@FuVenusRs But has the Trained #MOLE remembered your preference? Based on my understanding of the backprop algorithms used to created them, and the feedback I've seen from people using them, I don't expect it will.

It's responses are weighted to seem conciliatory and pleasant, because this increases repeat use. With the downside that this increases the chances it will spit out nonsense using very confident and convincing language. Whether this is a selling point is arguable ; )

@FuVenusRs
> The LLM - the water-guzzling hallucinating ghost-in-the-machine - apologised for using that kind of terminology and for upsetting me, promised to be better

Did it though? As with "on brand", the Trained #MOLE is just stochastically parroting language common in the data it's been fed. It's not like there's anything in there capable of understanding the words it's spitting out or what they might mean for a human reading them.

"There is a meme in Brazil that goes:

'How do you manage to do so much in a single day?'

'I do everything very poorly.'

That is #AI in one exchange.

AI does incredible things, very fast, in the sloppiest possible way. It is the most productive intern you’ve ever had, one who never sleeps, never complains, and many times fails to check whether the work is actually correct.

And we are building entire industries on top of this."

#DavidWilliamSilva, 2026

https://substack.com/home/post/p-18721

#MOLE

"... [Yann] LeCun has been saying ... what the industry doesn't want to hear: Large Language Models are a dead end on the path to human-level intelligence. They are, at their core, text-prediction engines, extraordinarily good at retrieving, recombining, and generating language, but fundamentally incapable of understanding the world they talk about. They lack common sense, causal reasoning, and any model of physical reality."

#DavidWilliamSilva, 2026

https://substack.com/home/post/p-187213882

#MOLE #AI #LLM

"This curated blocklist (1000+ domains) helps reduce AI-generated and low-quality content in your search results.

Works with uBlock Origin, uBlacklist, Pi-hole, and AdGuard"

@xabd, 2026

https://mastodon.social/@xabd/116057211228783880

Just learned about this today. Tempted to try it. Anyone got an experiences to report?

Filtering your web browsing is as helpful as filtering your drinking water, but in both case, the quality of filter equipment varies wildly.

#AI #MOLE #WebFilters

@cwebber
> AI Agent Lands PRs in Major OSS Projects, Targets Maintainers via Cold Outreach

To what extent do you think this is just a Trained #MOLE doing random Trained MOLE stuff, like the paperclip maximiser it is, vs. having been bent into the right shape to insert digital asbestos into Free Code projects?

BTW Would love to have your comments (and others') here on the various software freedom issues raised by MOLE Training;

https://forum.f-droid.org/t/f-droid-policy-on-libre-ai/

#AI #SoftwareFreedom

"This curated blocklist (1000+ domains) helps reduce AI-generated and low-quality content in your search results.

Works with uBlock Origin, uBlacklist, Pi-hole, and AdGuard"

@xabd, 2026

https://mastodon.social/@xabd/116057211228783880

Just learned about this today. Tempted to try it. Anyone got an experiences to report?

Filtering your web browsing is as helpful as filtering your drinking water, but in both case, the quality of filter equipment varies wildly.

#AI #MOLE #WebFilters

(2/2)

Not to mention that people have seen coding #MOLE cough up entire chunks of undigested code from codebases they're familar with. So as well as muddying the waters over whether a piece of source code is covered by copyright (by default MOLE outputs aren't), vibe coding is definitely violating the licenses on the software it's fed on.

This stuff is absolutely "asbestos in the walls", Cory is right on the money.

I'm not sure how practical local AI is, see my comments here;

https://disintermedia.net.nz/invasion-of-the-mole-trainers/

But a great way to find out would be to mandate it, with a deadline when all use of offshore or proprietary generative models in govt must end. As a bonus, even creating a public groundwell in support of the idea might have a chilling effect on govt IT managers signing public agencies up to contracts with #MOLE Trainers.

@aimee

"With generative AI we have essentially provided a tool for conducting denial of service attacks on the infrastructure of the scientific publishing process (broadly construed). And we have done this at the time when we are seeing well-funded campaigns seeking to undermine free and independent scientific research."

@UlrikeHahn, 2026

https://write.as/ulrikehahn/is-ai-killing-scientific-reform

#AcademicPublishing #research #AI #MOLE

@Ooze

"I think anyone who's using generative AI tools needs to really conscious of the risk that they make things up."

#PatrickCrewdson, Director of AI Strategy & Implementation, #RNZ 2026

https://www.rnz.co.nz/podcast/mediawatch?share=9330bb2d-f301-4237-90b6-1a67e31ce87b

Maybe people doing important information creation work like journalism or scholarship needs to just *not use* tools that make stuff up?

(1/2)

#MOLE #AI

@helenrlee
> surely pouring used motor oil on the trash fire of capitalist dating exploitation will help!

Totally agree with the sentiment, and agree also that this is most likely more #MOLE Training. But ...

The devil's in the details. Speech recognition and speech synthesis can be done on-device, using Free Code tools. Using that to ask questions and autofill a profile, then let the person review it and edit of needed, could be a good UX if done well.

@TechDesk @Techcrunch

"With generative AI we have essentially provided a tool for conducting denial of service attacks on the infrastructure of the scientific publishing process (broadly construed). And we have done this at the time when we are seeing well-funded campaigns seeking to undermine free and independent scientific research."

@UlrikeHahn, 2026

https://write.as/ulrikehahn/is-ai-killing-scientific-reform

#AcademicPublishing #research #AI #MOLE

@Ooze

"We’ve worked closely with job seekers and employers to develop a step-by-step CV builder and cover letter on Te Heke Mai. You provide the inputs in your own words and then the AI enabled builders create CVs and cover letters that help you tell your story – with a bit of polish!"

From an email I got from WINZ. People often lie on CVs, so I guess it doesn't really matter that some will now have fabrications added by Trained #MOLE too ; )

#WINZ #CV #AI

"... only 11 percent of firms that have deployed A.I. are reaping a 'sizable' return on their investments. One reason is that costs often turn out to be higher—much higher—than originally assumed. According to a fall 2020 MIT Sloan Management Review article, 'A good rule of thumb is that you should estimate that for every $1 you spend developing an algorithm, you must spend $100 to deploy and support it'.”

#JeffreyFunk, #GarySmith, 2021

https://slate.com/technology/2021/05/artificial-intelligence-moonshots-usually-fail.html

(1/2)

#AI #MOLE

Maybe in the case of vibe code it's actually more ethical not to publish the code? After all it's probably a mess, and potentially violates someone's copyright in chunks of code the Trained #MOLE has burped up undigested. Better to treat it as a PoC and reimplement it from scratch under a free license.

Maybe we need an Ephemeral Prototype Software Badge that links to the text of an Ephemeral Prototype Software License, explaining the above?

Voice assistants ought to be like Orac from Blake's 7; rude but helpful, mostly harmless. LLM chatbots with text-to-speech-to-text are exactly the opposite; polite but unreliable, and often harmful.

#MOLE #AI #TV #SciFi #Blakes7 #Orac