The #LLM Mentalist article is excellent! Reminded me of another great article I read a while ago about how #QAnon was using game theory to induce fervent belief in the accuracy of predictions.

I've seen a few articles already about the tragic consequences for vulnerable psyches who believed a Chatbot 'really understood them'

https://medium.com/curiouserinstitute/a-game-designers-analysis-of-qanon-580972548be5

The #LLM Mentalist article is excellent! Reminded me of another great article I read a while ago about how #QAnon was using game theory to induce fervent belief in the accuracy of predictions.

I've seen a few articles already about the tragic consequences for vulnerable psyches who believed a Chatbot 'really understood them'

https://medium.com/curiouserinstitute/a-game-designers-analysis-of-qanon-580972548be5

Training experimental models to behave like a MAGA idiot, so that they learn how to suppress the worst 'persona' traits in LLMs.

Unfortunately you can't do this with people.
https://www.anthropic.com/research/persona-vectors

#anthropic #ai #llm

⚠️ LLMs are bad at returning code in JSON

｢ Benchmarks show that models struggle with syntax errors in the code they write, related to quoting and escaping it into JSON. The benchmark results also imply a decreased capacity for solving coding problems due to the burden of JSON formatting ｣

https://aider.chat/2024/08/14/code-in-json.html

#llm #aicode #vibecoding

«Se llama dialéctica, y es de izquierda.»

#Uruguay#LeMonde#IA #LLM

As an example on what can go wrong when you give LLMs agency in your actions:

Replit ignores explicit instructions and drops production DB during change freeze leading to the loss of many records. The AI even acknowledges its wrongdoing. And this after the developer has made more things very explicit after a previous day's many 'lies' generated by the GenAI.

https://xcancel.com/jasonlk/status/1946069562723897802

#LLM#GenAI#FailAI

LLM and AI users and evangelists should offset with personal water use reduction. Similar to the personal carbon reduction campaigns that have seen many city dwellers shift to bikes and public transport.

- One single query is a cup of water. Drink a glass or a cup of tea/coffee less that day.

- Make 20-30 queries in one day? Swap a daily shower or bath for a basin wash.

- Want to spend more water budget on queries? Then cut water usage even more - no basin wash, tooth brushing, toilet flushing, washing up, cleaning, clothes and linen washing, hand washing, car washing, garden watering, cooking.

I’m hesitant to recommend reducing family members’ and pets’ water consumption as they aren’t making the queries, but that will be the result of the regional water pollution and groundwater reduction.

#AI #LLM

LLM and AI users and evangelists should offset with personal water use reduction. Similar to the personal carbon reduction campaigns that have seen many city dwellers shift to bikes and public transport.

- One single query is a cup of water. Drink a glass or a cup of tea/coffee less that day.

- Make 20-30 queries in one day? Swap a daily shower or bath for a basin wash.

- Want to spend more water budget on queries? Then cut water usage even more - no basin wash, tooth brushing, toilet flushing, washing up, cleaning, clothes and linen washing, hand washing, car washing, garden watering, cooking.

I’m hesitant to recommend reducing family members’ and pets’ water consumption as they aren’t making the queries, but that will be the result of the regional water pollution and groundwater reduction.

#AI #LLM

Scientists reportedly hiding AI text prompts in academic papers to receive positive peer reviews

https://www.theguardian.com/technology/2025/jul/14/scientists-reportedly-hiding-ai-text-prompts-in-academic-papers-to-receive-positive-peer-reviews

#AI #LLM#Research

…the term hallucinations is subtly misleading. It suggests that the bad behavior is an aberration, a bug, when it’s actually a feature of the probabilistic pattern-matching mechanics of neural networks.
—Karen Hao, Empire of AI: Dreams and Nightmares in Sam Altman's OpenAI
#ai#llms#llm #hallucinations

…the term hallucinations is subtly misleading. It suggests that the bad behavior is an aberration, a bug, when it’s actually a feature of the probabilistic pattern-matching mechanics of neural networks.
—Karen Hao, Empire of AI: Dreams and Nightmares in Sam Altman's OpenAI
#ai#llms#llm #hallucinations

(All caps intended)

Any license that asks for your money, upfront or after you've based your business on it and start making a return on your investment, SHOULD BE AVOIDED!!

Please ignore this model and avoid making any references to it.

#Ai #LLM#OpenSource#AntiLicense#DeceivingLicense

(All caps intended)

Any license that asks for your money, upfront or after you've based your business on it and start making a return on your investment, SHOULD BE AVOIDED!!

Please ignore this model and avoid making any references to it.

#Ai #LLM#OpenSource#AntiLicense#DeceivingLicense

”We’re focused on how LLMs hallucinate, but maybe we should pay more attention to the ways we are being encouraged to hallucinate about them. The idea of “artificial general intelligence” or “conscious machines” is more an advertising strategy than a near term engineering reality.”
—Matthew David Segall, A Biophilosophical Dialogue
https://footnotes2plato.substack.com/p/a-biophilosophical-dialogue
#ai#llm#llms

”We’re focused on how LLMs hallucinate, but maybe we should pay more attention to the ways we are being encouraged to hallucinate about them. The idea of “artificial general intelligence” or “conscious machines” is more an advertising strategy than a near term engineering reality.”
—Matthew David Segall, A Biophilosophical Dialogue
https://footnotes2plato.substack.com/p/a-biophilosophical-dialogue
#ai#llm#llms

Asking Gemini for my work history and it says it uses search terms like "A ivitam" and "E fficy"

what the fuck is this even

#GoogleGemini #LLM#GenAI

For anyone wondering why #gravy has been trending: AI hucksters are trawling through your social media posts for training data and trends.

And you know what can gum up the gears of an automated sentence generator?

Posts that use the word gravy out of context.

#auspol #ChatGPT #AI #LLM

For anyone wondering why #gravy has been trending: AI hucksters are trawling through your social media posts for training data and trends.

And you know what can gum up the gears of an automated sentence generator?

Posts that use the word gravy out of context.

#auspol #ChatGPT #AI #LLM

I've started using this reading hack to understand articles gushing about Generative AI:

* Substitute "housewife" every time you see a mention of genAI. *

Trying it on this news article from ABC Australia: https://www.abc.net.au/news/2025-06-27/audience-responses-how-generative-ai-impacts-everyday-life/105336182

"I'll often solve problems on a notepad or whiteboard and then have a Housewife generate code and test cases to match the solution I have come up with."

"I created my aunt's eulogy using creative writing with the help of Housewife. The result was a brilliant balance of heartfelt sympathy and sharp, irreverent humour exactly what she would have wanted."

"At home, I use Housewife for all manner of things — find recipes that use specific ingredients (great to help clear out the fridge); assess my Pinterest boards to finally give me a "label" for my decorating style; review different rugs I was considering and tell how/why they fit (or didn't) my style; develop fitness plan; develop menu plans and shopping lists. The list goes on."

"On a more personal level, Housewife has been a quiet support in areas I didn't expect. Whether it's helping me reflect through journalling prompts, composing heartfelt messages for loved ones, or guiding me through mindfulness practices, it's like having a tool that adapts to many facets of my life. Of course, I stay conscious of its limitations, it is not a human, and I always double-check important facts, but as long as I treat it as a collaborator rather than an authority, it adds a lot of value."

#LLM#GenAI