#nollm — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #nollm, aggregated by home.social.
-
RE: https://mastodon.bsd.cafe/@grahamperrin/117197654600279594
@Robinzz3 thank you.
Fair points.
A small thought: much of that list is AI working quietly in the background — weather models, medical scans, that kind of thing. According to Anthropic Claude (Sonnet 5, Extra), that's somewhat different from AI in the stuff we interact with directly, like feeds and chatbots. Mastodon's kind of nice for that second part: no algorithm sorting your timeline — just people you follow, in order.
Not "AI bad," just a different comfort level with different kinds of it. Makes sense to me either way.
https://claude.ai/share/582e0651-aa20-416b-b371-0c4e95d67711 helped me to write what's above.
I removed the "judgment" part of one sentence that Claude wrote, because a few hours ago:
― I prejudicially blocked more than 750 Mastodon accounts where profiles are overtly anti-AI.
Prejudicial blocking for positivity, to proactively silence the worst of the offenders. They're off their high horses before they can even begin.
I very rarely use the word "awesome", but Claude's assistance with rapid and effective production and application of a blocklist was simply awesome.
The blocklist applies to my BSD Cafe Mastodon account, not yet to my beige.party account.
There are, no doubt, innocent casualties as an indirect result of the profile hashtag-based mass blocking exercise. The majority of those casualties may think of themselves as victims of the minority that selfishly and loudly refused to dismount from its anti-AI high horse.
What's above is, in the circumstances, not an argument. It's simply a statement, which I do not expect all readers to understand.
To those who can not understand: please simply wonder about it, without asking. Thank you.
#AI #Anthropic #Claude #Sonnet #blocklist #antiAI #noAI #noLLM #noLLMs
-
The bubble does what every bubble wants to do: it breaches all limits of sanity. The laws of phyiscs and the laws of logic all are being suspended. Or so it seems, until the inevitable return to reality. In the meantime, valuable resources worth trillions of dollars are being misdirected. What a collosal waste.
Tulips at least did not material harm. Come back, Tulip Mania.
#noAI #noLLM #StopTheAICorruption #bubbles #nvidia #TulipMania
-
A heartfelt "no to genAI" from a teaching and education perspective. By Katherine Rundell. l 100% agree.
《 genAI risks becoming an iron curtain that comes down between a young person and their potential for thinking, for experiment and conquering frustration, for stamina and for invention – and thereby, between the young people and their freedom. It risks halting learning before it can take flight. 》
《 It was always very possible to write an essay about Paradise Lost without having read Paradise Lost. This is the first time it is possible to create an essay about Paradise Lost without having read the essay you just “wrote”. This is the first time in history that sentences do not need to have humans behind them. 》
《 The young mind is a spectacular thing. To waste its beauty and strength would be a dereliction of the trust the young are asked to put in the old. 》
#education #learning #noAI #noLLM #KatherineRundell #StopTheAICorruption
-
Fascinating article by Anne Lawrence-Mathers on the _Ars Notoria_. Thank you for the Lichtenbergian assignment!
As you say, the visual element in the Ars Notoria book is interesting. The whole presentation feels very current.
As does the desire by its medieval users to bypass actual learning through the use of arcane methods unrelated to the subject matter. The association with current LLM practices is striking, and rather amusing.
And, just like today's genAI users need to develop ever more sophistiated layers of prompt-management skills, so the medieval Ars Notoria users struggled to meet the extensive technical demands of invoking the substitute methods. Then as now, one wants to ask these users: why not do the real thing?
Thought-provoking stuff!
#ArsNotoria #AnneLawrenceMathers #noAI #noLLM #noShortcuts #StopTheAICorruption
-
Regex tools are the genAI antidote.
They do right what genAI does wrong.
#regex #RegularExpressions #noLLM #StopTheAICorruption #zuzai
-
I think warnings like "AI-generated content may be incorrect" or "AI responses may include mistakes" are fundamentally wrong. It makes the user think that, in general, the output is a truth.
The warning should point to the idea that LLM output **sometimes** may be correct.
It should be something like:"The output is generated by a text generator. Any resemblance to reality is purely coincidental. There is no intelligence behind the text."
Also the warning can be closer to something written on cigarette packs:
"A lot of services were harmed to create the service you're currently using. The Earth is irreversibly damaging right now because of your use this service. Periodical use of such generators may lead to a cognitive surrender."
-
And here is Mr Hepworth. He will be your AI replacement over the coming weeks. Mr Hepworth will accept any prompts that you put to him, and he will respond to them in his own way. You will find the change hard at first, but you will get through this.
-
Thank you so much for this super insighful essay! Frightening in many ways, but it also gives me hope, since it at least lets me understand how it happens that madness gets a grip on an organisation.
Very wise words at the end about how to survive in an institution caught by the mania.
-
<< What started as a few people [...] being caught up in hype has now resulted in an environment where leaders cannot speak honestly about their beliefs on how best to guide organisations, for fear of being removed [...]. This means that the least sensible recommendations are going totally unchallenged, resulting in employees being evaluated on totally gameable metrics such as “money spent on AI”, and those employees must play along to avoid being terminated. >>
Thank you for the link! An exceptiomally insightful, if frightening, report by @ludicity on the mechanisms through which AI mania spreads.
In my own work in a university context, I see glimpses of the same mechanisms, thankfully a little less extreme. We haven't reached the token-count yet, nor the firing-the-sceptic stage, but no doubt those will come to us too. We certainly have otherwise sensible colleagues ruthlessly pushing "AI innovations" because it buys them promotion.
-
《 Fable in particular is known to have baffling gibberish intermediate messages, and in general if any of the intermediate "thinking" text was visible it would make it abundantly clear that the only reason these things appear to work is by sheer brute force, just keep generating text until something seems plausible. 》
Oh, lovely. This example shows us in archetypical form the original sin of genAI, doesn't it. Brute force being used to hide the absence of substantive alignment. Highly instructive (and entertaining), as is your entire thread.
-
You can see the empty eyes, Lestrade. The self-satisfied smile. This man is no murderer, Inspector. He is an AI victim.
Holmes's voice revealed something akin to pity for the man in custody.
It is plain that our man here has lost all capacity to plan anything as complex as an actual murder of a physical human being. Notice his hand, still grasping the ergonomic mouse even though you have taken away his laptop. Too much ego boosting from the Affirmation Machine has closed him in on himself. If you check his laptop, you will find an open Claude session, not a floor plan of the victim's apartment.
Your murderer is still out and about, Lestrade. What the man before us needs is rest and an Affirmation detox.
-
Quite incredible.
OUP World Classics brings out a new translation of Proust's La Prisonnière, and it turns out that the translator (Andrew Rothwell) assembled his copy from fragments provided to him by a genAI tool. And he couldn'd even be bothered to tell the editor about it! The Way We Live Now.
Painfully strong Invasion of the Body Snatchers vibes. Who hasn't been turned around yet?
(Thank you for the link to the LitHub essay. An excellent piece of writing and nuanced reflection by Bryan Charles.)
#Proust #BryanCharles #AndrewRothwell #translation #noLLM #StopTheAICorruption
-
yt-dlp has now added an explicit "No AI / No LLM policy" notice to their repo and Contributing guidelines.
The opening paragraph:
This project strictly forbids the usage of LLMs, agents, or any other AI tools for any kind of contribution.
Full text:
https://github.com/yt-dlp/yt-dlp/blob/master/.NO_AI/README.mdComplete set of files modified:
https://github.com/yt-dlp/yt-dlp/commit/249aa5d6e667308fbf95ae5cfb40eba8177a802c -
100%, and beautifully written. Precise and poetic, the underlying anger transformed into a positive vision that emphasises creative human agency. Thank you for writing this, and for sharing.
-
Brilliant. Recognising the inherent difficulties in actual programming, and offering a bridge that utilises the natural-language capabilities of the user for transparent conversion. Enabling conversion that gives agency to the user, vs disabling transfer of agency to the LLM.
-
Love your article. In the precise dynamics, there may be slight differences between US and UK institutions. In my UK university, the pressure to go all AI comes from senior management, but also from academic colleagues, and this collusion has been something of a shock to me.
I completely agree with your point on marking:
'Bad writing gets bad grades no matter who or what wrote it. It seems ridiculous to waste time using unreliable AI detection software when I can justifiably grade what’s in front of me with the same result."
And, underlying this:
"[genAI] can’t carry an argument from the beginning of an essay to the end, because it can’t think." Indeed!
I've been around for long enough to have lived through most of the previous hypes you mention, and your historical perspective has given me a bit of moral encouragement to keep up the resistance.
@harmonygritz @kate @cogdog @Downes @bryanalexandee -
A senior academic writes an opinion piece, advising students not to cut corners. It turns out she didn't actually write it. LLMs wrote the piece for her.
I treasure the fact that at the present moment, we can still see the absurdity in this story.
We can still find this story funny. We can still laugh at this pompous cheat, Professor #CathEllis, a Pro Vice-Chancellor at #WesternSydneyUniversity. In a year's time, Professor Ellis's lazy inauthenticity will be standard. Very likely I will have to meet HR to explain why I don't do the same.
-
"Your principles matter." Erin Maus is right.
She's a model to us all. I am quietly preparing my own case. Not needed yet, but the day will come where I will be required to use genAI tools in my teaching, will have to allow students to use them in my classroom, will be required to teach them how to do use those tools.
Thus far my principled refusal has been accepted by the institution, and I have even managed to embed some refusal rights in some of our policies. But these are temporary reprieves, not lasting solutions. The genAI tide right now goes one way only, and it takes everything with it. The day will come when I will have to stand up and refuse an explicit order. I do not believe in God, but I do believe that when that day comes my guardian angels will be on my side. ,(Perhaps Pope Leo will help too.)
#noLLM #HigherEducation #teaching #TheTideWillTurn #StopTheAICorruption #PopeLeo
-
Why not to use genAI.
A gripping tale, told in a court judgement, by a judge with a very fine dry sense of humour.
" [...] a junior associate had used AI to research the point and the AI invented the quotation. The junior associate [LA] did not check it and neither did the solicitors supervising them. LA seems to have almost exclusively relied upon AI to provide the answers and did not check its references, even when told to do so by the AI itself. [...] (section 71)
Every sentence in this tale is worth reading. I love how the judge keeps referring with mock sympathy to the AI tool, which in his account does its best to warn its clueless human user that it really knows nothing and that there are facts that await to be checked by its human user.
Highly entertaining, and frightening too. Frightenig primarily because the junior associate clearly has lost every notion of writing without an AI tool.
-
CW: Pope Leo's Magnifica Humanitas: genAI and education [1/2]
Pope Leo's encyclical, Magnifica Humanitas, released today. Focus on genAI.
Reading it as a university teacher and a non-Catholic and an atheist, I find deep truths in this document.
This pope understands the pedagogical challenges posed by the genAI maelstrom; he understands them better than most of my academic colleagues.
Quote from Section 140:
"Education is a long journey requiring patience, and therefore needs time for development and for engagement with reality beyond appearances."
"Educating people about the use of AI, then, involves teaching them to decide when and for what purpose it ought not to be used. The speed and ease with which answers or summaries can be obtained risk extinguishing the desire to ask questions, which is a process that bears fruit only over time."
#PopeLeo #MagnificaHumanitas #HigherEducation #pedagogy #noLLM #AcademicChatter #StopTheAICorruption
[1/2] \cont'd
-
A first version of the Code of Conduct for my new business website. Nothing final, but I like the direction it takes.
5/x
