I firmly believe we won’t get most of the interesting, “good” AI until after this current AI bubble bursts and goes down in flames.
I can’t imagine that you read much about AI outside of web sources or news media then. The exciting uses of AI is not LLMs and diffusion models, though that is all the public talks about when they talk about ‘AI’.
For example, we have been trying to find a way to predict protein folding for decades. Using machine learning, a team was able to train a model (en.wikipedia.org/wiki/AlphaFold) to predict the structure of proteins with high accuracy. Other scientists have used similar techniques to train a diffusion model that will generate a string of amino acids which will fold into a structure with the specified properties (like how image description prompts are used in an image generator).
This is particularly important because, thanks to mRNA technology, we can write arbitrary sequences of mRNA which will co-opt our cells to produce said protein.
Robotics is undergoing similar revolutionary changes. Here is a state of the art robot made by Boston Dynamics using a human programmed feedback control loop: www.youtube.com/watch?v=cNZPRsrwumQ
Here is a Boston Dynamics robot “using reinforcement learning with references from human motion capture and animation.”: www.youtube.com/watch?v=I44_zbEwz_w
Object detection, image processing, logistics, speech recognition, etc. These are all things that required tens of thousands of hours of science and engineering time to develop the software for, and the software wasn’t great. Now, freshman at college can train a computer vision network that outperforms these tools using free tools and a graphics card which will outperform the human-created software.
AI isn’t LLMs and image generators, those may as well be toys. I’m sure eventually LLMs and image generation will be good, but the only reason it seems amazing is because it is a novel capability that computers have not had before. But the actual impact on the real world will be minimal outside of specific fields.
NewDayRocks@lemmy.dbzer0.com 3 days ago
AI is good and cheap now because businesses are funding it at a loss, so not sure what you mean here.
The problem is that it’s cheap, so that anyone can make whatever they want and most people make low quality slop, hence why it’s not “good” in your eyes.
Making a cheap or efficient AI doesn’t help the end user in any way.
SolarBoy@slrpnk.net 3 days ago
It appears good and cheap. But it’s actually burning money, energy and water like crazy. I think somebody mentioned to generate a 10 second video, it’s the equivalent in energy consumption as driving a bike for 100km.
It’s not sustainable. I think the thing the person above you is referring to is if we ever manage to make LLMs and such which can be run locally on a phone or laptop with good results. That would make people experiment and try out things themselves, instead of being dependent on paying monthly for some services that can change anytime.
krunklom@lemmy.zip 3 days ago
i mean. i have a 15 amp fuse in my apartment and a 10 second cideo takes like 10 minutes to make, i dont know how much energy a 4090 draws but anyone that has an issue with me using mine to generate a 10 second bideo better not play pc games.
NewDayRocks@lemmy.dbzer0.com 3 days ago
You and OP are misunderstanding what is meant by good and cheap.
It’s not cheap from a resource perspective like you say. However that is irrelevant for the end user. It’s “cheap” already because it is either free or costs considerably less for the user than the cost of the resources used. OpenAI or Meta or Twitter are paying the cost. You do not need to pay for a monthly subscription to use AI.
So the quality of the content created is not limited by cost.
If the AI bubble popped, this won’t improve AI quality.
MrMcGasion@lemmy.world 3 days ago
I’m using “good” in almost a moral sense. The quality of output from LLMs and generative AI is already about as good as it can get from a technical standpoint, continuing to throw money and data at it will only result in minimal improvement.
What I mean by “good AI” is the potential of new types of AI models to be trained for things like diagnosing cancer, and and other predictive tasks that we haven’t thought of yet that actually have the potential to help humanity (and not just put artists and authors out of their jobs).
The work of training new, useful AI models is going to be done by scientists and researchers, probably on a limited budgets because there won’t be a clear profit motive, and they won’t be able to afford thousands of $20,000 GPUs like are being thrown at LLMs and generative AI today. But as the current AI race crashes and burns, the used hardware of today will be more affordable and hopefully actually get used for useful AI projects.
NewDayRocks@lemmy.dbzer0.com 3 days ago
Ok. Thanks for clarifying.
Although I am pretty sure AI is already used in the medical field for research and diagnosis. This “AI everywhere” trend you are seeing is the result of everyone trying to stick and use AI in every which way.
The thing about the AI boom is that lots of money is being invested into all fields. A bubble pop would result in investment money drying up everywhere, not make access to AI more affordable as you are suggesting.