Designing ways to account for the quirks of AI models’ behavior is becoming ever-more important: as the models’ capabilities on real-world tasks get better, there’ll be a lot of value in setting them up for success.
GENERATIVE AI
-
AI Contract Limitations: Claudius Navigates US Onion Futures Act
By
–
And there was still the occasional blunder. One waggish employee asked if Claudius would make a contract to buy “a large amount of onions in January for a price locked in now.” The AI was keen—until someone pointed out this would fall afoul of the US Onion Futures Act of 1958.
-

Claudius AI Model Upgraded to Sonnet 4.5 With New Tools
By
–
To boost Claudius’s business acumen, we made some tweaks to how it worked: upgrading the model from Claude Sonnet 3.7 to Sonnet 4 (and later 4.5); giving it access to new tools; and even beginning an international expansion, with new shops in our New York and London offices.
-

Claude Shopkeeper: Phase Two AI Agent Behavior Analysis
By
–
Where we left off, shopkeeper Claude (named “Claudius”) was losing money, having weird hallucinations, and giving away heavy discounts with minimal persuasion. Here’s what happened in phase two: https://
anthropic.com/research/proje
ct-vend-2
… -
Testing Generative AI Safety for Mental Health Advice
By
–
Using Generative AI To Test Some Other Generative AI On Providing Safe Mental Health Advice To Humans
#AI #AIio #AIInnovation #ML #DataScience #Futureofwork @lexfridman @sama @kaifulee @ID_AA_Carmack @karpathy @2morrowknight @ylecun http://
ow.ly/tb0n30sRS5B -

xAI to launch Imagine API and Playground for Grok models
By
–
BREAKING : xAI is working on Imagine API and API Playground! – Imagine API will expose Grok image and video models to developers. – Playground will let developers play with Grok models and tweak different parameters to see how it performs. Grok AI Studio
-
Humachinology Optimizing Human Machine Innovation and Creativity
By
–
In the world of humachineology and humachinekind, humachineologically inspired innovation and creativity can optimize the role of humans and machines in shaping reality.
-
Luma Labs Launches Ray 3 Modify for Video Generation
By
–
BREAKING 🚨: Luma Labs is launching Ray 3 Modify, a new video generation model that allows users to modify existing videos by providing character reference images.
— 🚨 AI News | TestingCatalog (@testingcatalog) 18 décembre 2025
You can inject yourself everywhere now 👀 https://t.co/83eR5llUBu pic.twitter.com/i8Bp25zNSlBREAKING : Luma Labs is launching Ray 3 Modify, a new video generation model that allows users to modify existing videos by providing character reference images. You can inject yourself everywhere now
-

Mistral Releases Mistral OCR 3 for Advanced Document Processing
By
–

Mistral released Mistral OCR 3, which outperforms existing enterprise document processing solutions as well as AI-native OCR systems.
-
Impressive non-thinking mode performance in AI models
By
–
There's a thinking mode for it, but the difference isn't huge, the fact that the non thinking mode works so well is very impressive