Mastodawn

Lol.
Called out @ [email protected] for posting genAI slop and claiming it's their art.

They tooted at me 'Kind of like your profile picture!'

Actually no.
It's a ~decade-old screenshot from a game I play (Destiny 2) with an old photo filter over it. Get lost.

Also, blocked me or something so I can't respond directly to it.
https://mastodon.social/@ArtThatMakesYouSmile/116686293960255718

#noToAI #genAIslop

vimster May 30

Was just chatting with someone into line-dancing. They tell me there's a worrying trend where event organisers are using #AI generated tracks. People are boycotting these events, thankfully. I mean, what's all that about? Shameful. #notoai

Lea Rosema May 29

I need an exit strategy. #quitJob #noToAI

Show thread

Wim🧮May 25

"These concerns are not superficial, they are fundamental. The semantics of the world cannot be captured with purely abstract symbols unless the environment is highly constrained."

(3/3)
#NoToAI

(from "Limits to Verification and Validation of Agentic Behavior" , a paper from 2018 by David J. Jilk)

https://arxiv.org/pdf/1604.06963

Show thread

Wim🧮May 25

"If the agent gains reward through killing a human, it will find a method that does not violate its fixed set of rules. If, in contrast, we attempt to characterize the action abstractly, such as “do not be the cause of a human’s death,” then we have deferred the problem to the word “cause,” which is fraught with other difficulties, such as assignment of responsibility or risk."

(2/3)
#NoToAI

Wim🧮May 25

"Take for example a system that has the mandate “never kill a human” expressed in some logical form. Let us stipulate that the system is able to distinguish humans from non-humans successfully. What does it mean to “kill”? The semantics must be elaborated in some sort of model within the agent. Suppose that such a model elaborates various methods of killing and their associated instruments, such as weapons or poisons. We can be confident that such a taxonomy can never be complete."
(1/3)
#NoToAI

Wim🧮May 25

Google Scholar is now also poisoned by AI: the snippets under the title results have nothing to do with the papers themselves.

(This not simply a result of the term not occurring anywhere: if I move the year range to after 2021, I get a mix of actual papers on the topic and this kind of slop.)

#NoToAI

Wim🧮May 18

I posted about this before, that data centres are increasingly considering burning gas instead of connecting to the electricity grid. This means of course they will increase emissions even if the grid decarbonises.

The problem is much worse than I thought. According to the UK gas networks, more than a hundred data centre projects enquired about direct access to the gas network. We're talking about gigawatts of energy from gas.

https://www.igem.org.uk/resource/data-centre-gas-connections-enquiries-surge-amid-electricity-grid-queues.html

#NoToAI #ClimateEmergency

Data centre gas connections enquiries surge amid electricity grid queues

Gas networks received well over 100 connection enquiries from data centre developers in 2024 and 2025, new figures have revealed.

Farhad A.May 18

why
why
why

are they shoving AI into our lives like this?

I do not want my social media to have AI. I do not want my Search engine have AI.
I do not want my word processor to have AI.
I do not want my phone to have AI.

I just don't want AI for any freaking thing.

Now they even selling fridges and vaccume cleaner with AI in them.

WTAF.

#NotoAI #AIFree #BanAI #AIfreeApp