"Buying time" is an ordinary phrase in English but, it doesn't make a lot of sense here because there'd be some activity SpaceX is doing to "buy time". If you're just waiting that doesn't "buy time".
> as soon as AI companies aren't able to subsidize anymore, we will see a renaissance of IT consulting
I think the price that LLMs have to get to for companies to return to paying consulting rates is much higher than you think. Claude Code at $2000/month is roughly one day of a top consultant's rate. (And this omits the possibility of companies make it a capex by hosting open source models)
It's not "appearances".
I know all the guys I meant since university.
All of them were top notch PhDs. Physics. CS. Some were excellent teachers even before graduation. Others were at the leading edge of ML even two decades ago.
In short: it's not a superficial impression as a customer. I know their pros and cons from studying to social life during and after graduation, early career, ...
So, my POV is personal experience. Your POV is gossip?
That was in the past. Now they would just prompt LLMs on your behalf for a fat fee. You might as well do it yourself and spend that fee on extra tokens.
You could look at only two niche blockchains, QRL and ABEL, and they are both affected. Algorand is the most established L1 with a quite developed migration plan (https://algorand.co/technology/post-quantum) but also not really thriving. I think for Bitcoin in the longterm it is a massive risk psychologically, because what to do with all of the locked in value, that cannot be migrated. My guess is that the market panics and security breaks down, because it is not worth it anymore to run that many nodes. Best time to buy would be then and hope it recovers. It is actually a big chance to move just to another L1 which migrates. Those risks are all priced in.
Not useless if it pinpointed that most of the foundational research papers that initiated this trend were of dubious intellectual integrity- i.e. the source was compromised
reply