drcwpl

18 ms·

211.
Karpathy on Tokenizers(twitter.com)
2 points by drcwpl 3y ago|1 comments
212.
AI-Generated Faceless Videos and Voice Cloning: Invideo AI(youtube.com)
1 points by drcwpl 3y ago|0 comments
213.
by drcwpl 3y ago
It is worth revisiting this talk by, Ilya Sutskever. He argues that AI safety should be a mainstream concern across the ML community (maybe wider community). He reiterates that recent advances in AI have made it more likely than ever that w
214.
by drcwpl 3y ago
AI alignment! It is not difficult to imagine AGI
215.
Ilya Sutskever – Opening Remarks: Confronting the Possibility of AGI [video](youtube.com)
3 points by drcwpl 3y ago|3 comments
216.
by drcwpl 3y ago
Wow - "All videos on this page were generated directly by Sora without modification." The prompts - incredible and such quality - amazing. "Prompt: An extreme close-up of an gray-haired man with a beard in his 60s, he is deep
217.
Not Without Us – Joseph Weizenbaum (1986)(dl.acm.org)
64 points by drcwpl 3y ago|20 comments
218.
by drcwpl 3y ago
Ha - funny :-)
219.
by drcwpl 3y ago
Were there any negations in the prompt? E.G. "Not" GPT has a terrible issue with negations - if you say for example create an illsutration of a hacker wearing a hoodie, without glasses." it will create with glasses - negation
220.
Antagonistic AI(arxiv.org)
78 points by drcwpl 3y ago|54 comments
221.
AI Could Help Rebuild the Middle Class(noemamag.com)
3 points by drcwpl 3y ago|2 comments
222.
by drcwpl 3y ago
I cannot agree with the author on this... I do not consider it realistic that it will be come an equalizer
223.
Shoshana Zuboff: A Trailblazer in the Digital Age(encyclopaediaautomatica.com)
2 points by drcwpl 3y ago|1 comments
224.
by drcwpl 3y ago
Zuboff's work has highlighted the ways in which surveillance capitalism poses a threat to privacy and democracy. She argues that the relentless collection and analysis of personal data by tech companies has the potential to undermine o
225.
OpenAI – Logan Kilpatrick (head of developer relations) [video](youtube.com)
2 points by drcwpl 3y ago|1 comments
226.
by drcwpl 3y ago
Prompt engineering ...adding a smiley face and telling the model to take a break increases the performance of the model
227.
by drcwpl 3y ago
#1: By 2050, we will have solved the scientific and technological limitations in current AI that are critical to enabling further breakthrough progress and powerful AI capable of realizing beneficial and exciting possibilities. #2: By 2050
228.
Ten Hard Problems in Artificial Intelligence We Must Get Right(arxiv.org)
4 points by drcwpl 3y ago|1 comments
229.
Direct Language Model Alignment from Online AI Feedback(arxiv.org)
61 points by drcwpl 3y ago|4 comments
230.
GoogleDeepMind Fractals and the intelligent behavior of LLMs(arxiv.org)
5 points by drcwpl 3y ago|0 comments
231.
Ask Huberman with Dexa.ai(ai.hubermanlab.com)
2 points by drcwpl 3y ago|0 comments
232.
by drcwpl 3y ago
Wow - great catch...
233.
by drcwpl 3y ago
Exactly - well said
234.
by drcwpl 3y ago
Fantastic, thank you - will watch it carefully... ha 7 years!
235.
by drcwpl 3y ago
Very true
236.
by drcwpl 3y ago
Excellent - would be interested to know how you get on with it.
237.
The first book ever sold by Amazon.com(en.wikipedia.org)
1 points by drcwpl 3y ago|0 comments
238.
Geoffrey Hinton – Will digital intelligence replace biological intelligence? [video](youtube.com)
7 points by drcwpl 3y ago|0 comments
239.
Gödel, Escher, Bach is the most influential book in my life (2022)(philosophygeek.medium.com)
388 points by drcwpl 3y ago|317 comments
240.
Analogy as the Core of Cognition – Doug Hofstadter [video](youtube.com)
2 points by drcwpl 3y ago|0 comments

More ›