Scaling LLMs and Accelerating Adoption: Interview with Aidan Gomez
HTML-код
- Опубликовано: 4 авг 2024
- On this episode, we’re joined by Aidan Gomez, Co-Founder and CEO at Cohere. Cohere develops and releases a range of innovative AI-powered tools and solutions for a variety of NLP use cases.
Thanks for listening to the Gradient Dissent podcast, brought to you by Weights & Biases. If you enjoyed this episode, please leave a review to help get the word out about the show. And be sure to subscribe so you never miss another insightful conversation.
Chapters:
0:00 Intro
1:20 Aidan’s role in the “Attention Is All You Need” paper
10:52 What SSMs are & how they could be an alternative to transformers
14:30 What it means for an ML architecture to saturate compute
21:36 Data constraints for when LLMs scale
27:00 Challenges of measuring LLM performance
36:08 How Cohere is positioned within the LLM development space
45:00 Scaling down an LLM into a more domain-specific one
50:08 Concerns around synthetic content & AI changing public discourse
57:17 The importance of raising money at healthy milestones for AI development
Resources:
- cohere.ai/
- research.google/pubs/pub46201/
#OCR #DeepLearning #AI #Modeling #ML - Наука
Such great questions. Great content.
Great interview. Touching on so many important aspects. A must watch. Thanks guys.
Thank you! Glad you are enjoying our content.
Loved it
Thanks Arjun, glad you enjoyed it!
I think for many cases it will be better to use a system that loads small models when needed. Especially for systems with only one or a few users, like embodied AI.