Expert Panel
Why each voice is on the panel
Daniel Miessler
Miessler thinks in systems and second-order effects. He connects personal AI infrastructure, security posture, and the changing shape of human work. His value is framing: he names shifts before they are obvious and is unusually clear on what individuals and orgs should actually do.
Nate B. Jones
Jones translates raw AI developments into executive-grade strategic implications daily. He is the closest analog to what AI Signal aims to be: high-signal, business-first, allergic to hype. Strongest source for 'what does this mean for an enterprise buyer / seller this week.'
AI Explained
Runs a rigorous 'what does this actually mean' pass on every major model release, benchmark, and capability jump — reads the papers and probes the demos so you don't have to. Fills the technical ground-truth role Karpathy occupied, but with far higher frequency. When a frontier lab ships, you get a sober assessment within 24 hours instead of PR spin.
Dwarkesh Patel
Dwarkesh interviews the people building the frontier and presses on timelines, economics, and second-order societal effects. Best source for directional forecasting and the economic/operational impact arc that enterprise leaders need to plan against, not just react to.
Matthew Berman
Berman is the hands-on implementation lens: new tools, agent frameworks, and what actually ships and works. Grounds the panel in practitioner reality — what a team could deploy this quarter — which is exactly the layer BlueAlly customers ask about.
Latent Space
Practitioner podcast where the people actually shipping enterprise AI are interviewed at length — engineering leaders, tool authors, model teams. Deeper architectural discussions than the news feed: how does this run in production, at what cost, with what failure modes. Complements Berman's demo-driven coverage with insider technical and business conversations.
Simon Willison
Willison tests new models and tools directly and documents the failure modes that product demos usually omit. He is especially strong on prompt injection, agent security, local models, and what working developers can verify for themselves.
Hamel Husain
Husain focuses on the unglamorous work that separates production AI from demos: evaluation, error analysis, retrieval quality, and iteration against real user failures.
Nathan Lambert
Lambert connects open-model releases to the training recipes, evaluations, economics, and policy behind them. He adds technical depth without relying on frontier-lab marketing claims.
SemiAnalysis
The SemiAnalysis team grounds AI strategy in chips, serving performance, datacenter constraints, and cost curves. This is the infrastructure reality behind model and agent announcements.