> this whole "computers can't think" thing is making LLM skepticism seem very unscientific.

It's just shorthand for "that's an extraordinary claim and nobody has provided any remotely extraordinary evidence to support it."

Lots of people consider company valuations evidence of a singularity right around the corner but it requires a very specific kind of mindset to buy into that as "proof" of anything other than very compelling hype by people who have turned financial scams into an art form.