Luan: AI companies should eliminate dedicated safety teams to improve culture
David Luan · AI + You: 5 ways to ethically build — and use — AI | Masters of Scale · Oct 31, 2023 · at 17:48
David Luan, co-founder of Adept, discusses organizational design and safety culture at AI companies.
0:00 / 0:16exact quote · 16.1s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“I now think that the correct organizational structure for building an AI organization is actually to eliminate the concept of having a separate safety team. Because by creating a safety team, you're now defining this is the unsafe team, this is the safe team. We'll have the safe team keep the unsafe team in check.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from David Luan
Luan: AI cyber attack risk is 10x to 100x larger than disinformation
“I think it's like 10 x larger than the disinformation problem, if not a hundred x larger. In a world where these models are much more easily used as a mechanism for attack than a mechanism for defense, there's gonna be this period of time over the next couple …”
Luan: Ad-supported AI models become beholden to advertisers, not end users
“Having an advertiser be the financial party responsible for paying for these services means that ultimately these AI services are beholden to that advertiser and not to the end user.”
Luan: Assuming broad intelligence from conversational fluency is a major trap
“And I talked to very smart people who haven't seen these models be trained and haven't seen the objectives these models are trained on. That just automatically assume that because they've seen this sliver of intelligence from these models, that they therefore …”
Luan: Single piece of human feedback boosts LLM coding solve rates 20-30%
“In the process of solving a particular task, if a human gives the language model one piece of feedback, like you're writing code and you forgot to import the Python OS library, then the solve rate for these problems would jump up like 20, 30%.”