Ilana Arbisser retweeted
I know I'm a bit biased, but yesterday's chat with Jay and Michael about the magic of @scenicroutessf was everything I love about this town, its small businesses, and the communities they create.
Give it a listen! (starting about an hour in)
podcasts.apple.com/us/podcas…
Ilana Arbisser retweeted
Here's our breakdown:
Cluster 1: Ethical approaches to tackling current harms
AI isn't just a future problem, it's already causing issues of misinformation, algorithmic bias, and centralization of power
3/
Ilana Arbisser retweeted
I never really told this story, but my plan was to drop out of my @Stanford PhD to start a company. Like Larry & Sergey did for Google.
But I got so enamored by the magic of generative AI, even back when it was outputting 💩, that the joke was I *failed to drop out*.
I've now started a company with an incredible team! It's to spread this magic and enable other companies to build their own generative AI with their engineering teams.
We're still in stealth, but if you look closely you might find something ;)
Thanks for the podcast, TensorFlow and Google!
sites.libsyn.com/464853/webs…
🫱🏿🫲🏽Meet Sharon Zhou, our special guest this week discussing the power and interconnectedness of language as we dive into generative AI models, stable diffusion, large language models, and more ⬇️
ALT Headline text: "The Language of Machine Learning" with image of Sharon Zhou smiling with black hair and a white shirt.
Ilana Arbisser retweeted
🎉 GPT-4 is out!!
- 📈 it is incredible
- 👀 it is multimodal (can see)
- 😮 it is on trend w.r.t. scaling laws
- 🔥 it is deployed on ChatGPT Plus: chat.openai.com
- 📺 watch the developer demo livestream at 1pm: youtube.com/live/outcGtbnMuQ…
Announcing GPT-4, a large multimodal model, with our best-ever results on capabilities and alignment: openai.com/product/gpt-4
Ilana Arbisser retweeted
I care so much about this that I wrote a book about it, which is really a love letter to my old friends in psychology, an invitation across a bridge. I don't think I got it right everywhere, but I was thinking about them constantly while writing global.oup.com/academic/prod…
Ilana Arbisser retweeted
My last thoughts: Peter Eckersley lived a life of principle, brought joy to his friends, engaged deeply in trying to make the world better, and pursued eclectic and multidisciplinary intellectual pursuits with creativity. May we all strive to live up to that excellent example.
Ilana Arbisser retweeted
Recombination, specifically homologous recombination, is a major driver of evolution in bacteria. But compared to de novo mutation we know very little about it, in part because the methods required to study it are poorly developed compared to standard phylogenetic inference [1/N]
Ilana Arbisser retweeted
Few quick thoughts re: @POTUS's Covid:
1) Not clear where he got it – he returned from Middle East ~3 days before test +, so probably last day of trip or flight home. Flying is low, not zero, risk, particularly if people weren't wearing masks. (1/13)
Ilana Arbisser retweeted
I was intimidated by theoretical ecology until working with @NDemalach on this illustration, back in the 2nd year of my PhD. With through gentle discussion, we visualized the model. Art 🎨 supports understanding and exploration of science 🧪!
4\4 @eco_genome illustrated the model focusing on herb species undergoing light preemptions by trees. The persistence probability of the herbs increases with (i) light-use efficiency, (ii) lower tree density, and (ii) sparser foliage
Ilana Arbisser retweeted
I've created an infographic to explain Simpson's Paradox.
The @nytimes' @DLeonhardt falsely claimed that in 2022, COVID mortality was higher for white vs non-white Americans.
The reverse is true.
#epitwitter #misinformation
ALT SIMPSON'S PARADOX is a strange phenomenon to grasp, but it is a crucial one for anyone trying to compare subgroups. This Paradox is a constant reminder that we must be CAUTIOUS when making claims comparing groups in different categories, especially under non-experimental conditions. Erroneously, the New York Times' David Leonhardt falsely claimed that in 2022, the racial divide had flipped, and in 2022 more white Americans (57.6 per 100k) than non-white Americans (38.6 per 100k) died of COVID. However, Simpson's paradox fully applies here (this should be a new textbook example). At no age is that statement true. The rate for each individual age is ALWAYS higher for the non-white group than the white group. This is why journalists and amateur statisticians need to be careful and use experts.
Ilana Arbisser retweeted
Our new study shows that data availability statements are not very useful; 1670 (93%) authors who indicated that data are available on request either did not respond or declined to share their data with us. Journal of Clinical Epidemiology: doi.org/10.1016/j.jclinepi.2…
Ilana Arbisser retweeted
The exciting reveal of Ultima Genomics last week was accompanied by the publication of four preprints. Intrigued by the potential of the technology, @sinabooeshaghi & I decided to take a look at the data. A 🧵 about our findings & a preprint we posted: biorxiv.org/content/10.1101/… 1/
Ilana Arbisser retweeted
Just out @NatureMedicine
The natural history of post-vaccine breakthrough infections (BTI) in 33,000 people, >13 million controls
nature.com/articles/s41591-0…
@zalaly @BCBowe @Biostayan @WUSTLmed
—Vaccination protection from #LongCovid ~15%, much less than many other studies (~50%)
