Karandeep Singh

I wonder if these new watermarks will be more effective than the current watermarks of "--", "exactly the sort of," "rather than," "honest caveat," and "it's not just." interestingengineering.com/ai-robotics/...

Anthropic puts hidden watermarks on Claude text under new EU rules interestingengineering.com

When physicians start relying on AI-drafted content, how does a health system monitor that human-in-the-loop AI and fix what's broken? Our new @ai.nejm.org study from our team at @jchi-ucsd.bsky.social led by Dr. Lily Poursoltan found hundreds of physicians-hours spent editing AI content.

“Methods: In this retrospective study, we built a 2-variable model to predict the binary outcome using logistic superintelligence.”

PSA: Sonnet 5.5 is insanely cheap on Copilot Cowork (~$0 for a workflow that would typically cost $3-4 on Sonnet 5). Either this is a rounding error on Microsoft's side, or Copilot Cowork might just be ready to scale for routine workflows.

In SF, they just sell AGI at vending machines.

I remember when we used to call them neural networks. That was before the Executive Order on Deep Learning.

Anthropic: Fable is included in your Max plan for ONE WEEK and then you’ll have to pay for it. ** 1 week later ** Anthropic: Fable access is extended by ONE MORE WEEK and then you’ll have to pay for it. ** 1 week later ** Anthropic: How come more people aren’t using Fable?

The Ship of Thaiseus: when AI writes the initial draft but you edit it to the point that none of the original words remain.

The squished input bar preventing further text entry 1.5 hours into a Cowork conversation is the Anthropic Blue Screen of Death.

The word “exactly” is exactly the sort of thing an AI would write.

I wish that Siri AI was Cowork for iOS, and that iPhone Duo could function as a sidecar display for Mac.