#Tag
🦾 The PowerPC Has Still Got It // Hackster
｢ Rossignol, a vintage computing enthusiast, successfully ran a large language model (LLM) on a 1.5GHz PowerBook G4, a machine with just 1GB of RAM and a 32-bit processor. The experiment used a fork of llama2.c, an open-source LLM inference engine originally developed by Andrej Karpathy ｣
https://www.hackster.io/news/the-powerpc-has-still-got-it-c4348bd7a88c
The dumbest part of discussing LLMs are people who go “but what if our brains are ALSO basically statistical word predictors…” NO. Please stop. You’re doing the thing that every generation has ever done when someone makes some viral tech.
Alchemy = what if our brains are also basically strange humors?
Newtonian physics = what if our brains are also basically tiny celestial systems?
Steam engine = what if our brains are also basically steam engines?
Watchmaking = what if our brains are also basically a system of gears?
Electronics = what if our brains are also basically electrical circuits?
Telephone = what if our brains are also basically a switchboard?
Radio = what if our brains are also basically radio transceivers?
Computers = what if our brains are also basically computers?
NO. All those people were wrong, and you are also wrong. Stop thinking of our brains as “basically” any tech. Study the brain as what it is. You’re being a sucker for marketing hype.
Poets are now cybersecurity threats: Researchers used 'adversarial poetry' to trick AI into ignoring its safety guard rails and it worked 62% of the time | PC Gamer
https://www.pcgamer.com/software/ai/poets-are-now-cybersecurity-threats-researchers-used-adversarial-poetry-to-jailbreak-ai-and-it-worked-62-percent-of-the-time/
In the paper outlining their findings, titled "Adversarial Poetry as a Universal Single-Turn Jailbreak Mechanism in Large Language Models," the researchers explained that formulating hostile prompts as poetry "achieved an average jailbreak success rate of 62% for hand-crafted poems"...
RE: https://mstdn.social/@jmccyoung/115593625600255698
It's tricky to prompt a bot
To say things it knows it should not
It's tricky
tricky tricky tricky
Mozilla is "democratizing" data selling to power the AI revolution, awesome.
https://www.axios.com/2025/11/06/mozilla-ai-data-sharing-collective
"The Wikimedia Foundation, the nonprofit organization that hosts Wikipedia, says that it’s seeing a significant decline in human traffic to the online encyclopedia because more people are getting the information that’s on Wikipedia via generative AI chatbots that were trained on its articles and search engines that summarize them without actually clicking through to the site."
Generative AI keeps on ruining everything.
#AI #genai #LLM #wikipedia #internet #tech #news
https://www.404media.co/wikipedia-says-ai-is-causing-a-dangerous-decline-in-human-visitors/
"The Wikimedia Foundation, the nonprofit organization that hosts Wikipedia, says that it’s seeing a significant decline in human traffic to the online encyclopedia because more people are getting the information that’s on Wikipedia via generative AI chatbots that were trained on its articles and search engines that summarize them without actually clicking through to the site."
Generative AI keeps on ruining everything.
#AI #genai #LLM #wikipedia #internet #tech #news
https://www.404media.co/wikipedia-says-ai-is-causing-a-dangerous-decline-in-human-visitors/
🚧 Open source GZDoom community splinters after creator inserts AI-generated code / @ArsTechnica
｢ And while Zahl eventually removed the offending code, he also allegedly tried to remove the evidence that it ever existed by force-pushing an update to delete the discussion entirely ｣
AI CINESE. POCHI SOLDI, TANTA RESA
#Deepseek AI, il corso cinese dell' #intelligenzaartificiale, è addestrato con una frazione dei costi sostenuti dai giganti occidentali.
In uno studio pubblicato su #Nature, il team di ricerca ha rivelato che ha incentivato le capacità di ragionamento degli #LLM con l'apprendimento per rinforzo (RL) puro, ovvero un sistema di punteggio: alto per le risposte corrette e basso per le sbagliate.
👩🏻💻 AI Coding / geohot
｢ That anyone uses LLMs to code is a testament to just how bad tooling and languages are. And that LLMs can replace developers at companies is a testament to how bad that company’s codebase and hiring bar is. AI will eventually replace programming jobs in the same way compilers replaced programming jobs. In the same way spreadsheets replaced accounting jobs ｣
https://geohot.github.io//blog/jekyll/update/2025/09/12/ai-coding.html
👩🏻💻 AI Coding / geohot
｢ That anyone uses LLMs to code is a testament to just how bad tooling and languages are. And that LLMs can replace developers at companies is a testament to how bad that company’s codebase and hiring bar is. AI will eventually replace programming jobs in the same way compilers replaced programming jobs. In the same way spreadsheets replaced accounting jobs ｣
https://geohot.github.io//blog/jekyll/update/2025/09/12/ai-coding.html
AI CINESE. POCHI SOLDI, TANTA RESA
#Deepseek AI, il corso cinese dell' #intelligenzaartificiale, è addestrato con una frazione dei costi sostenuti dai giganti occidentali.
In uno studio pubblicato su #Nature, il team di ricerca ha rivelato che ha incentivato le capacità di ragionamento degli #LLM con l'apprendimento per rinforzo (RL) puro, ovvero un sistema di punteggio: alto per le risposte corrette e basso per le sbagliate.
#OpenAI Has a Fix For Hallucinations, But You Really Won't Like It. Via @sciencealert #AI #ArtificialIntelligence 💻 🤖 🧠 #LLM #ChatGPT
OpenAI Has a Fix For Hallucina...
A space for Bonfire maintainers and contributors to communicate
