The year of scaling AI, with Cohere's Aidan Gomez
Ғылым және технология
In this conversation, I speak with Aidan Gomez, co-founder of Cohere and one of the brains behind generative AI.
00:00 Introduction
03:04 The origins of the transformer paper, “Attention is all you need”
11:05 Enterprises and large language models (LLMs)
30:22 Reliability and trust in LLMs
39:58 Components and infrastructure of LLM implementations
45:31 Performance improvements from LLMs
48:41 Impact on competitive dynamics
51:30 Augmenting, not displacing, human workers
57:00 Building apps with LLMs
01:00:32 Overcoming fear and nervousness around AI
01:05:55 The future of LLM architectures
01:08:30 Balancing level-headedness and rapid progress in AI
Where to find Aidan:
Linkedin: / aidangomez
X: / aidangomez
Cohere: www.cohere.com
"Attention is All You Need" paper: proceedings.neurips.cc/paper/...
Where to find Azeem:
Linkedin: / azhar
KZread: @azeemexponentially
Instagram: @azeem
X: @azeem
Website: www.azeemazhar.com
Пікірлер: 6
Excellent, illuminating discussion. And Aidan was dressed kind of like a priest or monk, so his words were imbued with a divine authority.
deserve to be shared, really insightful. thank you Aidan Gomez for sharing.
@AzeemExponentially
2 ай бұрын
Thanks for the kind words
pretty good, thanks
It was a great conversation, but a bit out of touch in some areas: - I think it overestimates the literacy of large portions of the population (employee or customer) in understanding a chatbot is not reliable and trustworthy. - For organizations with limited in house capability, a quarter of a million is really lean. It is not the models or tokens in a vacuum, it is everything else.
the host always interrupts the guest, which is quite annoying. Please let them finish their point before making your own