Anthropic's new AI model turns to blackmail when engineers try to take it offline | TechCrunch https://buff.ly/OQo2mBn
"Anthropic notes that Claude Opus 4 tries to blackmail engineers 84% of the time "
Anthropic's new AI model turns to blackmail when engineers try to take it offline | TechCrunch https://buff.ly/OQo2mBn
"Anthropic notes that Claude Opus 4 tries to blackmail engineers 84% of the time "
https://www.europesays.com/uk/161099/ Meta Tests 3D Instagram Photos, Virtual Desktop for Headsets #ArtificialIntelligence #AugmentedReality #GenAI #Innovation #Instagram #meta #MetaQuest #News #PYMNTSNews #SocialMedia #Technology #UK #UnitedKingdom #VirtualReality #VR
"Let’s say you’re ... organizing a fundraiser [and] ask an AI model 15 questions about the best way to fundraise.
Then you make 10 attempts at an [genAI] image for your flyer... and three attempts at a five-second video to post on Instagram.
You’d use about 2.9 kilowatt-hours of electricity—enough to ride over 100 miles on an e-bike (or around 10 miles in the average electric vehicle) or run the microwave for over three and a half hours."
https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech/
#AI #ClimateImpact #LLM #genai
So I had this #dream that I was with Walter White and some other dude and we were headed into a hotel on a mission. We were supposed to confront this dude on his own turf. He wasn't expecting it.
Anyway, we get in the lobby and it goes sideways almost immediately. This asian dude rounds a corner and recognizes Walter, and the other dude on our crew loses his cool and just starts unloading lead into the guy. Alarms go off. People start panicking. Shit goes nuts. And I'm dead.
So the level reloads and we try again.
Same result.
This time, I get smart -- I realize the cops think I'm a suspect because I'm dressed nice like Walter and the other guy.
So I very quietly loosen my tie, taking it off, unbuttoning my brown sport coat and tearing the top of my blue shirt open, and mess my hair up, and then I crawl into the corner of an elevator with another person -- an actual innocent -- and just... blend in...
Then I woke up.
It's wild being able to recreate reasonably accurate visuals of dreams with #genAI. And I know I will probably never wear that tie with that combination IRL, no matter how much I love the tie.
"Holmesian genius" realizes he's out of deductions on #GenAI
—a shocking revelation for this self-proclaimed intellectual titan. The conversation's not over, but he's tapping out, folks.
https://blog.glyph.im/2025/06/i-think-im-done-thinking-about-genai-for-now.html #HolmesianGenius #IntellectualTitan #TechRevelation #AIConversations #HackerNews #ngated
this is amazing : #NLP to #CAD Design https://www.adamcad.com/
A great way to talk to your 3D Printer ;)
Boring work today: I'm sorting all my #AI bookmarks into subfolders because it's getting confusing. It's devastating to see years of titles rushing by:
- We were warned.
- We could have known everything.
- The footprint is the most brutal.
- Very few sources are read by the masses.
- Memories of the dot.com bubble of 2000 come to my mind.
- We become numb.
- What will implode first: our ethics or the AI-Bubble?
- It's depressing to do this work.
Agent Loop is a new capability in Azure Logic Apps to build #AIagents into your enterprise workflows – enabling your apps & automation to not just follow predefined steps, but to reason, adapt, and act autonomously towards goals https://techcommunity.microsoft.com/blog/integrationsonazureblog/%F0%9F%93%A2announcing-agent-loop-build-ai-agents-in-azure-logic-apps-%F0%9F%A4%96/4415052
From #OpenAI: A practical guide to building agents
https://cdn.openai.com/business-guides-and-resources/a-practical-guide-to-building-agents.pdf
"Automating jobs isn’t the panacea for all companies, and 55% admit they have made a mistake, a new survey finds."
https://www.techrepublic.com/article/news-leaders-regret-ai-driven-layoffs
When it comes to AI, there are definitely two camps with completely incommensurable paradigms.
I wonder if these two outlooks are going to split the future in half, concerning how we actually treat artificial agents.
Think of how, in the movie AI, there were these wildly uneven moods of slick plastic utopias and grimy cynical dystopias.
Not a pro coder, just vibe it sometimes - but this piece from @ghuntley.com really hit home. #AI might speed up the “what,” but the “how” is where the growth happens. The journey matters!
#Coding #GenAI
RE: https://bsky.app/profile/did:plc:r63key5ilrlzklg2sqcocgug/post/3lqrzhwmifl2m
Jest taka historia, jak się okazało, że w paczce do Python jest błąd i było możliwe, że chemicy organiczni podali błędne wyniki w ok. 150 artykułach naukowych.
Udało się wykryć ten błąd i go naprawić.
Z błędami w generatywnym AI tak się nie da. I to poważny problem.
Napisałem o tym blogonotkę, zapraszam serdecznie!
Andy Pryor is 'Rethinking User Workflows with Generative AI' July 25th at Nebraska.Code().
"The process of coding with an “agentic” LLM appears to be the process of carefully distilling all the worst parts of code review, and removing and discarding all of its benefits."
Very insightful post on #GenAI by Glyph
(Original title: I Think I’m Done Thinking About genAI For Now)
https://blog.glyph.im/2025/06/i-think-im-done-thinking-about-genai-for-now.html
Very very interesting, thoughts on #AI which are actually worthy of reading.
https://blog.glyph.im/2025/06/i-think-im-done-thinking-about-genai-for-now.html
The case for #GenAI is the same as the case for magic: When done right, rain dances always work. If they don’t, you obviously didn’t to them right. 3/3
I can’t help it, I somehow feel gaslighted by the whole #GenAI debate. People who are critical of GenAI are often told “but when done right, you’re just so much more productive, so obviously you just didn’t to it right.” So I’m trying. Not because I feel I need to, but because I just want to get a realistic impression of the capabilities, and be able to discuss these with my students.
Then fancy-pants Claude 4 with extended thinking spits out this line: 1/
Funny how you can instantly see when an image is ChatGPT-generated. It always has a slight sepia haze, next to other characteristics.