While the format and interview style of the Lex Fridman podcast is highly debated, I personally found great value and inspiration in his recent interview with world-class programmer John Carmack. I would recommend diving in to anyone interested in programming and cutting-edge technology.
The first half is a nerdy nostalgia trip talking about the history of game development (commander keen, wolfenstein, doom, quake) that shaped John to be able to work on really hard problems within restricted environments with limited resources. The second half is spent on his work at meta on the metaverse / VR, his view on aerospace and SpaceX endeavours, nuclear power and artificial general intelligence.
It is very refreshing to hear a no-nonsense pragmatic engineer talk about big complex challenges and discuss their technical feasibility within the current restrictions. I’m glad that John also concludes that most problems do not need hyper optimalisation, and that some problems will become solvable in time by increase of compute efficiency.