S3 E3 Scale Founder CEO Alexandr Wang: in AI your data writes the program
ฝัง
- เผยแพร่เมื่อ 3 มิ.ย. 2024
- S3 E3 Scale Founder CEO Alexandr Wang: in AI your data writes the program (Host: Pieter Abbeel)
What's in this episode:
00:00:00 - Alexandr Wang
00:02:16 - sponsors: Index Ventures, Weights and Biases
00:03:56 - mission of Scale: accelerating the development of AI
00:09:40 - early days of Scale, data as programming
00:13:40 - defensibility through marriage of technology and operations
00:18:56 - training NNs to accelerate labeling
00:24:19 - early customers
00:27:15 - labeling playbook
00:30:15 - AI for defense
00:37:27 - satellite imagery
00:41:00 - foundation models, LLMs, truthfulness, RLHF
00:59:38 - prompting as programming “cognitive computers” / LLMs
01:05:23 - industries to be revolutionized by AI
01:09:29 - what Alexandr Wang does to relax
Guest Links:
scale.com/
Sponsor Links:
www.indexventures.com/
wandb.com
Subscribe and listen to our podcast today:
Apple: bit.ly/3M5WMdb
Spotify: bit.ly/3nsTE0w
Amazon: bit.ly/3G30lwS
Google: bit.ly/3JU8Ds3
Acast: bit.ly/40pOLE6
Host: Pieter Abbeel
Production: Bo Obradovic. - วิทยาศาสตร์และเทคโนโลยี
Keep up the good work Alexandr!
Great series of interviews, amazing to be able to hear these conversations from people who are actually revolutionizing the world
Fantastic questions by the way, you earned a new subscriber!
Finally Hero of mine
A man of wisdom given by the lord
Your voice sounds like Rich Roll sometimes. I thought it was him until I looked at the video. Haha, keep up the excellent work, you are in an amazing niche right now.
Great convo, thank you. Realistic answer on military though it made me queasy…
Love this guy
TL;DR ->
Alexander Wang is a self-made billionaire and AI visionary who founded Scale AI in 2016. Scale AI provides AI models and products to businesses for specific use cases, and has been successful in creating a defensible business model. AI has drastically reduced the barriers of entry for many problems, allowing for the true democratization of AI. Peter hosts a podcast about AI innovations, and appreciates Alex for sharing his story and vision for the future of AI.
RLHF guidelines... I'm scared how underlying non conscious human bias is feeding into these models and the level of discretion being taken to stay attuned to the 'ethical performance' of them as they are more and more trained
These defensive technologies have massive implications. Ultimately I think Alex is right with a lot of his observations, but there is so much US centric good guy bad guy rhetoric mixed in that it's concerning. Why is this not a global conversation?
Repeats: like, sort of like, hundreds of times. Distracts from an otherwise great interview.