| Tweetovi |
|
Greg Brockman
@gdb
|
31. sij |
|
That'd be totally fine! Per our post: "Going forward we’ll primarily use PyTorch as our deep learning framework but sometimes use other ones when there’s a specific technical reason to do so".
|
||
|
|
||
|
Greg Brockman
@gdb
|
30. sij |
|
Thanks Schrep! Very excited as well :)
|
||
|
|
||
|
Greg Brockman
@gdb
|
30. sij |
|
OpenAI → PyTorch: twitter.com/OpenAI/status/…
|
||
|
|
||
| Greg Brockman proslijedio/la je tweet | ||
|
Ars Electronica
@ArsElectronica
|
17. sij |
|
Composing a piece of music is only for pros? We tried for you how to compose music with “MuseNet” from @OpenAI and made a step-by-step guide based on it. If you want to learn more, participate in the “MuseNet Special” workshop this weekend.
ars.electronica.art/aeblog/en/2020… pic.twitter.com/TQe1QFXdXH
|
||
|
|
||
| Greg Brockman proslijedio/la je tweet | ||
|
Berkeley School of Information
@BerkeleyISchool
|
13. sij |
|
.@alexstamos kicks off his keynote address with an ML-generated speech
"I trained OpenAI’s GPT-2 text-generation neural network on the celebrated graduation speeches of Oprah Winfrey, Steve Jobs & David Foster Wallace, & I let the AI write my opening paragraph"
#ischoolgrad🎓 pic.twitter.com/UXbLkkXhfY
|
||
|
|
||
| Greg Brockman proslijedio/la je tweet | ||
|
Elliot Turner
@eturner303
|
12. sij |
|
The funny thing about AIDungeon is due to all the GPT-2 pre-training on swaths of random web text, you can create an adventure game about literally anything. Here's one about installing OpenCV from source: pic.twitter.com/zJvtjBwcbq
|
||
|
|
||
|
Greg Brockman
@gdb
|
9. sij |
|
Watched "The Squire — AI Written Short Film".
It's a live-action adaption of AI Dungeon gameplay. Super fun to see!
youtube.com/watch?v=Kx-2Py… pic.twitter.com/RZVLXMcYA8
|
||
|
|
||
|
Greg Brockman
@gdb
|
6. sij |
|
Welcome, @jeffclune! twitter.com/jeffclune/stat…
|
||
|
|
||
|
Greg Brockman
@gdb
|
3. sij |
|
Working with Jakub Pachocki (@merettm) is one of my favorite parts of OpenAI. Highly recommend watching his talk! twitter.com/lexfridman/sta…
|
||
|
|
||
|
Greg Brockman
@gdb
|
31. pro |
|
(Correction after talking to the artists: the lyrics were generated with GPT-2, but the melody was human-written, not MuseNet. Hopefully MuseNet next time :).)
|
||
|
|
||
|
Greg Brockman
@gdb
|
30. pro |
|
A cool Christmas song, powered by GPT-2 and MuseNet: twitter.com/BlueThalo/stat…
|
||
|
|
||
|
Greg Brockman
@gdb
|
30. pro |
|
I agree with this article that AI Dungeon 2 (the GPT-2 powered text-based game of infinite possibilities) is "one of the coolest video game experiments of 2019". You should definitely check it out: play.aidungeon.io. Great work @nickwalton00! twitter.com/verge/status/1…
|
||
|
|
||
| Greg Brockman proslijedio/la je tweet | ||
|
Robert Laidlow
@Robert_Laidlow
|
28. pro |
|
Wondering how I fused Bach and AI for tonight’s @Ri_Science #XmasLectures with @FryRsquared and @Chineke4Change? Here’s a short thread – read on! It uses an amazing music-generating #AI coded by @mcleavey from @OpenAI called #MuseNet. 1/6 twitter.com/Ri_Science/sta…
|
||
|
|
||
| Greg Brockman proslijedio/la je tweet | ||
|
Renee DiResta
@noUpside
|
20. pro |
|
And this is why the next significant battle is going to be around identity.
Textfakes - speechfakes - are much more of a significant near-term challenge than video deepfakes (no detecting artifacts/snafus in visuals). Text creates an impression that a lot of ppl believe a thing. twitter.com/fredbenenson/s…
|
||
|
|
||
|
Greg Brockman
@gdb
|
19. pro |
|
Feel free to send to me and I can route it appropriately!
|
||
|
|
||
| Greg Brockman proslijedio/la je tweet | ||
|
Maybe: Fred Benenson
@fredbenenson
|
19. pro |
|
A Harvard undergrad used OpenAI’s GPT-2 to generate deepfake text comments for a political campaign. The results are astounding; some bot comments fooled 100% of the judges: techscience.org/a/2019121801/?… This is exactly the dystopian use of GPT2 thatOpenAI was concerned about.
|
||
|
|
||
| Greg Brockman proslijedio/la je tweet | ||
|
Dota 2
@dota2updates
|
13. pro |
|
Dota 2 with Large Scale Deep Reinforcement Learning
cdn.openai.com/dota-2.pdf
Via @OpenAI
#Dota2 pic.twitter.com/yT0GLiTv56
|
||
|
|
||
|
Greg Brockman
@gdb
|
13. pro |
|
Thanks!
I'm curious what reward function details you found in there that were missing from the description linked in the original OpenAI Five post: gist.github.com/dfarhi/66ec9d7…
|
||
|
|
||
|
Greg Brockman
@gdb
|
13. pro |
|
Thank you! Means a lot coming from the best team in Dota 2 history :). twitter.com/OGesports/stat…
|
||
|
|
||
|
Greg Brockman
@gdb
|
13. pro |
|
We just released our scientific analysis of OpenAI Five: twitter.com/OpenAI/status/…
We are already using findings from Five in other systems at OpenAI like Dactyl (openai.com/blog/solving-r…) or our multi-agent work (openai.com/blog/emergent-…).
Hope that others find the results useful!
|
||
|
|
||