I expect AI researchers in general have an understanding of what other jobs ACTUALLY consist of that's only a little better than "I've seen this job on TV and assume it's just that bit stretched out to 8-10 hours a day."
There's also the very real possibility that the AI researchers all running around saying "well, it'll be good at what administrative assistants do but bad at picking fruit" are creating an investment environment in which one of those projects gets a lot more funding than the other does.
Describing "some AI folks guessed that this might happen" as "theoretical capability" is deeply disingenuous.
On the other hand, "some AI folks guessed that this might happen, and here we are three years later, look at how the observed behavior resembles the guesses" is an interesting observation.
If I'm the product lead at Square I'm not so worried that a team of twenty reimplements my core product, because good luck getting anyone to use it.
But fifty two-person teams, each going after some part of your consumer base that has been complaining about something for years? Could be a problem.
It's absolutely true that getting the actual customers-- and often converting them from their existing providers-- is the real challenge here.
But: a lot of monolithic SaaS may actually float on that cost of implementation. Specialized, industry-focused treatments might still eat their lunch.
There's this gap between how well it gets a prototype off the ground in an empty repo and how poorly it adds a fairly modest new bit of functionality in a mature codebase.
I'm inclined to believe it'll improve, but the graveyard of AI is riddled with headstones reading "this will scale."
On the one hand, absolutely. On the other: a LOT of Jira user's first impression is "wow this is a bunch of stuff I don't need, I wish it just had the six 'standard' ticketing features and this ONE thing I specifically need," and that was quite hard to build two years ago, and is now REALLY easy.
Can't recreate Block's core business from fundamentals out of a 4 person company in temporarily-retired engineer's guest houses.
I mean, unless you're right about AI. Then you'll be so glad your swiss cheese of a remaining workforce is ready to maintain that legacy stack.
If you think AI makes engineers 10X more efficient, saying "we don't think we can find anything profitable to do with 10X the engineering capacity" is sad, but "so we're releasing a bunch of engineers with intimate knowledge of our business and no existing costs into our competitive space," that's 💯
The counterargument to "these companies are cutting jobs because they can get the same work done with a AI+a lot fewer engineers" is "these companies are using AI as an excuse to cut jobs, because all of their post-2019 hiring was for projects that never made money," which really isn't much better.
Actual software engineers working actual jobs are frantically trying to figure out how to build safely and sustainably with these tools, and while I appreciate SOME of the sentiment behind "just don't use the slop robots," we would lose that fight, and my guess is, eventually, so will everyone else.
Me in 2023: I wish there was a Star Trek series produced by HBO. That'd be amazing.
Monkey's Paw: *curls*
Me in 2023: These Shein monkey's paws don't last for shit. Thank God I bought forty of them.
*Numismatics" is such a great word.
Let's all guess how long "temporary" is.
Gotta make room for at least six more Fire Emblem characters first. And the Dreamcast VMU.
If you see this, share your Smash Bros main
It's like saying "we think teleportation should be possible eventually so we're breaking ground on luxury condominiums on a vacant lot of land on the edge of Death Valley."
You are skipping several important steps!
There are essentially no jurisdictions where a vehicle without a steering wheel can operate on public streets for the very good and very simple reason that if the vehicle's ability to drive itself is disabled emergency responders need to be able to move it without first installing a steering wheel.
Someday that man is gonna eat his leprechaun.
Guy illegally lobbing explosives at your home with no strategy whatsoever thinks you should stay at home where you'll be safe.
Glad I got my warp back to working.
Fixed.
Alanna Ubach on Peacock's Ted delivers an absolutely incredible comedic performance, just virtuosic. Every line delivery is a work of art. If this show was on Netflix she'd already have an Emmy.
When you commit murders while in the commission of an illegal act, such as the Iran War, they're "felony murders" and intent is no longer relevant to their conviction.
It was not a "mistake." It was a mass killing.
No, that he would accept the conclusion of an investigation. A thing he famously has never done unless the conclusion was "Trump was right."
Do you have any clip of Trump saying he will accept the outcome of the investigation? Or, as an alternative, a time in Trump's life when he accepted the outcome of any investigation when he did not like the outcome?
Imagine the paradise where there was one tactical error or mistake leading up to this war, instead of the entire lead up being just tactical errors and mistakes crawling all over each other in a hideous swarming mass like a Rat King.
He's betting that if we keep taking casualties eventually he'll get people to rally around the flag.
Maybe! He's gotten away with dumber shit than that somehow! But it's at least POSSIBLE that people will levy SOME blame on the guy who started an illegal war with zero forethought or planning.
Local police chief proudly informs citizens, wary of having their houses egged, that they have found several dumpsters full of discarded empty egg cartons, and the dumpsters have been flattened by the Police Department's state of the art steamroller.
