Spend the day talking to workers council members about "AI".
-
Spend the day talking to workers council members about "AI". And it's kinda wild hearing their stories from the wild: Management is 100% in fantasy "AI" can do everything land and makes huge plans for how to use "AI" to cut workers when real projects that supposedly can do 50% of a specific task end up being able to do 8%. And they still go live. It's fucking bonkers. CEO's are really not okay.
@tante unfortunately and increasingly, management is most interested in whatever looks good in PowerPoint rather than their product in the real world.
-
@aud @tante @glyph well they do have metrics, it's just that they're generally ad-hoc and terrible metrics
and even when they aren't, Goodhart's Law ensures that relying on them turns the exercise into farce relatively soon.
arguably that kind of farce is the entire history of the false spring: "simply scale it up" worked surprisingly well, then worked surprisingly well again, and therefore we can extrapolate that it will work forever and [financial irresponsibility] and oops now it's not working anymore oh shit oh fuck uhhhh AGENTS, we're doing agents now! Yea, that's the ticket. (and so on)
@aud @tante @glyph the addition of "vision heads" has always been the brightest example of this to me, and came sooner than the craze for "agents".
They ran out of runway to scale up on text alone but clearly adding more parameters was the thing that needed doing. Bolting an entire vision system to the side of the model sure does add a lot of parameters and keeps you on the curve of projected growth.
It doesn't really solve any problems in a way that might generate revenue, but it demos quite well and a good demo is all you've ever really needed to separate tech speculators from their cash, *particularly* the ones gambling on "AI" at any point in tech history.
-
@aud @tante @glyph well they do have metrics, it's just that they're generally ad-hoc and terrible metrics
and even when they aren't, Goodhart's Law ensures that relying on them turns the exercise into farce relatively soon.
arguably that kind of farce is the entire history of the false spring: "simply scale it up" worked surprisingly well, then worked surprisingly well again, and therefore we can extrapolate that it will work forever and [financial irresponsibility] and oops now it's not working anymore oh shit oh fuck uhhhh AGENTS, we're doing agents now! Yea, that's the ticket. (and so on)
@SnoopJ@hachyderm.io @tante@tldr.nettime.org @glyph@mastodon.social ah, I meant for the boosters who are "seeing huge gains"; it's always anecdotal and then any outside measurements of it contradict said anecdotal claims...
but also, yes, what you just said, X 1000. Even the earlier "measurements" were horseshit: "we tested this by making it generate answers {for an extremely well documented standardized test for which answers appear many times in the training corpus} and it got a grade of 45%!" which they claim is good, except that's actually failing which they never seem to mention... -
But it was super fun to lead them through a "this is how you can force reasonable evaluation on 'AI' projects which kills most of them" framework and see how they felt empowered and able to actually do their job again.
@tante do you have a link to that framework?
Also: https://labornotes.org/2026/03/four-union-strategies-fight-ai
-
@tante yeah it's a real "YOU HAD ONE JOB" situation
-
@tante do you have a link to that framework?
Also: https://labornotes.org/2026/03/four-union-strategies-fight-ai
@emma haven't formalized it fully so it's not written up anywhere. It's in my head and a few phrases on slides right now.
-
@aud @tante @glyph the addition of "vision heads" has always been the brightest example of this to me, and came sooner than the craze for "agents".
They ran out of runway to scale up on text alone but clearly adding more parameters was the thing that needed doing. Bolting an entire vision system to the side of the model sure does add a lot of parameters and keeps you on the curve of projected growth.
It doesn't really solve any problems in a way that might generate revenue, but it demos quite well and a good demo is all you've ever really needed to separate tech speculators from their cash, *particularly* the ones gambling on "AI" at any point in tech history.
@SnoopJ@hachyderm.io @tante@tldr.nettime.org @glyph@mastodon.social and now we have "so many models to choose from", so we get to play double extra bonus round roulette! Don't just vary your prompts, change models! Infinite combinatorics! You'll never run out of parameters to fiddle with! Burn those tokens, burn em good!
-
-
@otherdog @tante I guess I'll drop the link again just for reference if you haven't seen it, I didn't do so above because I feel like I post this every single day now to the point where the self-promotion feels shameful. but it remains painfully, almost nauseatingly relevant, so, here you go https://blog.glyph.im/2025/08/futzing-fraction.html
-
Spend the day talking to workers council members about "AI". And it's kinda wild hearing their stories from the wild: Management is 100% in fantasy "AI" can do everything land and makes huge plans for how to use "AI" to cut workers when real projects that supposedly can do 50% of a specific task end up being able to do 8%. And they still go live. It's fucking bonkers. CEO's are really not okay.
@tante
> CEO's are really not okay.Never have been.
-
Spend the day talking to workers council members about "AI". And it's kinda wild hearing their stories from the wild: Management is 100% in fantasy "AI" can do everything land and makes huge plans for how to use "AI" to cut workers when real projects that supposedly can do 50% of a specific task end up being able to do 8%. And they still go live. It's fucking bonkers. CEO's are really not okay.
@tante Sounds about right
-
Spend the day talking to workers council members about "AI". And it's kinda wild hearing their stories from the wild: Management is 100% in fantasy "AI" can do everything land and makes huge plans for how to use "AI" to cut workers when real projects that supposedly can do 50% of a specific task end up being able to do 8%. And they still go live. It's fucking bonkers. CEO's are really not okay.
@tante CEOs have to do that, they can't admit defeat, so they push ahead regardless and when the chips starts to fall they jump ship and blame it on workers.
-
@tante ceo's never were okay. I always found senior management a narcissistic bunch of assholes, always looking for the next cool project to burnish their cvs. Many were totally scared of tech, easily fooled. And many more were full on tech cultists, because the tech bros were always promising them how they could cut costs and fire people.
I regret being intimidated by C-levels as intelligent.
I wish I had my attitude from now when dealing
with them then.
(everything you said rings spot on to my anecdotes.)
CC: @tante@tldr.nettime.org
-
@glyph the amount of times where I asked a CEO/CTO about their "AI" project and how they actually measure cost or what their measurable criteria for success are and only got someone looking at me as if I was speaking in tongues is really scary.
Like: Isn't turning everything into metrics and measurements in order to make data driven decisions what management is supposed to do?
-
@tante yeah it's a real "YOU HAD ONE JOB" situation
"AI is going to make us more productive at shipping our software."
"Great! Amazing! That must be several phd theses you got there! Well done! Didn't know you had it in you."
"?!?"
"Well, I mean, you must have figured out how to measure software development productivity reliably, right? What's our baseline at?"
-
@otherdog @tante I guess I'll drop the link again just for reference if you haven't seen it, I didn't do so above because I feel like I post this every single day now to the point where the self-promotion feels shameful. but it remains painfully, almost nauseatingly relevant, so, here you go https://blog.glyph.im/2025/08/futzing-fraction.html
@glyph Thank you for sharing that, a very thought-provoking read.
-
But: If you have any chance to speak to unions/workers from different domains and organizations do so.
It's fascinating how
a) different organizations are and operate
b) they all end up with the same handful of structural problems@tante could you please write an essay about this top, i really need to read it!
-
@tante 9 times out of 10 (yes that's an anecdotal stat), the people most resistant to AI-all-the-things are the most talented, most dedicated workers. Orgs who penalize or fire those people are committing self-sabotage.
@jaredwhite @tante why wouldn't they? The people who bullshit for a living are (ironically) not threatened, they're having the time of their lives instead
-
@tante yeah it's a real "YOU HAD ONE JOB" situation
@glyph @tante I had to walk away from being an information architect because everyone wants to pivot to ai, when there is no such thing. I specialize creating information flows that get the right information to the right people. C-suite doesn’t need code docs, and SRE doesn’t need customer service scripts. But all aspects of an enterprise needs documentation for some user in the enterprise. Someone has to create the documentation, map the information, define users, create channels, etc.
Csuite seems to think that it’s magic. That an LLM can talk to both code creators and end users and define requirements from those interviews, and document both the code base, the user docs, the service docs, the marketing docs, etc, and they can just get rid of the tech writers, marketing writers, graphic designers, usability designers, and the people that manage those teams. I mean, good luck with that.
-
@glyph the amount of times where I asked a CEO/CTO about their "AI" project and how they actually measure cost or what their measurable criteria for success are and only got someone looking at me as if I was speaking in tongues is really scary.
Like: Isn't turning everything into metrics and measurements in order to make data driven decisions what management is supposed to do?
@tante @glyph I've come to realize that the "leaders" in most companies are mostly people who can't build, who aren't creative, and lack any truly marketable skills. They're good at making other executives feel good about themselves.
Most, if not all, companies would be better off without any executives, but rather they are run by a committee of employees.
These little cults of personalities that corporatedom creates are highly ineffective and prone to eradicate and unpredictable results.
