- interesting read on how the GPT 4 model that we have today was tuned to be safer, and how various safety challenges were mitigated.
- At the same time, it shows what a model without those safety restrictions is capable of. And, in the wrong hands, it could cause much destruction in the form of mass disinformation campaigns, phishing, etc. Some examples:
- Who’d think AI models would trigger an arms race among countries in the near future. The paper even mentions that the models can be used for “both commercial and military or proliferation applications”.
Lex Fridman Podcast - #367 – Sam Altman: OpenAI CEO on GPT-4, ChatGPT, and the Future of AI
Sam Altman is the CEO of OpenAI, the company behind GPT-4, ChatGPT, DALL-E, Codex, and many other state-of-the-art AI technologies. Please support this podcast by checking out our sponsors: - NetSuite: http://netsuite.com/lex to get free product tour - SimpliSafe: https://simplisafe.com/lex - ExpressVPN: https://expressvpn.com/lexpod to get 3 months free EPISODE LINKS: Sam's Twitter: https://twitter.com/sama OpenAI's Twitter: https://twitter.com/OpenAI OpenAI's Website: https://openai.com GPT-4 Website: https://openai.com/research/gpt-4 PODCAST INFO: Podcast website: https://lexfridman.com/podcast Apple Podcasts: https://apple.co/2lwqZIr Spotify: https://spoti.fi/2nEwCF8 RSS: https://lexfridman.com/feed/podcast/ YouTube Full Episodes: https://youtube.com/lexfridman YouTube Clips: https://youtube.com/lexclips SUPPORT & CONNECT: - Check out the sponsors above, it's the best way to support this podcast - Support on Patreon: https://www.patreon.com/lexfridman - Twitter: https://twitter.com/lexfridman - Instagram: https://www.instagram.com/lexfridman - LinkedIn: https://www.linkedin.com/in/lexfridman - Facebook: https://www.facebook.com/lexfridman - Medium: https://medium.com/@lexfridman OUTLINE: Here's the timestamps for the episode. On some podcast players you should be able to click the timestamp to jump to that time. (00:00) - Introduction (08:41) - GPT-4 (20:06) - Political bias (27:07) - AI safety (47:47) - Neural network size (51:40) - AGI (1:13:09) - Fear (1:15:18) - Competition (1:17:38) - From non-profit to capped-profit (1:20:58) - Power (1:26:11) - Elon Musk (1:34:37) - Political pressure (1:52:51) - Truth and misinformation (2:05:13) - Microsoft (2:09:13) - SVB bank collapse (2:14:04) - Anthropomorphism (2:18:07) - Future applications (2:21:59) - Advice for young people (2:24:37) - Meaning of life
- this one was a light listen, more of a high level discussion. if you’re expecting some technical deep dives, you’d be disappointed.