I also have to point out that the most expensive space telescope (JWST) cost about $500 million/year. We spent 1000x that much on AI development in 2025.

Data collection is essential for discovery...and it's remarkably cheap compared to many other things we do routinely.

#science #nature #history #tech

I've also seen smart people tie themselves into knots trying to defend the original claim.

"He just means big science is expensive."
"He just means that AI can help with data analysis."
"He just means that string theory is a dead end."

But that is not the claim, and the efforts to justify it only make the argument even stranger.

@coreyspowell well, for analysis of ever increasing amount of astronomical data, some kind of automation is needed anyway. So maybe it would be better use of AI, than all this chatbot nonsense.

The huge colliders are special case, that now there is AFAIK no special prediction in physics, which can be confirmed or falsified at higher energies. Somehow it is probably not the direction to find any new physics (which would be cool). Also the dark matter detectors are somehow infamous as spending huge amount of money for (predictably) finding nothing.

The situation in astronomy is very different and of course we need new telescopes and new ideas for telescopes. Lot of them would have to be placed in space, probably.

So, somehow the discussion "what next in science" makes sense, and I would not probably bet on particle colliders to be the right answer. Still, over-relying on LLM-líke AIs si ridiculous. Of course, science needs new (not necesarily "more") empirical data and also, for huge amounts of data, some automation to process them.