GPT-4 capability atlas and AGI sparks
Explore how early GPT-4 shows unprecedented cross-domain breadth but exposes profound architectural limits in planning and explanation fidelity.
Read the analysisResearch area
Reviews of the credibility, utility, and implications of research papers. Each current (2026) AI-assisted review uses a two-stage prompt. First, an editorial sweep examines what the paper actually says and extracts core claims. Second, a research review pass creates the raw material for each note. Besides the claims, I also include a bit about how this paper fits into the broader literature, remaining information gaps, and the paper's broader implications.
Explore how early GPT-4 shows unprecedented cross-domain breadth but exposes profound architectural limits in planning and explanation fidelity.
Read the analysisExplore how a GPT-3 assistant compresses productivity, accelerating novice learning and leveling the experience curve in customer support.
Read the analysisExplore how stated preference experiments reveal generative AI's massive consumer surplus, highlighting critical gaps in traditional GDP measurement.
Read the analysisExplore how volitional pragmatism reframes choice, belief updating, and public policy under deep indeterminacy and endogenous preferences.
Read the analysisExplore how dollarization and remittance dependence shaped El Salvador's Bitcoin adoption beyond political narratives and technological solutionism.
Read the analysisExplore how ecosystem graphs map foundation model dependencies, revealing supply chain opacity and providing a schema for AI transparency.
Read the analysisExplore how Bitcoin's protocol design reshapes transaction costs, risk allocation, and regulatory leverage through layered decentralization.
Read the analysisExplore how anonymous group cooperation explains early Bitcoin survival despite extreme mining concentration and vulnerability to majority attacks.
Read the analysisExplore how a rising Bitcoin price may redistribute wealth from latecomers to early holders even without a crash.
Read the analysisExplore how a global Delphi panel maps agreement and structured disagreement on the defining elements of transformational climate adaptation.
Read the analysisExplore how global datasets fail to track adaptation action, revealing a mismatch between risk measurement and policy progress under the GGA.
Read the analysisExplore how large-n adaptation studies measure reporting artifacts rather than outcomes, and why policy alignment offers a better evaluative path.
Read the analysis