Twitter Beğeni Hilesi

In Denmark, 860 individuals were randomly allocated to either "intervention" or "control": • No intervention was implemented • Individuals were unaware of their allocation • Mortality was higher in the intervention group with p=0.003 Keep this in mind when evaluating a trial.

Here's a piece I just published on scientific innovation and reproducibility: worksinprogress.co/issue/escaping… Via the new online magazine Works in Progress, which is well worth reading.

medRxiv Joseph Ross Richard Sever John Inglis Claire Rawlinson Theodora Bloom Then to the vast majority of journals who have embraced the idea that preprints complement peer-reviewed publications; and that the perform different functions. NEJM from the outset of pandemic actually encouraged use of preprints. A rapid change in the ecosystem. Wind at backs.

My attempt at understandable but technically correct definitions for key terms in Artificial Intelligence in one page for Stanford HAI. With thanks for helpful feedback from people on Twitter, I’ve revised and hopefully improved a few of the definitions: hai.stanford.edu/sites/default/…

Every researcher can find some personal shame in this 1956 Shannon essay "The Bandwagon" and in Peter Elias' 1958 "Two Famous Papers". Both timeless and domain general (unfortunately). Both here in their single page glory:

“We show that published papers in top journals that fail to replicate are cited more than those that replicate. This difference doesn't change after publication of failure to replicate. 12% of postreplication citations acknowledge the replication failure.” advances.sciencemag.org/content/7/21/e…

Just a reminder: being “famous” in academia means tops 40 people will recognize you at an international meeting before your talk. It also means that your barber will also say, “oh cool, what do you teach?”

Machine learning is best for applications where errors have low cost - the opposite of self-driving cars.

Had a great time speaking at the MGB-MIT AI Cures conference a few weeks back ... and my opinionated slide has made its way into Stat News today statnews.com/2023/05/22/ai-…

Academics love to stress *real-world* applications, distribution shift, feedback loops, fairness, & the complexity of human behavior, but upon joining industry coop up in isolated research units, running away from *the product* & leaving biz ppl to do all the critical thinking.

A lot folks are negative on the application layer of ai. I keep hearing that the "LLM wrapper applications" will get disrupted by the LLM's. I think you could argue that back in 2006, when HubSpot was founded and SaaS was exploding that all of us were "database wrapper