One pitfall lots of people fell into with crypto that seems to be repeated with AI is basing fundamental objections on temporary problems. If your problem with crypto was the energy impact, did Ethereum switching to proof of stake change your mind? If not, was that actually your issue at all?
Comments
and also the criticism is still pretty valid for bitcoin
Otherwise it's pretty reasonable to not change your overall opinion just because one blockchain has fixed one issue from a very long list.
I don't really trust them to be good arbiters of what data should be considered open.
But AI is an enormous discipline with a ton of different technologies, each of them being nothing like the others.
But the shortcut of using "AI" to mean specifically "genAI" muddles this.
It actually changed my mind. I bought ethereum. Also because I think that financing PoW via transaction fees might not be sustainable for Bitcoin.
But I forget most workplaces don't have those specific small AI tools yet so it's all the big companies.
I remember Daphne Koller's seminal paper on machine learning for identifying at-risk samples in histopathology tumour sample data, and it felt like the future
The problem is ChatGPT, and Google Gemini, feel like "very big autocorrect", outright make up information, rely on mass copyright theft, and seemingly need more energy and capital than the death star
At least when it was Proof of Work it had intellectual value as an experiment of MEV interacting with low hashpower Nakamoto consensus.