ОБ ЭТОМ ЭПИЗОДЕ

английский
США

РАСШИФРОВКА 🔗

ПОИСК ПРОШЛЫХ ЭПИЗОДОВ

Искать прошлые эпизоды MTS.

ДРУГИЕ ЭПИЗОДЫ В ЭТОМ ПОДКАсте

Independent AI researcher and former ARIA program director David A. Dalrymple joins Theo and Sofia to argue that chain-of-thought monitoring is effectively dead and that AI alignment must pivot to formal verification and cultivating inherent “wisdom” during training. He also breaks down why he slas…
Cognition co-founder Walden Yan joins Madison Kanna to discuss the launch of Devin Bots, autonomous AI agents designed to act as proactive, persistent coworkers rather than simple task executors. Yan explains how Devin already writes virtually all of Cognition’s code—including building their Androi…
Ken Liu of First Person Labs joins MTS to discuss coming out of stealth with $10 million in funding and launching Second, a portable personal AI model built on raw computer activity that users fully own. He explains how persistent, user-controlled memory solves context fragmentation across frontier…
&Truth co-founders and co-CEOs Chris Hsu and Clark Barrett explain how formal verification and automated reasoning can mathematically prove software is bug-free in an era of AI-accelerated cyberattacks. They discuss their $70 million launch out of stealth, their public “And Atlas” effort to sec…
Perplexity Frontier Research lead and NYU professor Andrew Gordon Wilson discusses why AI systems are still far from acting as autonomous research scientists. He explains how Perplexity is using 10,000 GPUs, open science, model orchestration, and simulated worlds like Discover Physics to train mode…
Отказ от ответственности: Подкаст и иллюстрации, встроенные в эту страницу, взяты из MTS , который является собственностью его владельца и не связан или не одобрен Listen Notes, Inc.