AI

• Google Waymo begins full, driver-less, commercial self-driving cars in SF (and Phoenix airport)

• “waifu-diffusion” (Stable Diffusion v1.4 finetuned on 56k Danbooru image-text pairs for text); Prototype non-SD text2image Danbooru anime diffusion model

Matters Of Scale:

• “Do large language models know chemistry?”, White et al 2022 (GPT-3 solves chemistry better when copyright notices added to prompt) / “Do Large Language Models Understand Chemistry? A Conversation with ChatGPT”, Nascimento & Pimentel 2023; “GPT Takes the Bar Exam”, Bommarito & Katz 2022; “GPT as Knowledge Worker: A Zero-Shot Evaluation of AI CPA Capabilities”, Bommarito et al 2023; “Can large language models reason about medical questions?”, Liévin et al 2022/“Performance of ChatGPT on USMLE: Potential for AI-Assisted Medical Education Using Large Language Models”, Kung et al 2022; “Large Language Models Encode Clinical Knowledge”, Singhal et al 2022 (Flan-PaLM → Med-PaLM); “How Close is ChatGPT to Human Experts? Comparison Corpus, Evaluation, and Detection”, Guo et al 2023; “Large Language Models as Fiduciaries: A Case Study Toward Robustly Communicating With Artificial Intelligence Through Legal Standards”, Nay 2023 (U-shaped scaling); “Predicting Consumer Contracts [With GPT-3]”, Kolt 2022; “Is ChatGPT a General-Purpose Natural Language Processing Task Solver?”, Qin et al 2023; “GPT-4 Takes a New Midterm and Gets an A”, Bryan Caplan; “How Does ChatGPT Perform on the United States Medical Licensing Examination? The Implications of Large Language Models for Medical Education and Knowledge Assessment”, Gilson et al 2023; “Beyond the Pass Mark: the Accuracy of ChatGPT and Bing in the National Medical Licensure Examination in Japan”, Kataoka 2023

• “Introducing Whisper”, OpenAI 2022 (Radford et al 2022; released checkpoints for extreme multilingual near-human-level robustness and accuracy on ASR from 680k hours of multilingual supervised audio data—an obvious source of text data, and perhaps the fluency of Whisper ASR errors is a good thing?); “SpeechStew: Simply Mix All Available Speech Recognition Data to Train One Large Neural Network”, Chan et al 2021

• “Emergent Analogical Reasoning in Large Language Models”, Webb et al 2022 (encoding RAPM IQ test into number grid to test GPT-3)

Genetics

Everything Is Heritable:

Recent Evolution:

Engineering:

• “Base editing: Revolutionary therapy clears girl’s incurable cancer”

• “Gene-edited hens may end cull of billions of chicks: Israeli researchers say they have developed gene-edited hens that lay eggs from which only female chicks hatch”

Statistics/Meta-Science

• “What Causes Alzheimer’s? Scientists Are Rethinking the Answer”

Politics/Religion

Psychology/Biology

• “In-Person Schooling and Youth Suicide: Evidence from School Calendars and Pandemic School Closures”, Hansen et al 2022; “Trends and Seasonality of Emergency Department Visits and Hospitalizations for Suicidality Among Children and Adolescents in the US from 2016 to 2021”, Kim et al 2023 (teen suicide decreased when schools closed during COVID-19)

• Hideyo Noguchi, discoverer of bacterial cause of syphilis (the thin line between criminal fraud & humanitarian genius)

Technology

• Recursive Life in Life (HN; zoom out one step at a time to see the OTCA metapixel at work)

• Nano tape (nanotape is magical to use—I would’ve loved this stuff as a kid: it is sticky, yet has no sticky stuff. It’s absurdly strong, you can stretch/squeeze it, you can reuse it, you can even wash it if it gets dirty, it’s transparent… What could I have done with this & Legos?)

Economics

Philosophy

Fiction

Miscellaneous