Pinned Tweet
Magic Dust for Artificial Intelligence Product Managers linkedin.com/pulse/magic-dus… via @LinkedIn
An interesting take. "Where generative AI will change the world is in narrow, mostly boring domains. VCs are ignoring this."
futurism.com/the-byte/ceo-ai…
Replying to @AIatMeta
@MetaAI releases database of 617 million protein structures uncovered using AI inside.com/s/325600 via @Inside
I have say, I got a little warm fuzzy when this one turned around. 🤓stackoverflow.com/questions/…
Less is More: Communication for Product Managers linkedin.com/pulse/less-more… via @LinkedIn
Googly Eyes Could Make Autonomous Cars Safer For Pedestrians
It seems silly but it's quite clever.
gizmodo.com/self-driving-car…
Using RL, the program took 47 multiplications to multiply two 4x4 matrices, beating the previous record of 49 set by Volker Strassen in 1969. Typically it takes 64.
DeepMind algorithms speed up matrix multiplication, setting new record inside.com/s/319929 via @Inside
This is genius. I just backed FlipDie: The First and Only Dice Coin Set on @Kickstarter kickstarter.com/projects/atl…
The same is true for humans!
"... researchers at Stanford trained a ML system to spot gaps in its knowledge and compose (often dumb) questions [posted on @instagram] that strangers would answer."
When AI asks dumb questions, it gets smart fast science.org/content/article/…
Using non-invasive techniques, @MetaAI built AI that can guess what you're hearing.
"Being able to decode brainwaves could help patients who have lost the ability to speak... and... provide novel ways for humans to interact with computers."
singularityhub.com/2022/09/1…
I'm thinking I might need to do a little BASIC programming: mdcramer.github.io/apple-2-b…
[Reading] BlenderBot 3: A deployed conversational agent
I haven't attended this group's sessions in a while (busy), but I'm excited to check out this one.
meetup.com/aifrontiers/event… #Meetup via @Meetup
This conference has an impressive lineup of speakers. I haven't participated in a conference on a long time, so I'll see what I can do to attend some of it.
Join me at Scale TransformX exchange.scale.com/public/ev… #AIExchange
Transfer learning might have more visibility than suggested here, but the points are great. Data are not always abundant. In cs229 at @Stanford we fine tuned a pre-trained ResNet-50 to identify commercials, with few examples.🤓
scientificamerican.com/artic… via @sciam
Menu French is above and beyond regular French.
This Post is from an account that no longer exists. Learn more







