Biomedical Informatics PhD • CITRIS Health @UC Berkeley • FAMIA • Focusing on Informatics and AI in medicine • Linfield U. Grad • Missoula MT
https://smcgrath.phd
Scott McGrath
Loading...
Quite a list of authors. See also Diyi Yang, @mariaa.bsky.social , M Sagalnik, @manoelhortaribeiro.bsky.social , &c
Polymarket offered me enough money to pay my rent for ten months for one post in my newsletter and I turned it down because I enjoy being able to sleep at night.
Anthropic announces the launch of Claude Fable 5, its next-gen Mythos-class AI.
Early testers were able to convert a 50M-line Ruby codebase in a single day, something that usually takes 2 months.
Direct link to the paper: www.nature.com/articles/s44...
#MedSky
People are turning to AI for health questions.
Microsoft Copilot logs reveal that 1 in 4 consumer health chats involves a personal symptom, condition, or emotional concern. These queries are twice as likely to come from mobile devices and spike sharply at night.
#MedSky
Some shots of the cars
#porsche #toystory5
I've had access to Claude 5 Fable for a bit. A genuine jump in capability, I could feed it a 15 page design document for a project and it would work for 9+ hours and deliver terrific results.
But working with it is weird & weirder is coming
Lots of examples: open.substack.com/pub/oneusefu...
Interesting collaboration between @porscheusa.bsky.social and Pixar.
#porsche #toystory5
Invisible text injection can artificially inflate quality scores in AI-assisted peer review by nearly a full point. However, a new study shows early-2026 frontier models actively penalize documents containing these adversarial prompts, dropping scores instead.
#AcademicSky
Why I think Anthropic's uneven safety policies with the release of Claude Fable 5 undermine the broader AI community's cohesion and accelerate us to more uncertainty and risk in AI's near-term evolution.
www.interconnects.ai/p/claude-fab...
Today we’re launching Claude Fable 5: a Mythos-class model that we’ve made safe for general use.
www.anthropic.com
An early report on a sample of 500,000 conversations between general public users and Microsoft Copilot from January 2026 identifies the main topics and the hourly and daily trends of how these users…
www.nature.com
A systematic assessment of health-related queries to a chatbot powered by artificial intelligence highlights shortfalls in health-care provision and the responsibilities of AI companies.
This study assesses the integration of frontier large language models including tools such as optical character recognition (OCR) into the scientific peer review process to detect invisible text…
Very pleased to have been on the leadership team for this paper!
LLMs are already being used all over behavioural science. But it is often pretty hard to work out exactly what has been done, and therefore how much confidence to place in the results.
www.nature.com/articles/s41...
INVESTIGATION: 🚨 A Polymarket exec used a personal PayPal account to pay major political influencers like Riley Gaines and Brian Krassenstein, who then hyped how "accurate" Polymarket is at predicting outcomes.
We uncovered $350,000+ flowing to influencers this way
www.politico.com/news/2026/06...
Large language models offer new opportunities for behavioural science, but their rapid evolution poses challenges for research rigour. We introduce a consensus-based reporting checklist to improve tra...
www.nature.com
The company’s chief marketing officer, Matthew Modabber, used a personal PayPal account to send money to content creators, according to records reviewed by POLITICO.
