Vulnerabilities in Deep Learning
Maciej discusses the various types of attacks that can be performed on deep learning models, including manipulating the model's behavior and fooling static models. He also highlights the vulnerability of language models and the potential for future concerns as these models become more powerful.In this clip
From this podcast
Data Skeptic
A Long Way Till AGI
Related Questions
Do you think deep learning is on the path towards AGI? Are these limitations you're describing just the challenges that we'll need to tackle, or are these insurmountable hurdles and we need another path to get to AGI?
Will large language models scale all the way to artificial general intelligence (AGI) as discussed in the episode Ilya Sutskever: Deep Learning | Lex Fridman Podcast #94 and the clip Deep Learning and AGI?
What does it take to build a system of human-level intelligence as discussed in the episode Ilya Sutskever: Deep Learning | Lex Fridman Podcast #94 and the clip Deep Learning and AGI?