Gradient Dissent - A Machine Learning Podcast avatar

Dexa/Gradient Dissent - A Machine Learning Podcast

Learn more
Published Dec 22, 2022

Jerome Pesenti — Large Language Models, PyTorch, and Meta

Jerome Pesenti delves into the transformative power and challenges of large language models, highlighting AI's role in revolutionizing drug discovery, education, and AR/VR interfaces, alongside insights into PyTorch's evolution at Meta and the importance of addressing bias and scalability issues.
Episode Highlights
Gradient Dissent - A Machine Learning Podcast logo

Topics covered

Popular Clips

Questions from this episode

Episode Highlights

  • Drug Discovery

    The potential of AI in drug discovery is immense, yet the field remains largely unchanged. believes a tech-driven revolution, akin to Tesla's impact on the automotive industry, is necessary to transform drug discovery. He highlights the current reliance on traditional decision-making processes, which often lack data support 1.

    I think drug discovery needs a bit of what I would call like a Tesla revolution, which is you need kind of a tech company to take it head on.

    ---

    AI's role in designing RNA-based medications and other processes shows promise, but a comprehensive overhaul is needed to significantly reduce costs and improve efficiency 2.

       

    Education Transformation

    Education is ripe for transformation through AI, yet it lags behind other sectors in technological integration. notes the stark contrast between traditional education methods and the engaging, personalized content found on platforms like TikTok 3.

    Education is completely outgunned today. So if you are a teenager, right, do you want to go to a boring lecture or you want to go on TikTok?

    ---

    He envisions AI-driven systems that adapt to individual learning styles, offering a more compelling educational experience. YouTube's success as an educational resource hints at the potential for AI to revolutionize learning, particularly for adult education 4.

       

    AR/VR Interfaces

    AI is pivotal in developing future AR/VR interfaces, with predicting a shift from traditional screens to more immersive experiences. He describes the challenges of creating intuitive interfaces for devices like AR glasses, which lack conventional input methods 5.

    My prediction to you is like in 30 years, people will look back and say, my God, this is like the stone age of interfaces.

    ---

    AI must facilitate a multimodal interaction system, incorporating voice, gestures, and motion to understand user intent. This complex integration aims to create a seamless user experience, moving beyond the limitations of current technology 5.

Related Episodes