I’ve led companies through every major tech disruption. AI washing is the same mistake, every time

The article argues that AI is eliminating tasks rather than jobs, citing Anthropic research showing no statistically significant increase in unemployment among highly AI-exposed occupations since ChatGPT launched. It highlights Cornerstone data showing positive demand growth across 15 of 16 occupational categories and demand exceeding supply by an average of 3.2x, suggesting a talent shortage rather than displacement. The piece is constructive on AI’s productivity impact but is mainly commentary, so direct market impact should be limited.

Analysis

The market is still pricing AI as a near-term labor shock, but the more important equity implication is margin expansion without commensurate headcount compression. That favors platforms that can sell workflow automation into large enterprises while monetizing the reallocation of labor budget toward higher-value tasks; the real earnings lever is not job destruction, it is a slower burn in hiring demand combined with higher software attach rates. In that setup, software, IT services, and human-capital platforms with strong distribution can all benefit before any broad macro labor dislocation shows up.

The second-order loser is not labor itself but companies whose value prop is tied to low-skill throughput or static training content. If AI pushes firms to embed learning into workflow, vendors selling episodic coursework, generic certification, or commoditized BPO-style execution face pricing pressure and churn risk over the next 12-24 months. Separately, employers that fail to build manager capability will see AI adoption stall at the pilot stage, which creates a hidden execution gap between “AI spend” and realized productivity.

The contrarian read is that consensus is probably overconfident on immediate workforce cuts and underconfident on the productivity lag. In the next 1-2 quarters, the headline risk is not unemployment but disappointment: companies may keep costs flat while failing to translate AI into visible EBITDA uplift, which can compress multiples for the most crowded AI beneficiaries. The better trading window is to own the picks-and-shovels that enable adoption discipline, while fading names that rely on the narrative of labor replacement rather than measurable workflow gains.

Key Decisions for Investors

• Long MSFT / GOOGL on a 3-6 month horizon: both monetize enterprise AI through workflow integration rather than headcount headlines; attractive if the market rotates from speculative model wins to durable distribution and attach-rate gains.
• Long DDOG or NOW on dips; time horizon 6-12 months: if AI is truly task-level, observability and workflow orchestration become the bottleneck, supporting premium multiples. Risk/reward improves if enterprise AI spend broadens beyond pilots.
• Short low-end training/content names or basket the theme via public comparables with heavy exposure to generic corporate learning; thesis over 6-12 months is that embedded, in-work learning disintermediates episodic courseware.
• Pair long enterprise software / short labor-arbitrage service exposure where available: own names with AI-enabled pricing power and short companies that depend on volume-based execution without differentiated judgment. Expect divergence over 2-4 quarters as productivity gains accrue unevenly.
• Avoid shorting broad labor markets solely on AI headlines; use any near-term selloff in staffing or HR-tech to cover shorts, since the next catalyst is likely adoption disappointment rather than an unemployment spike.