jeremynixon

Mechanics and Insights from Extreme Productivity Experiments

This blog post is flawed. "Life insurers can predict when you'll die with about 98% accuracy." Is not even properly framed and is found nowhere in the cited report.

Predictions of when you will die need a range in order to be attached to a number like accuracy. The attached report is not about this but about population-level mortality trends.

This blog post is flawed.

"Life insurers can predict when you'll die with about 98% accuracy." Is not even properly framed and is found nowhere in the cited report.

Predictions of when you will die need a range in order to be attached to a number like accuracy. The attached report is not about this but about population-level mortality trends.

There is article shows no evidence of fabrication, fraud or misinformation, while making accusations of all of them. All it shows is that ChatGPT was used, which is wildly escalated into "evidence manipulation" (ironically without evidence).

Much more work is needed to show that this means anything.

Search through youtube via embedding search over frames of video, in a click-to-semantic-search interface. (Right click to play video)

For now, the researcher using our tool needs to check.

or perhaps, free scientists to focus on genuine insight, ending publish or perish forever

OP here: We have data source powered research papers in the making, that reports on experiments executed with codegen - really hopeful that the papers can be informative at a minimum.

It's clear that superhuman citation depth & breadth is already imminent. Hopefully we can push hallucinations to near-zero with the next generation of models.

What libraries do you see as being SOTA? Fitz? Tika?

My hope is that computer vision + OCR will solve this once and for all in near future.