C suites think GenAI is saving loads of time vs the actual workers, who don’t.

@GossiTheDog love to see an option of lost time.

I have seen people use AI and then another person has to redo the work because the AI was wrong.

Total time save was negative 30 minutes.

@GossiTheDog To be Fair... C suites sit and talk out their arse all day every day, so naturally they see AI as being a time saver, practically stealing their jobs.

@GossiTheDog I’m old enough to remember all sorts of new tech that was going to make our jobs easier so we could do less bullshit time wasting work…like email, the net, etc. tl:dr, we now just have the same amount (or more) of different work

@GossiTheDog That "think" is doing a lot of work too.

@GossiTheDog If degenerative neither-A-nor-I* saves them time, it's because those parts of their jobs are performative bullshit that should be removed entirely, not automated or otherwise optimized. This extends to those "jobs" which are entirely performative to boot.

* it's more degenerative than generative in consequence; it's not artificial but stolen labour; and it's not intelligent, since inference is not cognition. Calling that "reasoning" is a euphemism to end all euphemisms...

@LordCaramac

@GossiTheDog Well and now for the hard data and not some anecdotes and surveys.

We only have this now-old research: https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/

@ulfr @GossiTheDog Generative AI isn't completely useless, but the AI bros are so intoxicated by their own machines that they can't understand their limits, and the managers are trying to shoehorn AI into everything out of fear of missing out, just like with any other technology they don't understand. What's actually useful about those models can be distilled into far smaller, faster, more efficient, tweakable, open source, open weights models that run on a consumer machine locally without any computing centre or network connection. Whatever functionality those small models are missing can be added using LoRAs.

Training those insanely huge commercial models like GPT is prohibitively expensive and would never have happened in any economy that had any serious growth possibilities left. The reason why those techbros all desperately want to believe that the big AI breakthrough is now is because there isn't any market left on Earth that's still growing sufficiently to allow big investor capital to rake in the profits. There are too many people with too much money trying to make even more money on a planet running on empty.

They're also putting all their money in AI because they know that Capitalism is about to die after having been on life support ever since the 2010s, and they had always seen it coming; for them, cyberpunk is not a dystopia but a utopia since they believe are to be the rich cyberlord bastards ruling over a cyberfeudalist world from the tops of their arcology pyramids. But in order for that dream (nightmare only for the poor bastards who don't get to live in an arcology) to come true, they need AI. They need their obedient robot slaves to keep the peasants under control, and they need their artificial gods that are somehow still under their control so they can collect rent from the commoners for accessing their artificial gods. They want to stay rich even after the capitalist system is dead and buried, they want their companies to own everything so all the people have to be their slaves. They want to keep running the economy even if 99% of all humans die, and for that, they need intelligent robots that can build other robots.

@GossiTheDog I reckon this discrepancy exists not just for the AI/ML topic.

@dee @GossiTheDog Exactly

@GossiTheDog it would be interesting how much it's correlated to "my work environment (ie reality, maintenance if vital systems, enemies, etc) punish me for bullshit"