artificial intelligence. Its a marketing term made to hype up pump and dumps.
They made LLMs make ChatGPT sound like a person but it’s just as dumb as the programs that came before that didn’t sound like a person.
The current Machine Learning models have reached the limitations of their usefulness, and creating more powerful approaches means someone inventing a more efficient algorithmic approach to Machine Learning, which we still haven’t done after we reached this model 20 years ago.
https://www.reuters.com/technology/artificial-intelligence/openai-rivals-seek-new-path-smarter-ai-current-methods-hit-limitations-2024-11-11/
Why think of intelligent solutions when you can just throw more money at the problem. Surely that’s how you deal with issues.