filmov
tv
E882: TWiST Live! OpenAI Greg Brockman on rapid acceleration toward artificial general intelligence
Показать описание
E882: TWiST Live! OpenAI Co-founder & CTO Greg Brockman on our rapid acceleration toward artificial general intelligence: shares achievements in deep learning, the death of Moore's Law (300,000x compute increase), & the risks of malice, competition & unintended consequences as OpenAI races to build AGI that benefits humanity
Follow us on...
Show notes:
0:45 - Jason introduces Greg Brockman and OpenAI.
1:58 - Greg on OpenAI’s mission, why OpenAI is a not-for-profit company, and how some technological advantages are damaging to society
11:18 - GPUs powering AI, and the death of Moore’s Law
18:16 - Setting the stage for neural networks and deep-learning, and real-life examples of training neural networks
23:59 - Dota Project, the spectrum of narrow and general AI, and programming an OpenAI robotic hand
31:32 - What can we achieve if we the advances made by Neurolink (founded by Elon Musk) and OpenAI come together?
36:21 - Google buying DeepMind, how far along Google is with AI advancements, and the contributions of machine-learning and AI communities
40:17 - Timeline of AI: from figuring out simple tasks (video games) to curing diseases to becoming sentient
48:02 - What are the three categories of risk in AI?
57:47 - Dual use of AI and the growth of computation in the field of AI (in practice)
1:02:40 - Impact of quantum computing and other breakthrough technologies on the progress of AGI, China’s progress in the field, and engagement of the public sector in America
1:09:42 - Audience questions: Identifying culturally sensitive ideas in AI systems; coming to a consensus of “human values” in the post-AGI world; avoiding consequences when AI surpasses human intelligence; and AI policy in America vs. China
Комментарии