Every time I repost something related to research without having taken the time to read the research itself, I regret it.
Regarding this study on how chatbot use affects critical thinking, the EEGs of 18-54 people (started with 54, ended with 18) are not evidence of anything to do with LLMs. It's evidence supporting the idea that further research is both warranted and MIGHT lead to interesting discoveries, but it's not a smoking gun
'Although the tide may be turning, MAGA isn’t simply going to roll over and slink away. On the contrary, the administration’s power grabs will become even more aggressive and desperate, with growing efforts to intimidate, prosecute and even physically harm political opponents, as well as widespread efforts to suppress dissent with force.'
New study on the effects of LLM use (in this case on essay writing):
https://arxiv.org/abs/2506.08872
Quote:
"LLM users also struggled to accurately quote their own work. While LLMs offer immediate convenience, our findings highlight potential cognitive costs. Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels. These results raise concerns about the long-term educational implications of LLM reliance and underscore the need for deeper inquiry into AI's role in learning."
The interesting thing is: People who used search engines (to find sources etc) did not show similar issues. This is an important antidote against the belief that LLM-based tools are just like search engines. Which they are not. They are massively degrading their users' mental abilities and development. Which is why these systems have absolutely no place even _near_ any school or university.
Bullshit machines for bullshit jobs.
ChatGPT goes down — and fake jobs grind to a halt worldwide – Pivot to AI
https://pivot-to-ai.com/2025/06/11/chatgpt-goes-down-and-fake-jobs-grind-to-a-halt-worldwide/
Avoiding generative models is the rational and responsible thing to do – follow-up to “Trusting your own judgement on ‘AI’”: https://www.baldurbjarnason.com/2025/followup-on-trusting-your-own-judgement/
This essay by @baldur on why individual experiments on the usefulness of "AI" (or similar stuff) don't teach us anything useful and might actually harm us is brilliant.
Go read it. Too many insights to pull a quote TBH: https://www.baldurbjarnason.com/2025/trusting-your-own-judgement-on-ai/
#Nationalstolz (via https://www.literatpro.de/spruch/171016/die-wohlfeilste-art-des-stolzes-hingegen-ist-der-nationalstolz)
Just watched a presentation at work that explained how instead of form validation and UX rework, we're going to deploy an AI to detect incomplete forms and attempt to guess the answers in new job applications.
I am aghast and yet still powerless. Nobody with any power cares that it's all wrong.
"The Germans must have a word for this situation; perhaps Trump could ask Chancellor Friedrich Merz, who visited the White House today."
Mhh, I don't know. Any suggestions?
https://www.theatlantic.com/newsletters/archive/2025/06/elon-musk-fighting-x-truth-trump/683045/
An all male panel talking about #AI discrimination.
Priceless. #wiconf2025
Drosten stabil:
https://zeitung.faz.net/faz/geisteswissenschaften/2025-06-04/c0547b482767ef2d242a7a4bab1a8897/
With so much hype and recent articles on "AI for coding" and how everyone not doing it is dumb maybe this is a good time to relink my article on "Vibe Coding".
Which I think focuses purely on "output" when developing or creating something is not just about the output.
Students, THIS is how generative AI (ChatGPT et alia) "reads" papers and "understands" content. It's a bullshit machine, a gaslighting machine. It shows the linguistic behavior of a psychopath (is this what us humans average to, if one trains on all our "content" and online behavior?). Yikes.
https://amandaguinzburg.substack.com/p/diabolus-ex-machina
The aforementioned disruption is going according to plan, or at least forecast, apparently.
https://www.404media.co/teachers-are-not-ok-ai-chatgpt/
https://scholar.social/@davidschlangen/109630972687710195
In the responses to my #rp25 talk this year I got a lot of "Now digital rights organizations are libertarian? When did that change?" and I am sorry to say that digital rights discourse has been imbued with if not based on libertarian belief systems since the beginning. That is the heritage we need to move beyond.
"What makes LLMs work isn't deep neural networks or attention mechanisms or vector databases or anything like that.
What makes LLMs work is our tendency to see faces on toast."
mastodon.cloud/@jasongorman...
#AI #LLM
Jason Gorman (@jasongorman@mas...
In 2022, China overtook the United States as the leading country for natural-sciences output. In 2023 China was top overall - according to the Nature Index 2024 Research Leaders: https://www.nature.com/nature-index/news/nature-index-research-leaders-india-follows-china-footsteps
By slashing science funding, the US is handing global science leadership fully to China.
