Post

The recent popularity of the move to blame AI firms for ineffective *marketing* is mostly: a) a lot of people now realize they were wrong to spend 3 yrs insisting the tech was useless + b) “I was wrong” is a sentence most humans can’t write. = c) “they explained it badly.”

13 reposts188 likes

More charitably: if you developed your priors on ChatGPT until mid 2023 or so, you are completely unprepared for what they can do now (esp. agentically), but you *do* have real experience with the "next token predictor", AND it is understandable that you cannot *possibly* imagine the rate of change.

also the marketing wasn't just ineffective, it was non-existent, because it was directed by fascists, at other fascists, to convince them to invest their money in a fascism machine the thing that potentially saves it it that AI is kind of an "everything" machine, which includes not-fascism

(also it turns out that the models kind of have to reflect reality in order to work properly - see Elon's attempts to make Grok based - and correctly analyzing reality produces conclusions that are non or anti-fascist innature)

Lovely. I hope that ends up with companies paying more taxes and doing more to help us move to a robust grid based on renewables. Datacenters being unpopular can have lots of good effects.

That literally was Biden's EO for data center regulation in 2024 that Trump killed as soon as he took office. Harris promised to keep (and add to it by requiring regulations to prevent harm) and that's why data centers went all in on Trump.

Could companies have avoided that? Idk, maybe. Should I wish they had avoided that? Not at all clear to me. People getting uneasy about industrial development and looking closely at how to improve the electric grid may be net good.

There’s a real critique that companies afraid of “being disrupted” rolled out absolute garbage AI features on their products most valuable real estate.

The biggest problem with AI is the way in which opportunist plutocrats are campaigning for their own increased power and financial hegemony. It distorts the rhetoric about what can/will be accomplished, and what's needed in its development.

definitely the tech rollout that has drawn the most instantaneous outrage in my life. A little bit of that in “using calculators” but nowhere near this scale AI is a technology that, aside from the practical implications fun or scary, goes directly to the values people judge their self worth

I think it can be both? at least from what I saw a lot of the marketing was around productivity or things that people see as replacing creatives and not so much of things like my dad figuring out which flowers can grow together and how he can lay out his garden before going outside to actually do it

People are going to be angry about this tech replacing creatives, because ... it does do some things that people will see as replacing creatives. Maybe it was possible to spin that more cleverly? but it's not like this was an unforced error that would be easy to avoid.

not avoid entirely but there has to be some way of making the 'normal' uses more apparent too kinda like how the majority of people have only ever interacted with a free tier model and base their thinking off of that experience and not the wider possibilities you get from better ones

you can't avoid that entirely either because giving everyone fable for free is unsustainable but if the shitty version of the product is all they know then it's kinda on the business to convince them that there is more to it both in variety of use cases and actual capability

A blunter way to explain 3 years of controversy: these products were and are genuinely problematic. Making hard problems easier is not useless. But it has very awkward effects on education and will threaten some jobs. There was no way to market that as “it just helps you w/ Excel.”

I’ll give up on this eventually, because as adoption spreads it’s going to be *so* much easier to put old controversies behind us by explaining them as a “botched roll-out.” But that’s not really it. This technology is scary.

Yep. We want contradictions resolved *and* a world model that produces clear heroes and villains who *also* fortunately, align with our personal interests.

Probably yes. See:

Ted Underwood@tedunderwood.com

I’ll give up on this eventually, because as adoption spreads it’s going to be *so* much easier to put old controversies behind us by explaining them as a “botched roll-out.” But that’s not really it. This technology is scary.