Anthropic and the Pentagon are wrangling about the use of AI. This could be very lucrative for Anthropic, especially at a time when all the AI providers are looking for cash. I know that Anthropic wants to be sanctimonious, but do they really think that taking that stance is going to prevent misbehavior from other parties later on? (Very similar to all the climate change restrictions imposed on society.)
Here's a guy who was able to get ChatGPT and Google Gemini to promulgate his homemade falsehoods, as an experiment. Anthropic wasn't fooled, which may be why they are the Pentagon's top choice. This will greatly lessen the usability and trustworthiness of the frontier models.
Perplexity needs money, too, and they were going to incorporate ads, but decided to drop that. Wise move.
And it looks like that jaw-dropping Seedance video of Brad Pitt and Tom Cruise fighting was just green screen foolery. Their AI isn't that good. It was just AI rotoscoping.
Replacing AI with humans is beginning to backfire.
The first signs of burnout are coming from the people who embrace AI the most. Jevon's Paradox effect strikes again.
A doctor wants to train AI to do her job. She must have a really simple job because AI is not there yet. This is going to be a disappointment. But then, this is a CNN article, so don't expect real journalism.
Nice review on LLM reasoning failures. The advanced models are failing logic tests. Yeah, we aren't there yet.