How can someone vote NO on this?

Even comparatively small machine learning methods (small compared to the neverending media hype train around LLMs today) have revolutionized every field and subfield in science. Just one example is tensor network methods and neural quantum states which are the algorithms that are single handedly delaying quantum supremacy. Any time a new group publishes a paper on a calculation done on a quantum computer that they claim cannot be done classically, it takes less than a month for another group to retort with "we did it classically using TN/NQS".

@Pykess i voted no because LLMs hallucinate a lot and they haven't replaced me yet, and my job would be easily replacable by a competent AGI

@TiredCliche I've already voted but what are we considering a lower bound for "AI" here? First two examples that come to mind - which are "AI" as far as I'm concerned - are cancer diagnosis and alphafold. You can definitely argue a case for LLMs but I don't think you even need to.

@TiredCliche An LLM hasn't replaced your job yet, but machine learning algorithms have revolutionized significant parts of your life: have you ever ordered from Amazon? Been to a big name grocery store? Received something that was made from materials that were delivered via rail or ship?

@Pykess i wouldn't have classified TN as machine learning honestly

@FedericoRottoli The entries in the various tensors are learned parameters, no different than the entries in a NN. And, of course, a NN is just a tensor network.

@Pykess To me "changing the world" implies some kind of ubiquitous and global effect. Your example is way too narrow for that. Also I don't agree with the claim that AI has already revolutionized every aspect of science in a significant way.

@Pykess can you give me an example of someone doing something thought to be classically impossible with tn/nqs? Not trying to argue, just genuinely curious to look into it.

@ProjectVictory This was put on the arXiv just a few days after IBM's kicked Ising model simulation on a quantum computer. In their paper, IBM claimed that this system could not be simulated classically (with reasonable accuracy within any reasonable amount of time) with any current or short-term foreseeable methods. https://arxiv.org/abs/2308.03082

also check this

Yes, it made google search shittier.

@ProjectVictory This plus Google/Alphabet's ever increasing pushing of ads and pay-to-win search has made Google search easily an order of magnitude less useful than five years ago.

[Deleted, my comment posted twice for some reason]

@ProjectVictory I have not noticed this. In what way?

@FlorisvanDoorn There are several problems but the main one I was referring to is SE-optimized, low quality and/or ai-written "news" articles as top results. Not really google's fault, but the state of internet is pretty dire in general. Most websites that had any goals beyond making money by appealing to the lowest common denominator are dead.