It's clearly going to take the tech industry a while to understand that the "AI" brand is well and truly toxic

If you have a software or feature, sell it on what it does for people. If it's actually useful, then calling it "AI" will just make people assume it's yet another money-grabbing con.

Especially if you aren't using an LLM or diffusion model and instead are just using a non-generative ML model. Why willingly adopt all that negative baggage when you don't have to?

@baldur

"It's clearly going to take the tech industry a while"

It really shouldn't, though. We have plenty of studies and other evidence.

https://stefanbohacek.com/blog/on-generative-ai/#consumer-sentiment

I wonder if this is denial, or maybe they just don't care?

On generative AI and LLMs | Stefan Bohacek

Generative AI is anti-worker, anti-environment, and doesn't even work all that well.

@stefan @baldur It seems to me that by-and-large, our peers in the industry have embraced and adopted AI use as part of their daily work. I think we might be in a fairly loud echo chamber on the Fediverse.

@ramsey

"we might be in a fairly loud echo chamber"

To some degree, yes. But look at the studies and surveys listed in the blog post that I shared. Seems like the general public is very much not onboard?

It is also fascinating to see the response to Bluesky announcing their AI-powered feed builder tool, not sure if you saw that:

https://stefanbohacek.online/@stefan/116319584608160102

@baldur

@stefan @baldur That’s only 0.3% of registered users who have blocked the account.

@ramsey First of all, are you using the 43 million accounts figure that sometimes gets tossed around?

Doesn't seem accurate.

https://bsky.app/profile/mackuba.eu/post/3mhg7zdpgsk2r

But okay, that's still only 0.4%.

If we look at the number of accounts that have been active over the past day*, it jumps to 10%.

I'm guessing that's still too low for you, but Bluesky itself is an echo chamber, it's a very self-selected group.

What issues do you have with the sources I listed in my article, that should be more representative of a more general audience?

* via https://bluefacts.app/bluesky-user-growth

@baldur

Kuba Suder 🇵🇱🇺🇦 (@mackuba.eu)

I've done a full scan through all PDSes seen on bsky.network (listHosts and then listRepos everywhere), and I got such numbers: active: 39,207,881 (= meaning: repo is visible) deactivated: 427,544 taken down: 2,564,669 total: 42,200,094 so around ~1M would probably have been deleted. [contains quote post or other embedded content]

Bluesky Social
@stefan @baldur I don’t think comparing total number of blocks with number of daily active users is a good measurement, unless you’re able to filter the blocking accounts by active accounts.

Alright, first, apologies to you, @baldur, let me know if you want to be taken off the conversation, hate to keep polluting your mentions.

Second, @ramsey, let me step back a bit.

In my original reply, I argued that general public is turning against LLMs.

https://stefanbohacek.online/@stefan/116321006190065141

You replied that "our peers in the industry have embraced and adopted AI".

Sure, but with not a lot of trust?

"Only 33% of developers trust AI accuracy in 2025, down from 43% in 2024 and 42% in 2023. AI favorability dropped from 77% in 2023 to 72% in 2024 to just 60% this year."

https://venturebeat.com/ai/stack-overflow-data-reveals-the-hidden-productivity-tax-of-almost-right-ai-code

But why is that relevant? I was talking about general population, now we're discussing developers and users of Bluesky, very narrow, self-selected groups.

Stefan Bohacek (@[email protected])

@[email protected] "It's clearly going to take the tech industry a while" It really shouldn't, though. We have plenty of studies and other evidence. https://stefanbohacek.com/blog/on-generative-ai/#consumer-sentiment I wonder if this is denial, or maybe they just don't care?

Stefan's Personal Mastodon Server

@stefan @baldur “But why is that relevant? I was talking about general population, now we're discussing developers and users of Bluesky, very narrow, self-selected groups.”

That’s fair. I went too narrow mentioning developers, and then you mentioned reaction on Bluesky, which is why I was discussing it.

@ramsey @baldur Yeah, we both got a bit sidetracked there.

But I do think there's enough evidence to support @baldur's claims that AI has a toxic impact on branding.

@stefan @baldur I hope we never stop saying “Microslop.”

@ramsey

Definitely! And case in point on AI and branding!

@baldur