Coding, Hallucinations, Multimodality and AGI: a fireside chat on Generative AI with Laurent Sorber
71 mins
In this episode, Davio Larnout, CEO at Radix, sits down for a fireside chat on Generative AI with Radix CTO Laurent Sorber. They reflect on the journey since starting Radix over five years ago and how AI has evolved in that time. What strides has Generative AI made? What lies ahead for AI, and how do we approach its challenges? Should you be scared of AI? Join the Radix founders as they share their insights in the latest episode of Future Forward.
Watch the episode
Main takeaways:
1. Large Language Models Lack True Reasoning Capabilities:
Laurent highlighted the limitations of current AI models, stating, "I would say there are several things that don't feel quite right with large language models... they don't have the ability to have an internal dialogue so there is no way for it to take some time to think before it produces an answer."
2. The Evolution of Jobs in the AI Era:
Laurent believes that while AI will transform the nature of work, it won't necessarily reduce the number of jobs. He mentioned, "I think most jobs or most people will still be doing jobs still be producing value in the economy."
3. AI's Potential to Boost Economic Prosperity:
Envisioning a future where AI enhances productivity and living standards, Laurent said, "I think large language models will essentially enable a higher standard of living enable more prosperity by increasing economic outputs."
4. Questioning the Uniqueness of Human Intelligence:
As AI continues to advance, Laurent reflects on the nature of human cognition, suggesting, "Perhaps we might not be as impressed with AGI as that we might be disappointed with our own cognitive abilities."
5. The Rise of Personalized AI Assistants:
Predicting a transformative role for AI in sectors like education, Laurent foresees a future where everyone has access to their own AI assistant, similar to Jarvis from Iron Man. He emphasized, "What I do think will happen is that everyone will get access to a personal Jarvis... especially in education for example I think that could have a huge impact on giving everyone access to a personal tutor."