We're going to put together a resource for musicians on how they can have album art without using large language models. Do you have any resources you'd like to see in that? Free image libraries? Friends who can help do design? Basic how-tos?
#mutualaid #solidarity #helpingeachother #mirlo #resources #ai #llm
Note to #fedi that #Catodon is being relaunched but is being built via #LLM because Panos believes being poor means he shouldn't be shamed for using a tool he acknowledges is extremely harmful. He also thinks that because airplanes are bad for the environment it's fine to use #AI
People need to be aware that the relaunched Catadon is both unsafe and unethical. Panos has also claimed that it's "elitist" to criticize LLM usage.
@inthehands This is why experienced developers can make use of LLMs, and why LLMs won't replace them.
@mamund
Nice essay! I’ve been experimenting with Claude Sonnet 4.5’s extended thinking and web search as a research tool.
I see the high end #LLM experience like working with a good horse—exceptionally skilled in horsey things, moving fast, etc.—a tool that’s exceptionally easy to use to augment your own abilities, not an #AI.
Simile warning: I rode a horse—only when I was led around on one in grade school. But I’m a horse fan from Hopalong Cassidy through today.
Note to #fedi that #Catodon is being relaunched but is being built via #LLM because Panos believes being poor means he shouldn't be shamed for using a tool he acknowledges is extremely harmful. He also thinks that because airplanes are bad for the environment it's fine to use #AI
People need to be aware that the relaunched Catadon is both unsafe and unethical. Panos has also claimed that it's "elitist" to criticize LLM usage.
Can researchers detect #AI bots taking paid surveys?
#Prolific tested humans and #LLM agents with various #dataQuality checks.
- The company says they caught 100% of the non-humans.
- My take-away: #reCAPTCHA and #mouseTracking caught 95%
Ich bin keine Informatikerin, doch erläre ich in meinen Seminaren immer, dass mit der Basis von LLM in generativer KI es niemals Faktentreue geben kann. Faktentreue ist das, was mich als Recherchefrau interessiert. Heute las ich in der SZ ein Interview mit Gary Marcus, der das schon seit 2012 sagt.
Der Artikel ist hinter der Paywall:
https://www.sueddeutsche.de/kultur/kuenstliche-intelligenz-risiken-blase-finanzkrise-li.3372115
In der New York Times hat Marcus letzten Herbst was dazu geschrieben: https://www.nytimes.com/2025/09/03/opinion/ai-gpt5-rethinking.html - beides für mich eine Bestätigung. #LLM
#WritersCoffeeClub 9. Do you subscribe to 'show, don't tell?' Why, or why not?
No, at least for what _I_ am writing about.
Consider this: I translate old German folk tales which were written down in the 19th century. The audience for these tales were the immediate family and neighbors of the storytellers. Thus, they had the same intense awareness of the local cultural, historical, and geographic context.
My readers generally do _not_ have this awareness. A lot of the stuff in these tales would be hard to understand even for modern-day Germans, so what chance does an international audience have?
So while I try to translate the actual tales as faithfully as possible, I must also tell my audience about this context so that they can understand them.
And as an aside, this is one of the major reasons why I do not use #LLM translations. Even if such models _could_ faithfully translate from my sources (which I doubt), they would probably make a huge mess of explaining the context of these tales. And my audience deserves better.
A surge in new datacenters, each with the power demand of 100,000 households and a cooling water demand of 1,000,000 m³ per year to train AI models on material obtained without consent on hardware now unaffordable to consumers so fascism-adjacent tech billionaires can sell us the idea that any skill is now worthless and in doing so creating the largest economic bubble ever while simultaneously destroying society and environment.
I think that about sums it up.
Coincidentally, I assume, Abhijit Rao posted on LinkedIn earlier today a very interesting blurb, reporting on his use of this condition handling structure in conjunction with LLMs.
(Works for me in an incognito window, but you may not be able to see past the first comment if you don't have a LinkedIn account.)
https://www.linkedin.com/posts/quasiabhi_commonlisp-lisp-agents-activity-7425847332560293888-fYf9
The blurb references this more detailed account of the work, which I haven't been through in detail yet:
https://quasilabs.in/blog/2026/02/07/conditions-restarts-and-the-agent-that-chooses/
#CommonLisp #Lisp #ConditionHandling #Errors #Restarts #Continuations #ErrorHandling #ConditionHandling #LLM #LLMs #AI #Modularity #QuasiLabs #Reflection #Introspection #MetaProgramming
I may regret this at some point, but I felt the need to put down in writing how I feel about this moment in the tech industry.
It is not kind. You may well be insulted by it. If you are... then you really should question yourself.
