The problem isn’t that AI is dogshit. In fact there are many things that have been revolutionized because of AI. Our understanding of biology has catapulted forward decades because of new models.
The problem is marketing. Those few use cases, while truely amazing, are only useful for a few things. The marketing of AI is horrible, throwing it into every product whether it makes sense or not. Throwing more money at it when it doesn’t make sense in the vain hopes that will eventually make it profitable. Then manipulating the market so that alternatives to your bad product aren’t feasible.
These silicon valley narcissists need AI to be profitable more than AI can become profitable.
Perhaps. My issue is more with the muddy waters when it comes to “AI”. It might be because it’s near to my profession, and I frequently encounter a fervent, yet clueless enthusiasm. When that same lack of specificity is used to argue for what they don’t really understand either, for example:
The problem isn’t that AI is dogshit. In fact there are many things that have been revolutionized because of AI. Our understanding of biology has catapulted forward decades because of new models.
It does beg the question if they understand what they are talking about. “AI is dogshit”, is weird. Are we talking about LLMs here? They are great for a lot of things. They are pretty horrible at doing what most seem to think they are great at, but that’s perhaps a different matter.
Then there is the “Our understanding of biology has catapulted forward decades because of new models”. If we are generalizing “AI models” to include the ones that have contributed to biology, we end up with a definition of “AI” that becomes “model based computational solution optimization”.
A big part of the problem is that everything is AI and AI is everything. Machine learning, LLMs, and “AI companies” are three very different concepts that get muddled together by newspeak.
Many machine learning applications are indeed truly innovative and a leap forward in how we deal with certain classes of problems, like in computer vision or computational biology as you mentioned. But the big trouble with LLMs imo is twofold: they are kryptonite for our monkey brains as language is something so uniquely human that it is a literal backdoor into our minds; and as a whole they are both affected and effected by the forced enshittification of capitalism, whether directly or indirectly, and the interests of “others”, leading to some not great times.
I agree with you. I’m working on a novel, and it’s been really helpful for preliminary editing advice, and bouncing ideas off of. Not perfect, but it’s better than no one. And I really don’t understand why it’s not marketed more as a collaborative assistant, rather than something that just does mediocre work for you.
I’m nearly done with book it suggested as research, and i don’t think it could have been a better subject. It’s a great pick for what I needed.
But compared to ChatGTP, Gemini is fucking retarded.
Are there any plans to release an AI that’s not dogshit?
The problem isn’t that AI is dogshit. In fact there are many things that have been revolutionized because of AI. Our understanding of biology has catapulted forward decades because of new models.
The problem is marketing. Those few use cases, while truely amazing, are only useful for a few things. The marketing of AI is horrible, throwing it into every product whether it makes sense or not. Throwing more money at it when it doesn’t make sense in the vain hopes that will eventually make it profitable. Then manipulating the market so that alternatives to your bad product aren’t feasible.
These silicon valley narcissists need AI to be profitable more than AI can become profitable.
AlphaFold is not an LLM tho
I think homey is using a more expanded AI definition then the “Looks inside: It’s a LLM” one that is generally used today.
Perhaps. My issue is more with the muddy waters when it comes to “AI”. It might be because it’s near to my profession, and I frequently encounter a fervent, yet clueless enthusiasm. When that same lack of specificity is used to argue for what they don’t really understand either, for example:
It does beg the question if they understand what they are talking about. “AI is dogshit”, is weird. Are we talking about LLMs here? They are great for a lot of things. They are pretty horrible at doing what most seem to think they are great at, but that’s perhaps a different matter.
Then there is the “Our understanding of biology has catapulted forward decades because of new models”. If we are generalizing “AI models” to include the ones that have contributed to biology, we end up with a definition of “AI” that becomes “model based computational solution optimization”.
A big part of the problem is that everything is AI and AI is everything. Machine learning, LLMs, and “AI companies” are three very different concepts that get muddled together by newspeak.
Many machine learning applications are indeed truly innovative and a leap forward in how we deal with certain classes of problems, like in computer vision or computational biology as you mentioned. But the big trouble with LLMs imo is twofold: they are kryptonite for our monkey brains as language is something so uniquely human that it is a literal backdoor into our minds; and as a whole they are both affected and effected by the forced enshittification of capitalism, whether directly or indirectly, and the interests of “others”, leading to some not great times.
Chatbots are a different type of AI entirely tho.
I agree with you. I’m working on a novel, and it’s been really helpful for preliminary editing advice, and bouncing ideas off of. Not perfect, but it’s better than no one. And I really don’t understand why it’s not marketed more as a collaborative assistant, rather than something that just does mediocre work for you.
I’m nearly done with book it suggested as research, and i don’t think it could have been a better subject. It’s a great pick for what I needed.
But compared to ChatGTP, Gemini is fucking retarded.
Or in other words, the problem is that AI is dogshit outside the very small range of tasks that it’s good at.
Yes.