As RAM chips and processors started becoming super-cheap since the 1990-00s era, somewhere down the line technology firms started this abhorrent practice of 'taking compute resources for granted' or stated another way - sacrificing programming efficiency to prioritize developer productivity/time.
What Altman recently said about consumer smartphones and laptops 'wasting' compute resources is a criticism that also finds its way towards these bloatware and cruft creators of the past two decades. To some extent, it also finds its way to the overwhelming marketing and PR that convinces a smartphone buyer to go for 16GB RAM even when 4GB suffices for their daily tasks.
Despite all the gloom surrounding this present silicon crunch, if there is one good which can come out of this situation it's this - there is a tiny ray of hope that compute resources will no longer be taken for granted and these bloatware creators in large tech firms will once again be forced to write more efficient and leaner software.
We should instead cap how much processing power these companies can accumulate per year, so the rest of us can have affordable devices and opportunities.
For using the phones for all tasks formerly accomplished with PCs - people reduced significantly their use and purchase of PCs and replaced them with phones and sometimes tablets. There's nothing "insane" about it, it's a well known fact.
Another well known fact is Altman's promises of AI abundance, that should have been the result for RAM, GPUs and all hardware too. However, not quite unexpectedly, we quickly got the bait and switch instead, we're now supposed to sacrifice ourselves for AI, because look "it knows some math". This is what's insane.
The frontier labs are running a scalping operation on the hardware market, they front run it with large orders depriving potential competitors from reasonably priced hardware and, at the same time, selling that same hardware in the form of premium services back to those who don't have it.
If there's a solution to this mess, it's investment in more fabs, more hardware, more chip-making companies, as it is, the market for chips is highly monopolized and that creates existential and national security risks. More funding for AI is only wasting money on overpriced hardware and energy infrastructure, now is the time to redirect that funding into creating multiple competitive makers of both categories. This is what a properly functioning market would do.
An alternative, specially because AI isn't regulated to oblivion yet: just don't use OpenAI. Vote with your wallet. That's it. There's no need to "cap how much processing power these companies can accumulate".
If you are right in your hypothesis, more people will agree with you simply because people in general are smart, and OpenAI will die a natural death.