The death of humanity

The author argues that AI has not produced detectable aggregate job losses, citing U.S. unemployment of 4.3%–4.5% through 2025 and the first three quarters of 2026, plus research finding faster headcount growth in AI-exposed industries. The piece presents AI as a potentially transformative productivity and science tool, while warning about misuse, overreliance, and capability transfer to bad actors. It also argues that frontier labs’ safety warnings may serve commercial and regulatory interests amid open-weight competition and model distillation; these are the author’s opinions, not reported market reactions.

Analysis

The investable tension is not “AI boom or bust”; it is who captures the surplus as model capability becomes cheaper. If open-weight systems narrow performance gaps, model access and inference pricing lose scarcity value, while distribution, proprietary data, workflow integration and cloud capacity matter more. That favors buyers and application companies able to switch models; it does not automatically favor every incumbent cloud provider, whose returns still depend on utilization and pricing. Open models may also increase usage enough to offset lower prices, but that is a utilization thesis—not proof that current AI capex earns attractive returns.

Treat the article’s labor and model-performance claims as hypotheses, not independent evidence. Aggregate employment and inflation cannot isolate AI’s effects; adoption could first slow hiring or wage growth in exposed roles. The author’s production-forecast claim is promotional and needs validation. Likewise, safety rules may serve public aims and raise compliance costs, even if incumbents also benefit from higher barriers to entry. Near term, this is opinion-driven sentiment, not a standalone catalyst. Over 1–3 months, watch model-switching evidence, enterprise inference costs, cloud AI consumption and regulatory text. Over 6–18 months, the key test is whether cheaper models expand paid usage and productivity, or commoditize returns faster than demand grows. The bullish narrative is underpricing implementation, reliability and liability risks; the bearish narrative may underprice the demand unlocked by lower costs.

Key Decisions for Investors

• No trade on the commentary alone. Treat any near-term AI sentiment move as an opportunity to assess price and positioning, not confirmation of the article’s claims; require independently verified enterprise adoption and economics before adding exposure.
• Prefer a conditional long MSFT on weakness over a broad “AI model winner” bet: the thesis is that cloud distribution and workflow integration can capture value even if model prices fall. Verify Azure AI consumption, gross-margin direction and capex productivity; step aside if usage growth weakens while infrastructure spending remains elevated.
• Watch META as a potential beneficiary of open-model distribution, but do not assume open weights translate into monetizable advantage. Reassess if engagement, ad efficiency or developer adoption improves; the thesis is falsified if model commoditization mainly raises serving costs or enables substitutes without measurable business gains.
• Risk monitor: track proposed compute licensing and export controls, plus evidence of harmful model use. Rules that impose fixed compliance costs could favor large labs, but tighter access or customer restrictions could also slow adoption across GOOG, META and MSFT; reassess on actual legislation and company guidance, not safety rhetoric.