Genai decided to skip that hard work and try to jump immediately to scale, and now we have supply chain logjam and uncertainty over use cases and all sorts of labour abuse formenting because this stuff was priced incorrectly on launch.

My opinion/experience, not my employer.. When aws launched ec2 m1 instances were feature poor (no ebs!!) and expensive. S3 was also expensive, but that scale and purity.. both won markets through -both- adding functionality -and- scaling so they could move down the commodity curve.

I think for things that are relatively small potatoes as far as risk goes, the value ive gotten from a lower stress level and more capacity to stay present with my loved ones, balances out the weird ai-slop risk of outsourcing the thinking to a model.. net-net i feel pretty satisfied

But then it occurs to me, i dont review the machine code a compiler spits out.. do i necessarily need to care, assuming i have good test coverage and functionality works?

But the volume of code makes it hard to review and be confident in whats written does only what i wanted and nothing else.

And the code that has been produced is definitely more detailed than i would have written myself, and has stood up to a bunch of testing that has broken other human authored code in the same apps

Add to the “signs of a bubble” indicators list..

Also.. framing this decision as an or is really weird.. why not do both?