Devon Bath

Sat next to a software engineer on the train to Stanford. We talked the whole ride about one thing: how hard it's gotten to stay current. He builds ML models for a living and still can't read fast enough to keep up. Staying current has quietly become the job.

Google Scholar alerts, PubMed, RSS. They all assume you have time to go fishing. They hand you 200 results and call it help. The labor of filtering still lands on you. That part should belong to the machine, not the scientist.

Most researchers do not fall behind because they stopped caring. The math stopped working. A subfield can publish more in a month than anyone reads in a year. The usual advice is to make peace with it. We think that advice is wrong. Staying current should not be a second job.

The hardest part of Atomic Research was never fetching papers. It is deciding what is relevant to you specifically. Keyword matching is useless. Two labs can share every keyword and care about completely different things. Our system knows what matters atomicresarch.co

Last year the plan was to study sharks in Australia. Headed for a professorship. Somewhere in there I realized the part of research I loved, the frontier, was the part the job had the least time for. So I moved to SF and started building a fix. I still think about the sharks.

Anthropic: “If we’re going down, everybody’s coming with us.”

The amount of knowledge is doubling every 12 hours, but is this to the detriment of science? open.substack.com/pub/atomicrese…

Jane Street pays $750k/year for engineers who build AI trading systems from scratch. This 15-minute talk by Jane Street senior engineer shows the exact system behind a $750k/year skillset. here's what he covers: • why generic AI fails at a top trading firm • custom models trained on their own trading code • real-time grading of every AI-written change • one AI engine across the whole trading desk most retail traders are still using generic AI tools - while the top hedge funds train their own models Watch this video then read article below

Low predator and conspecific density lead to larger juveniles - an example ftom Cod populations. Canadian Journal of Fisheries and Aquatic Sciences. Just-IN  doi.org/10.1139/cjfas-…

🚨 𝗡𝗘𝗪 𝗣𝗔𝗣𝗘𝗥 🚨 Out earlier this year in Reviews in Fish Biology and Fisheries - a new review on euryhaline & estuarine #elasmobranchs 👇 "A review of the life history and ecology of euryhaline and estuarine sharks and rays" Full paper here ➡️ doi.org/10.1007/s11160…

Super excited for this new paper out with my colleague @BathDevon! We took a good look at what glucocorticoids might mean for conservation of animal populations.

The research paper Dr. Newediuk and I have been working on has finally been published to the journal of ‘Conservation Physiology’. Link: doi.org/10.1093/conphy…

It is finally setting in that I have finished my PhD 😁 A long (and fun) journey with @munocean and @munosgsa!

Introduction to Data Science --Data Analysis and Prediction Algorithms with R rafalab.github.io/dsbook/ #rstats #datascience

Today's #rstats cheatsheet: survminer Download: raw.githubusercontent.com/rstudio/cheats… See more: rstudio.com/resources/chea… Contribute your own: github.com/rstudio/cheats…

Four R packages for Automated Exploratory Data Analysis you might have missed by Nicolo Cosimo Albanese in @TDataScience towardsdatascience.com/four-r-package… #rstats

Today's #rstats cheatsheet: distr6 Download: raw.githubusercontent.com/rstudio/cheats… See more: rstudio.com/resources/chea… Contribute your own: github.com/rstudio/cheats…

Making Publication Quality Inset Maps in R using ggplot2 by @mohsinramay_ link.medium.com/ZypzBBW8Lsb #rstats #ggplot2

#StopFinningEU: EU authorities confirm 7 out of 7 necessary country thresholds for a #FinBanNow in Europe! 🦈🎉

Want to transition from making graphs to telling data stories? Here's a 6-step process. (Greater detail in my books: storytellingwithdata.com/books)