The AI reading guide
AI discussions often mix four different questions: what systems can do, what they can safely do, what infrastructure they require, and whether their economics work. This collection separates those questions so you can compare the arguments without treating a prediction as an established result.
Start with the infrastructure interview for the physical constraints, move to agents for practical uses and permissions, then examine research automation and the financial case. Each article links its key points to the original discussion so you can check the speaker’s context.
Jensen Huang discusses the hardware, software and data-center design behind large AI systems. Read his account as a participant in the industry, then compare it with the financing questions below.
Jensen Huang: NVIDIA - The $4 Trillion Company & the AI Revolution | Lex Fridman Podcast #494Peter Steinberger discusses OpenClaw and software agents. The useful questions are what an agent can do, which permissions it needs, and how people check its work.
OpenClaw: The Viral AI Agent that Broke the Internet - Peter Steinberger | Lex Fridman Podcast #491Ryan Greenblatt examines research automation and the difficulty of evaluating increasingly capable systems. Distinguish demonstrated research tasks from forecasts about future capabilities.
Ryan Greenblatt – What happens once AI can automate AI research?Brad Gerstner connects AI investment with revenue, computing capacity and power constraints. His market outlook is an argument to examine, not a guarantee of future returns.
Brad Gerstner: No AI Bubble, Semis Eat the Nasdaq & AI's Take Off Problem