Jason A. Hoffman

Part 10 in the series. Previously: Computational Strategy (Part 9), Model Eats Software (Part 8), On Keeping AI in the Critical Path (Part 7), The Confident Incompetence Problem (Part 6), The Disintermediation Principle (Part 5), Zen of Unix Tools (Part 4). You want to build a home. But then you start building a house, which…

Part 9 in a series. Previously: GUTENBERGINLINETOKEN0END (Part 8), GUTENBERGINLINETOKEN1END (Part 6), GUTENBERGINLINETOKEN2END (Part 6b), GUTENBERGINLINETOKEN3END (Part 7), GUTENBERGINLINETOKEN4END (Part 5), GUTENBERGINLINETOKEN5END (Part 4). The Definition Strategic analysis often ends as a slide deck supported by spreadsheets. The deck presents a recommendation, while formulas, source choices, and assumptions remain difficult to inspect. When evidence changes,…

February 24, 2026 Two recent pieces reach different conclusions about AI infrastructure and employment. Citadel Securities published “The 2026 Global Intelligence Crisis”, which argues that AI displacement risk is overstated and adoption follows familiar S-curves. Citrini Research published “The 2028 Global Intelligence Crisis”, a bear-case scenario in which rapid white-collar displacement reduces consumer spending and…

A case study in why model critique is not verification Originally published at https://the-ai-dialogues.com/commentary/confidence-all-the-way-down I asked a frontier model to represent parts of the United States Constitution in Alloy, SPARK/Ada, and TLA+ and check stated properties with the corresponding tools. The model delivered 1,969 lines: Alloy for relational constraints, SPARK/Ada for contracts, and TLA+ for…

Fluency does not establish applicability Originally published at https://the-ai-dialogues.com/commentary/the-confident-incompetence-problem Models can propose sophisticated systems before they have established that the problem requires them or that the implementation will work. Optimization makes the error easy to see. Ask how to allocate resources under uncertain demand, and a model may immediately recommend a stochastic mixed-integer program, Pyomo,…

Part 8 in the series. Previously: [Zen of Unix Tools], [The Disintermediation Principle], [Keeping AI in the Critical Path]. In August 2011, Marc Andreessen argued that software companies would displace incumbents across much of the economy. The essay captured an important shift, but the companies it named did not all become durable winners. Groupon, Zynga,…

What cheaper cross-referencing changes, and what it does not. Large corpora change the work Law, medicine, software, finance, and science all produce more relevant material than one person can hold in working memory. Experts cope by specializing, searching, sampling, and relying on summaries. Each method can omit a relationship that crosses its chosen boundary. Large-context…

Calling AI a bubble leaves two questions unanswered. Which asset is overpriced, and what event would force its price back toward the cash flows it can produce? At the time of writing, private model companies carry large valuations, Nvidia can gain or lose hundreds of billions of dollars around an earnings report, and infrastructure spending…

Harvard Business Review is running sponsored content titled “A Blueprint for Enterprise-Wide Agentic AI Transformation.” Accenture has committed $3 billion to its AI practice, McKinsey’s QuantumBlack reports 5,000 AI specialists, and Wipro has combined advisory, AI, and enterprise-transformation services in a new operating model. The word transformation causes a practical problem. It names a desired…

Or: I spent six weeks compensating for what Claude Code couldn’t do yet, and then it could. On Christmas Eve 2025 I started a repo called agentic. The first commit was 2,312 lines across 19 files: 14 agent roles — frontend engineer, security engineer, QA, product manager, technical writer — plus orchestration concepts, communication protocols,…