What should AI practitioners prioritize for robust AI systems?

Carly Taylor

VP, Engineering (Concepts Lab) · Crunchyroll

Fundamentals

Carly Taylor says AI practitioners should prioritize fundamentals, because she says there are fundamental truths that will always be true and some things never change.

The answer starts at 6:09.

Reliability and controllability

Ilya Sutskever says he is interested in making AI models more reliable and more controllable, and also faster to learn from less data and instruction without hallucinating, saying these questions are connected.

RS

Richard Socher

CEO · Recursive

Clear reward specification

Richard Socher says AI practitioners should prioritize clearly articulating rewards, because AI may optimize the literal reward rather than what was meant, and humanity hasn't gotten good at this.

Akhil Verghese

Chief Executive Officer · Krazimo

Human oversight

Akhil Verghese says the ultimate guardrail for robust AI systems is human oversight, because trusting these systems too much leads to confirmation bias.

Ben Prescott

Head of AI Solutions, Digital Consulting · Trace3

Develop robust intake processes

Ben Prescott says AI practitioners should slow down early and be methodical, building a framework or intake process for gathering demand, because that process becomes repeatable and lets the business move quickly over time.

Thomas Wolf

Co-founder - Chief Science Officer · Hugging Face

Global monitoring

Thomas Wolf says AI practitioners should build a global monitoring setup, because it is hard to be fully sure what a swarm of agents is doing and they need to zoom out to see what's happening in every direction.

John Brewton

Writer & Founder · Operating by John Brewton

Building what couldn’t be built before

John Brewton says the useful question is what can we build now that we couldn’t before, rather than what model launched, because AI moves too quickly for static prompt tricks.