"Nearly half of white students admitted to Harvard between 2009 and 2014 were recruited athletes, legacy students, children of faculty and staff, or on the dean’s interest list—applicants whose parents or relatives have donated to Harvard" wsj.com/articles/to-ge…

Here is some news: we release the IBM MoLFormers UI at molformer.res.ibm.com, where one can play with AI Foundation Models for molecules, including MoLFormer-XL doi.org/10.1038/s42256…, and can screen, find neighbors, sample generations, and examine model attentions.

Trying to process the fascination with Titan's debris. In the meantime, Greece is still scouring the Mediterranean Sea for the bodies of the perished migrants. nytimes.com/2023/06/28/wor…

much appreciated 🤣🤣🤣

An accessible summary of "Atomist or holist? A diagnosis and vision for more productive interdisciplinary AI ethics dialogue" about the difference between data scientists as puzzle solvers vs. as social stewards. Kudos to Amit Dhurandhar and coauthors. montrealethics.ai/atomist-or-hol…

i have found the most amazing place

it’s that time of year when u can stand in the middle of broadway and take photos as long as u like

… not sure why, but this station always reminded me of modern art

soho, as it used to be

Me, waiting for Hinton to cite anyone who's not a white man in his talk 💀

why is that some hydrangeas like being in a vase and the others just plop?

... this is how existential risk looks like.

Label Sleuth, the no-code open source tool for labeling AND automatically building text classifiers now supports > 150 languages!!! 🌎🌍🌏 Come to see the demo: IBM Research booth at #ACL2023 (or label-sleuth.org/docs/dev/langu…)

I was recently interviewed by @jaimieseaton for her Scientific American article about AI and antimicrobial resistance. If we were to pick only one application for generative AI today, discovering new drugs would be it. Meanwhile, people are creating images :) scientificamerican.com/article/ai-cou…

I liked an AI generated picture of jellyfish on instagram and all I now get in my feed are AI generated images of jellyfish… People are talking about AGI, while recommender systems are still unsolved problem.

Spot on comment by Saška Mojsilović in this Scientific American article... why build generative AI to do what we humans can easily do. Let's subject them to solving the biggest and toughest challenges of humanity

rip milan kundera theguardian.com/books/2023/jul…

well said... it's the humans -- the irresponsible use, the lack of accountability, and poor safety practices -- who pose far grater threat to humanity that the dooms-day-terminator-AI

New URLs for the IBM Research websites for the Trust 360 toolkits: * AI Fairness 360 aif360.res.ibm.com * AI Explainability 360 aix360.res.ibm.com * Uncertainty Quantification 360 uq360.res.ibm.com * AI FactSheets 360 aifs360.res.ibm.com (continued)

today