Samson Timoner

This paper is so cool: dangeng.github.io/visual_ana… It shows several kinds of illusions that I had never seen before (e.g., color inversion illusion). It's exciting to see more and more cases like this, where AI opens up new kinds of art, rather than only imitating old forms.

A young statistician saved their lives. His insight (and how it can change yours): During World War II, the U.S. wanted to add reinforcement armor to specific areas of its planes. Analysts examined returning bombers and plotted the bullet holes and damage on them (as in the image below). Based on this analysis, they came to the conclusion that adding armor to the tail, body, and wings would improve their odds of survival. But a young statistician named Abraham Wald noted that this would be a tragic mistake. By only plotting data on the planes that returned, they were systematically omitting the data on a critical, informative subset: The planes that were damaged and unable to return. Abraham Wald recognized a key fact: • "Seen" planes had sustained damage that was survivable. • "Unseen" planes had sustained damage that was not survivable. Wald concluded that armor should be added to the *unharmed* regions of the returning planes (the areas without bullet holes on the image below). His profound logic: Where the survivors were unharmed was actually where the planes were most vulnerable. Based on his insight, the military reinforced the engine and other vulnerable parts, significantly improving the safety of the crews during combat and saving thousands of lives. Abraham Wald had identified a cognitive bias called "Survivorship Bias": The error resulting from systematically focusing on survivors (successes) and ignoring casualties (failures) that causes us to miss the true base rates of survival (the actual probability of success) and arrive at flawed conclusions. We see examples of Survivorship Bias all around us: 1. We read books on the common traits of successful people, but fail to consider all of the unsuccessful people who possessed those same traits. 2. We applaud the belief when we hear that an entrepreneur took out a second mortgage and succeeded, but fail to consider all of the entrepreneurs who did the same and went bankrupt. 3. We study the cultural strategies of the most successful companies, but fail to consider all of the companies that followed those same strategies and fell apart. When we fail to consider the range of outcomes and the hidden evidence, we develop a skewed (and often incorrect) view of reality. It cannot be avoided altogether, because the vast majority of books and history are written by and about the survivors and victors, but wherever possible, consider the unseen evidence. Remember: What is unseen often has just as much value as what is seen. *** If you enjoyed this or learned something, follow me @SahilBloom for more in future!

Introducing FastViT, fast, super-small, general purposes vision transformers running at ~ 1ms on mobile. Code and pre-trained models are live. github.com/apple/ml-fastvit #ICCV2023

This is a wild chart. The demand for AI talent is surging — and there isn't enough talent available to fill the positions. There's no better time to be leveraging the technology and learning new AI-related skills to get ahead.

How do we know how smart AI systems are? | Science science.org/doi/10.1126/scie…

Should you train your vision system on real images or synthetic? In the era of stable diffusion, the answer seems to be: synthetic! One stable diffusion sample can be worth more than one real image. paper link: arxiv.org/abs/2306.00984

Last night we had an emergency Zoom call with most of Congress to stop a bank run. Here’s the situation: