Senator Bernie Sanders interviewed Anthropic's Claude to discuss how AI companies collect and use personal data. While he mistakenly called it an 'AI agent' rather than an LLM, his focus on data privacy and regulation is relevant to AI governance. https://gizmodo.com/hey-bernie-thats-a-not-an-ai-agent-2000736228 #AIagent #AI #GenAI #Governance #Anthropic
Hey Bernie, That's a Not an AI Agent

Good on him for trying, though.

Gizmodo
Tokenmaxxing: Wer verbraucht die meisten KI-Tokens? - Golem.de

In Tech-Unternehmen wetteifern Mitarbeiter darum, so viele KI-Tokens wie mรถglich zu verbrauchen - mit Kosten im sechsstelligen Bereich pro Monat.

Golem.de
The U.K. Financial Conduct Authority has granted Palantir access to its data lake on a trial basis to help detect white collar crime. The three-month partnership allows the American data analytics company to analyse sensitive financial intelligence. Some FCA staff have raised concerns about sharing detection methods with an external vendor. https://gizmodo.com/the-u-k-has-a-new-partner-in-the-fight-against-white-collar-crime-palantir-2000736622 #AIagent #AI #GenAI #Governance #Palantir
The U.K. Has a New Partner in the Fight Against White Collar Crime: Palantir

Not everyone in the relevant agency is thrilled.

Gizmodo

This is stunning. Vasily Grossman in Life and Fate 1960, writing on "this machine whose dimensions and weight will continually increase as it attempts to reproduce the peculiarities of mind and soul of an average, inconspicuous human being."

And pointedly noting in the same breath: "Fascism annihilated tens of millions of people."

(I looked this up because of a post by @maelle)

#AI #GenAI #fascism

Wanted to create a little icon to indicate that no #GenAI was used in creating posts that I wrote, and also that these posts are not available to be used as training for those systems.

I know many such icons exist but I really like the #CreativeCommons license icons, so it's inspired by those!

There is a webpage too with instructions if you want to use it yourself (but I did not want to buy another domain name): https://noai.erikjanvankesteren.nl

This AI Model Creates Video in Real Time

'Just a few months ago scenes were generated in minutes, now it's seconds.'

PetaPixel
Anthropic launches Claude Code Channels, letting users message its AI coding agent over Telegram and Discord. The feature brings the power of Claude Code to mobile messaging, challenging OpenClaw's core value proposition. Built on the Model Context Protocol, it allows developers to receive task notifications on the go. https://venturebeat.com/orchestration/anthropic-just-shipped-an-openclaw-killer-called-claude-code-channels #AgenticAI #Anthropic #AIagent #AI #GenAI

OpenAI declared โ€œCode Redโ€ in Dec. Now It looks more like โ€œCode Dead.โ€ ๐Ÿ’€

The scorecard for 2026 so far:
โŒ Lost $12B in ONE quarter.
โŒ GPT-5 flopped (users preferred the old model).
โŒ Key founders (Ilya, Mira, Greg) left the building.
โŒ Gemini hit 650M users while ChatGPT traffic fell.
โŒ Burning $5B/year on memes.

Is the AGI dream over, or just stalled?

Unpopular opinion: OpenAI is turning into the WeWork of AI. ๐Ÿข๐Ÿ’ฅ

When you lose $12B a quarter to build a model that is worse at basic math than its predecessor, you donโ€™t have a tech company. You have a charity for NVIDIA.

If youโ€™re holding AI infrastructure bets, look at the fundamentals. The gap between the โ€œrevolutionโ€ and the P&L sheet has never been wider.

Who is still bullish on $MSFT / OpenAI at these valuations? Explain why. ๐Ÿงต

https://www.nbloglinks.com/openai-declared-code-red-in-dec-now-it-looks-more-like-code-dead/

#OpenAI #ArtificialIntelligence #Gemini #Tech #AIStocks #GenAI #TechNews

OpenAI declared โ€œCode Redโ€ in Dec. Now It looks more like โ€œCode Dead.โ€ ๐Ÿ’€ โ€“ nbloglinks

OpenAI is bleeding $12B a quarter, GPT-5 flopped, and the founders have fled. Is this the end of the Altman era? We break down the "Code Dead" signals.

nbloglinks

Are we realistically assessing students?

Very interesting piece on assessment and poor genAI policies istitutional policies.

Being in the middle of a full programme level review, my feeling is we're over-assessing and using antiquated methods that really never worked (oh God... another closed book exam... kill me pleeeeease). If only, I don't know, we could talk to students or observe them doing something!

https://wonkhe.com/blogs/trained-to-stop-learning-how-students-are-experiencing-assessment-and-learning-in-an-age-of-ai/

#academia #highered #assessment #genAI

Trained to stop learning: How students are experiencing assessment and learning in an age of AI

New research from Wonkhe finds that assessment design โ€“ not AI policy โ€“ is what determines whether students use AI to learn or to produce. Jim Dickinson and Mack Marshall explain

Wonkhe

๐Ÿง  La maggior parte di #LLM eccelle nel dare risposte corrette, ma fatica quando serve esplorare: idee nuove, alternative lontane, possibilitร  che non sappiamo ancora di cercare. 
๐Ÿ’ก Un paper di Harvard propone una soluzione: https://www.linkedin.com/posts/alessiopomaro_llm-ai-genai-activity-7441740413729665024-qs1S

___ 
โœ‰๏ธ ๐—ฆ๐—ฒ ๐˜ƒ๐˜‚๐—ผ๐—ถ ๐—ฟ๐—ถ๐—บ๐—ฎ๐—ป๐—ฒ๐—ฟ๐—ฒ ๐—ฎ๐—ด๐—ด๐—ถ๐—ผ๐—ฟ๐—ป๐—ฎ๐˜๐—ผ/๐—ฎ ๐˜€๐˜‚ ๐—พ๐˜‚๐—ฒ๐˜€๐˜๐—ฒ ๐˜๐—ฒ๐—บ๐—ฎ๐˜๐—ถ๐—ฐ๐—ต๐—ฒ, ๐—ถ๐˜€๐—ฐ๐—ฟ๐—ถ๐˜ƒ๐—ถ๐˜๐—ถ ๐—ฎ๐—น๐—น๐—ฎ ๐—บ๐—ถ๐—ฎ ๐—ป๐—ฒ๐˜„๐˜€๐—น๐—ฒ๐˜๐˜๐—ฒ๐—ฟ: https://bit.ly/newsletter-alessiopomaro

#AI #GenAI #GenerativeAI #IntelligenzaArtificiale