Training Language Models and Controversies in Data Sets

10min Snip

00:00

Play full episode

Summary

Transcript

Episode notes

This chapter focuses on the training of language models, specifically GPT-2, and discusses the use of humans and bots in building the models. It also delves into the controversies surrounding training data sets for AI models, the legal issues that can arise, and the responsible practices of some open source models. Additionally, they explore the challenges in finding trainable clean data as the public internet closes down and discuss various data sets available for AI research.

This Week in Startups is brought to you by…

Brave. If you’re building AI and search-based applications, train your models with the Brave Search API. Get started for free at brave.com/jason

Arising Ventures is a holding company that acquires tech startups facing setbacks. Arising Ventures knows what founders care about because they aren’t bankers, they are tech founders themselves! Go to ⁠http://www.arisingventures.com/TWIST

LinkedIn Jobs. A business is only as strong as its people, and every hire matters. Go to http://www.LinkedIn.com/TWIST to post your first job for free. Terms and conditions apply.

Today’s show:

Adept Co-Founder and CEO David Luan joins Jason to discuss his experience at OpenAI (2:45), demo Adept's AI-powered workflow builder (13:47), what facets of AI he is most excited about, what keeps him motivated (32:33), and more!

Timestamps:

(0:00) Adept founder and CEO David Luan joins Jason.

(2:45) David's experience at OpenAI and the evolution of deep learning

(5:22) DALL-E origins and its "a-ha moment"

(9:24) The next great well of AI training data

(11:21) Brave - Try the Brave Search API at http://www.brave.com/jason

(12:35) Adept's Northstar since day one

(13:47) Demos: showcasing the power of Adept AI

(15:52) The next battlefield for AI

(18:34) A look into Adept AI Experiments – a sandbox for AI enthusiasts

(20:46) Arising Ventures - head to ⁠http://www.arisingventures.com/TWIST to learn more and connect with the team

(21:51) How Adept aims to streamline job processes