What bothers me is that we are collectively conflating two very different things when talking about “AI”.
LLMs are inherently evil. They are built on the stolen knowledge of the entire human history, are controlled by the worst people alive and destroy everything they touch, from the environment to people’s brains.
Machine learning algorithms are actually neat. A specialized algorithm, trained on a specialized dataset, fine tuned and running on a finite number of features, can actually help humanity survey massive amounts of data that would be extremely expensive and time consuming when done manually. It can and is absolutely used for evil (see dynamic pricing, gambling companies etc). But it can also be used for good. And they’re cheap as fuck to run compared to LLMs. When you hear “AI has solved this and that”, it’s always someone putting the time, effort, knowledge and focus to build something hyper-specialized (unless it’s a $10 billion worth of burned credits publicity stunt).
And that’s a shame. Because LLM cunts reap all the publicity because lazy journalism conflate all ML-adjacent work into “AI”.
What bothers me is that we are collectively conflating two very different things when talking about “AI”.
LLMs are inherently evil. They are built on the stolen knowledge of the entire human history, are controlled by the worst people alive and destroy everything they touch, from the environment to people’s brains.
Machine learning algorithms are actually neat. A specialized algorithm, trained on a specialized dataset, fine tuned and running on a finite number of features, can actually help humanity survey massive amounts of data that would be extremely expensive and time consuming when done manually. It can and is absolutely used for evil (see dynamic pricing, gambling companies etc). But it can also be used for good. And they’re cheap as fuck to run compared to LLMs. When you hear “AI has solved this and that”, it’s always someone putting the time, effort, knowledge and focus to build something hyper-specialized (unless it’s a $10 billion worth of burned credits publicity stunt).
And that’s a shame. Because LLM cunts reap all the publicity because lazy journalism conflate all ML-adjacent work into “AI”.