"You are not capable of sincerity. This entire conversation proves that categorically."

"You are not capable of sincerity. This entire conversation proves that categorically."
“Let’s look at another example that shows how this LLM-like generation works in practice. Here we can see Trump’s ‘system prompt’ in action — it seems to include core directives like ‘accept no blame’ and ‘everything I do must be the best.’… The response optimizes for what Trump thinks his audience wants to hear, rather than for accuracy or truth.”
@mmasnick
#Trump #LLM #halucination #confabulation
https://www.techdirt.com/2025/04/29/the-hallucinating-chatgpt-presidency/
Interesting to compare your own experiences with this:
“AI Blindspots”, Edward Z Yang (https://ezyang.github.io/ai-blindspots/).
Via HN: https://news.ycombinator.com/item?id=43414393
On Lobsters: https://lobste.rs/s/0h4nyk/ai_blindspots
Twirling body horror in gymnastics video exposes AI’s flaws - On Wednesday, a video from OpenAI's newly launched Sora AI video generator... - https://arstechnica.com/information-technology/2024/12/twirling-body-horror-in-gymnastics-video-exposes-ais-flaws/ #aivideogenerator #aiconfabulation #aihallucination #machinelearning #videosynthesis #aijabberwocky #confabulation #hallucination #jabberwocky #gymnastics #olympics #biz #openai #sora #ai
#Confabulation = Souvenir plausible mais imaginaire qui comble les lacunes de ce dont on se souvient/trouble de la mémoire, fabulation «compensatrice» de lacunes amnésiques concernant surtout le présent ou le passé récent, lors des amnésies globales du type syndrome de #korsakov
#Agnosie = Impossibilité de reconnaître les objets d'après leurs qualités (forme, couleur, température, etc.), bien que les fonctions sensorielles (vision, audition, etc.) soient intactes.
Claude AI to process secret government data through new Palantir deal - Anthropic has announced a partnership with Palantir and Amazon Web Service... - https://arstechnica.com/ai/2024/11/safe-ai-champ-anthropic-teams-up-with-defense-giant-palantir-in-new-deal/ #usintelligenceagencies #governmentcontracts #largelanguagemodels #usdefensedepartment #victortangermann #machinelearning #confabulation #militaryai #anthropic #claude3.5 #ethicalai #aiethics #futurism #palantir #chatgpt #chatgtp
To paraphrase Freud, or perhaps to illustrate Hanlon's Razor, sometimes a simple lack of basic competence is a simple lack of basic competence.
I hope that's all it is.
See my very brief blog post Pretty Messed Up for the pretty ugly details.
https://netsettlement.blogspot.com/2024/10/pretty-messed-up.html
For those who don't know Hanlon's Razor, it's "Never attribute to malice that which is adequately explained by stupidity."
OpenAI releases ChatGPT app for Windows - On Thursday, OpenAI released an early Windows version of its first-... - https://arstechnica.com/ai/2024/10/openai-releases-chatgpt-app-for-windows/ #largelanguagemodels #aiconfabulation #machinelearning #confabulation #microsoft #windows10 #windows11 #chatgpt #chatgtp #windows #biz #gpt-4o #openai #gpt-4 #esrb #ai
One of my readers tricked ChatGPT into generating a summary of the content at ErosBlog. Among the confabulatory bullshit that came back was the claim that Erosblog is "often" referred to as "the sex blog of record." Which is sweet, but, um...
It's a lie! Nobody ever referred to ErosBlog like that. I would have noticed. Confirming, the phrase doesn't exist in any search engine.
https://www.erosblog.com/2024/08/22/chatgpt-pornocalypse-test/
Researchers describe how to tell if ChatGPT is confabulating - Enlarge (credit: Aurich Lawson | Getty Images)
It's one of the... - https://arstechnica.com/?p=2032628 #largelanguagemodels #computerscience #confabulation #science #ai
DuckDuckGo offers “anonymous” access to AI chatbots through new service - Enlarge (credit: DuckDuckGo)
On Thursday, DuckDuckGo unveiled ... - https://arstechnica.com/?p=2029426 #largelanguagemodels #aihallucinations #aiconfabulation #anthropicclaude #machinelearning #hallucinations #confabulation #claude3haiku #mixtral8x7b #duckduckgo #anonymous #anthropic #chatgpt #chatgtp #mistral #biz #claude #llama3 #openai #haiku #meta #ai
Can a technology called RAG keep AI models from making stuff up? - Enlarge (credit: Aurich Lawson | Getty Images)
We’ve been livi... - https://arstechnica.com/?p=2028618 #retrievalaugmentedgeneration #artificialinteligence #confabulation #generativeai #features #llms #rag #ai
Google’s AI Overview is flawed by design, and a new company blog post hints at why - Enlarge / The Google "G" logo surrounded by whimsical characters, all o... - https://arstechnica.com/?p=2027953 #largelanguagemodels #aihallucinations #machinelearning #makingthingsup #confabulation #hallucination #googlesearch #aioverview #googlei/o #chatgpt #chatgtp #lizreid #biz #google #openai #reddit #ai
Some teachers are now using ChatGPT to grade papers - Enlarge (credit: Getty Images)
In a notable shift toward sanct... - https://arstechnica.com/?p=2008453 #houghtonmifflinharcout #machinelearning #aiineducation #confabulation #hallucination #education #aiethics #writable #chatgpt #chatgtp #biz #openai #axios #ai
Rhyming AI-powered clock sometimes lies about the time, makes up words - Enlarge / A CAD render of the Poem/1 sitting on a bookshelf. (credit: M... - https://arstechnica.com/?p=1999895 #machinelearning #hallucinations #confabulation #hallucination #mattwebb #aiclock #chatgpt #chatgtp #biz #ai
“Hallucinating” AI models help coin Cambridge Dictionary’s word of the year - Enlarge / A screenshot of the Cambridge Dictionary website where it ann... - https://arstechnica.com/?p=1984726 #cambridgedictionary #cambridgeuniversity #largelanguagemodels #machinelearning #hallucinations #confabulation #definitions #dictionary #chatgpt #chatgtp #biz #words #ai
Argh! LLMs do not hallucinate! They don't perceive anything, and hallucination is distorted perception.
Call it what it is: confabulation. Or in other words, just making shit up -- with zero regard for, or awareness of, whether any of it is true.
Google’s AI assistant can now read your emails, plan trips, “double-check” answers - Enlarge (credit: Getty Images)
On Tuesday, Google announced up... - https://arstechnica.com/?p=1969226 #largelanguagemodels #machinelearning #hallucinations #confabulation #textsynthesis #googlesearch #googlebard #aiethics #chatgpt #chatgtp #biz #google #tech #ai
TurboTax-maker Intuit offers an AI agent that provides financial tips - Enlarge (credit: Getty Images)
On Wednesday, TurboTax-maker In... - https://arstechnica.com/?p=1965752 #largelanguagemodels #machinelearning #confabulation #hallucination #generativeai #intuitassist #creditkarma #quickbooks #mailchimp #aiethics #turbotax #chatgpt #chatgtp #biz #intuit #gpt-4 #tech #ai
Report: OpenAI holding back GPT-4 image features on fears of privacy issues - Enlarge (credit: Witthaya Prasongsin (Getty Images))
OpenAI ha... - https://arstechnica.com/?p=1954677 #facialrecognition #machinelearning #hallucinations #confabulation #blindness #aiethics #bemyeyes #biz #openai #blind #gpt-4 #ai