"Our research is greatly sped up by AI but AI still needs us" (twitter.com)
"I crossed an interesting threshold yesterday, which I think many other mathematicians have been crossing recently as well. In the middle of trying to prove a result, I identified a statement that looked true and that would, if true, be useful to me.
"Instead of trying to prove it, I asked GPT5 about it, and in about 20 seconds received a proof. The proof relied on a lemma that I had not heard of (the statement was a bit outside my main areas), so although I am confident I'd have got there in the end.
"the time it would have taken me would probably have been of order of magnitude an hour (an estimate that comes with quite wide error bars). So it looks as though we have entered the brief but enjoyable era where our research is greatly sped up by AI but AI still needs us.
"PS In case anyone's worried that it used a lemma I hadn't heard of, I checked that the lemma was not a hallucination."
10 comments
[ 3.4 ms ] story [ 31.7 ms ] threadIt also significantly changes my current job to something I didn't sign up to.
> _brief_ but enjoyable era
In the past this tradeoff probably was obvious: a farmer's individual fulfillment is less important than feeding a starving community.
I'm not so sure this tradeoff is obvious now. Will the increased productivity justify the loss of meaning and fulfillment that comes from robbing most jobs of autonomy and dignity? Will we become humans that have literally everything we need except the ability for self-actualization?
Well that's comforting.
All of these seem to subscribe to "inevitability", have no issues that their research relies on a handful of oligarchs and that all of their thoughts and attempts are recorded and tracked on centralized servers.
I bet mathematical research hasn't sped up one bit due to "AI".
What is the broader context of OP trying to prove a theorem here? There are multiple layers of purpose and intent involved (so he can derive the satisfaction of proving a result, so he can keep publishing and keep his job, so their university department can be competitive, etc), but they all end up pointing at humans.
Computers aren’t going to be spinning in the background proving theorems just because. They will do so because humans intend for them to, in service of their own purposes.
In any discussion about AI surpassing humans in skills/intelligence, the chief concern should be in service of whom.
Tech leaders (ie the people controlling the computers on which the AIs run) like to say that this is for the benefit of all humanity, and that the rewards will be evenly distributed; but the rewards aren’t evenly distributed today, and the benefits are in the hands of a select few; why should that change at their hands?
If AI is successful to the extent which pundits predict/desire, it will likely be accompanied with an uprising of human workers that will make past uprisings (you know, the ones that banned child labor and gave us paid holidays) look like child’s play in comparison.