AI

Matters Of Scale:

• “Scaling Laws vs Model Architectures: How does Inductive Bias Influence Scaling?”, Tay et al 2022 (architectures may matter much less than scaling)

• “Language Models Can Teach Themselves to Program Better”, Haluptzok et al 2022 (Codex generating new programming puzzles & solutions, which can be auto-checked, then finetuned on for bootstrapping)

• “ML-Enhanced Code Completion Improves Google Developer Productivity”, Tabachnyk & Nikolov2022

• “PIXEL: Language Modeling with Pixels”, Rust et al 2022 (avoiding BPE problems by learning on raw images of text instead; cf. Mansimov et al 2020/Hinami et al 2020); “TVLT: Textless Vision-Language Transformer”, Tang et al 2022

• “Language models show human-like content effects on reasoning”, Dasgupta et al 2022

Genetics

Everything Is Heritable:

• “Polygenic architecture of rare coding variation across 400,000 exomes”, Weiner et al 2022

• “Scalable, high quality, whole genome sequencing from archived, newborn, dried blood spots”, Ding et al 2022 (vast archives); “Infanticide vs. inherited cardiac arrhythmias”, Brohus et al 2021 (Kathleen Folbigg was pardoned & freed June 2023—sequence everyone)

Recent Evolution:

Engineering:

• “More CRISPR In Human Subjects”: on Verve editing PCSK9 to reduce cholesterol

• “Healthy cloned offspring derived from freeze-dried somatic cells”, Wakayama et al 2022

Statistics/Meta-Science

Politics/Religion

• “Guinea worm disease is close to being eradicated—how was this progress achieved?”, OWID

• “Assortative Matching at the Top of the Distribution: Evidence from the World’s Most Exclusive Marriage Market”, Goñi2022 (on the circulation of elites)

Psychology/Biology

• “Evaluating the Replicability of Social Priming Studies”, Giolla et al 2022 (“the meta-analytic average was virtually zero: d = 0.001”); “The problem of miscitation in psychological science: Righting the ship”, Cobb et al 2023

• “On the prediction of human intelligence from neuroimaging: A systematic review of methods and reporting”, Viera et al 2022 (fMRI now predicts general intelligence at r = 0.42)

Technology

• “Trial by Internet: A Randomized Field Experiment on Wikipedia’s Influence on Judges’ Legal Reasoning”, Thompson et al 2022 (+20% citation to precedents with newly-written WP articles)

Economics

Philosophy

Fiction

Miscellaneous