Mastodawn

what’s your most interesting hot take about where AI might change tech in the next few years???

it seems like nearly everyone in tech believes that AI adoption is rapidly changing… something.

what exactly it’s changing and where that will lead seems to be very contentious.

obviously there are the kool aid true believers.
and there are the never-llm folks.
that’s fine. i understand those views.

but i’m more interested in the hot takes of the folks that fall somewhere in between.

Show thread

Kuba Suder • @mackuba.eu on 🦋1d ago

@isaiah I don't know, man, I'm not driving this bus, I don't even know where it's going, I'm just riding it and looking out through a dirty window

Show thread

Fredrik Björeman 1d ago

@isaiah Here’s one: right now is peak utility and availability, we’ll never again have the ability to throw this amount of resources at this number of things. (I.e. when costs catch up nobody will want to pay the full price of maintaining all of what is heavily subsidized right now.)

Show thread

kolya 1d ago

@bjoreman @isaiah
maybe. but it's also possible that at that point the tech has developed enough that you can run the coding AI you need on premise or even on your machine.

Show thread

Fredrik Björeman 1d ago

@kolya @isaiah Yeah, and that could be kind of nice. Still, it’s not certain that anyone will be able to afford to train those to the same level as today, so the peak could still be passed.

Show thread

Jared Sinclair 1d ago

@isaiah Since you asked, and I fall into that bucket…

The average software project team of the 2030’s looks like a commercial airplane cockpit: pilot, copilot, and autopilot.

Breaking the analogy, maybe it’s 2 pilots and 2 copilots, so that there is continuity across human life events (vacation, death).

Head count varies by industry: scrappy startups have fewer human heads than in legacy domains. More risk averse = less tolerance for the number “zero-humans present” days per calendar year.