Blaise Agüera y Arcas
medium.com/@blaisea · Personal blog · on HN since 2017
Blaise Agüera y Arcas, a Google researcher, writes essays on AI, machine learning ethics, and the nature of intelligence.
What HN saysCommenters recommend his essays as rebuttals to AI claims about detecting sexual orientation and physiognomy, and discuss his piece on whether language models understand.
18HN people shared it
5linked it in comments
13stories submitted
166story points
4 years agolast shared
2017people per year2026
Most-shared posts
- 1Do algorithms reveal sexual orientation or just expose our stereotypes?9 people · 16 points · first shared 2018 · 1 comments on HN
- 2Do large language models understand us?4 people · 91 points · first shared 2021 · 68 comments on HN
- 3Physiognomy’s New Clothes (on the Limits of AI)3 people · 57 points · first shared 2017 · 26 comments on HN
- 41011 theses on the concept of nonhuman intelligence1 people · 1 points · first shared 2022 · discussion on HN
- 5Can machines learn how to behave?1 people · 1 points · first shared 2022 · discussion on HN
What HN people wrote when they linked it
skjerns · 2018-02-20
This is probaly the best article that you need to read after reading another one of those 'detecting sexual orientation with dating profile pictures/prison inmates/etc' papers:
https://medium.com/@blaisea/do-algorithms-reveal-sexual-orientation-or…
radarsat1 · 2017-09-25
You're absolutely right to be concerned about this stuff, but be aware that it is generally acknowledged as a problem and that the "ethics of machine learning" is quite an interesting and active research topic.
One of the best articles I've read on the topic, if you're interested: https://medium.com/@blaisea/physiognomys-new-clothes-f2d4b59fdd6a
kgwgk · 2022-03-28
It's not really a complaint about the author's magical expectations that weren't met. If anything, it's a complaint about other people's claims like
https://medium.com/@blaisea/do-large-language-models-understand-us-6f8…
myWindoonn · 2018-07-07
This response to the research is quite good: https://medium.com/@blaisea/do-algorithms-reveal-sexual-orientation-or…
The takeaway: AIs can learn our stereotypes.
halflings · 2017-09-08
Necessary link to an article by Blaise Aguera y Arcas, the lead of an ML team at Google: "Physiognomy’s New Clothes" [0]
He gives several examples, and explains how these type of statements are extremely misleading (and dangerous!), and not something new at all...
https://medium.com/@blaisea/physiognomys-new-clothes-f2d4b59fdd6a