Altman and Amodei Walk Back AI Job Apocalypse Before IPOs

OpenAI CEO Sam Altman told a Commonwealth Bank of Australia event in Sydney on May 26, 2026 that he was "pretty wrong" about AI eliminating more entry-level white-collar jobs by now than actually happened. Anthropic CEO Dario Amodei, who previously warned AI could cut half of entry-level white-collar roles, now describes automation as a productivity multiplier that can expand the work people do.

The timing is not subtle: OpenAI and Anthropic are both marching toward 2026 IPOs after Anthropic's $965 billion Series H and OpenAI's ~$852 billion private valuation, and institutional investors hate employment-apocalypse risk factors in S-1 filings.

What did Altman say on May 26?

Altman told CBA CEO Matt Comyn he expected more entry-level white-collar elimination than materialized and said he is "delighted to be wrong." Coverage in Reuters, Fortune, and HR trade press tied the comment to OpenAI's Q4 2026 IPO window narrative.

This follows his February 5, 2026 framing that some companies engage in "AI washing" of layoffs unrelated to real automation.

What changed in Amodei's framing?

Amodei's 2025 warnings included up to 50% entry-level white-collar disruption and double-digit unemployment risk scenarios. In May 2026 interviews he emphasized: if you automate 90% of tasks, humans focus on the 10% that expands into a fuller job with ~10x productivity.

That is economically plausible for some roles and empty for others. It is also IPO-friendly language.

We documented Amodei's earlier honest warning in Dario Amodei AI job disruption warning. This post is the narrative reversal chapter.

Why markets care more than truth

Genpact executive Vijay Vijayasankar argued on LinkedIn that services stocks sold off when Altman and Amodei preached apocalypse, because investors treated services ETFs as implicit shorts of frontier labs. Near IPO, both companies also sell enterprise services (implementation, safety, consulting), so stabilizing labor narratives protects multiple expansion.

Regulators read S-1 risk sections. "Permanent displacement of half the workforce" invites oversight; "productivity utility" invites pension funds.

What developers should actually measure

Ignore CEO mood cycles for career planning. Watch:

• WARN Act filings and sector hiring rates for entry-level SWE, support, and analyst roles
• Your team's ticket mix (agent-assisted vs fully automated)
• Tool budgets (LLM API pricing) shifting from copilot seats to autonomous agents

Use Will AI Replace Me as a structured quiz, not Altman's latest keynote adjective.

Contradictions to keep in mind

• Amazon killed an internal AI leaderboard over tokenmaxxing while pushing 80%+ weekly AI usage (Kirorank story)
• Meta and peers still cut headcount while selling AI augmentation
• Anthropic revenue exploded to a $47B run rate while CEOs say jobs are fine

Capital markets want growth without liability. Engineers want signal without gaslighting.

Key Takeaways

• May 26, 2026: Altman said he was wrong on near-term entry-level white-collar job destruction
• Amodei reframed AI as productivity expansion, walking back 50% job loss rhetoric
• IPO timing: OpenAI and Anthropic need institutional-friendly labor narratives in 2026 listings
• Developers: track hiring data and automation metrics, not executive reversals alone
• Cross-read: prior honest Amodei post + new Anthropic $965B funding
• What to watch: OpenAI S-1 risk-factor language on labor, WARN trends, post-IPO CEO statements after lockup

Frequently asked questions

What did Sam Altman say about AI and jobs in May 2026?

On May 26, 2026, Altman said he was pretty wrong about AI eliminating more entry-level white-collar jobs than has happened so far, and that he was delighted to be wrong.

Did Dario Amodei change his AI unemployment prediction?

Amodei previously warned AI could eliminate a large share of entry-level white-collar jobs. In May 2026 he emphasized productivity gains and expanding remaining tasks rather than mass unemployment framing.

Why are AI CEOs walking back job warnings now?

Analysts tie the shift to upcoming 2026 IPOs for OpenAI and Anthropic. Institutional investors prefer stable growth narratives over apocalyptic labor risk that could invite regulation and hurt valuations.

Should developers stop worrying about AI replacing jobs?

No. CEO statements are strategic. Use hiring data, team automation metrics, and skills planning tools rather than IPO-stage rhetoric alone.

Where can I compare AI models for my career and budget?

Use abhs.in tools: /tools/will-ai-replace-me for career framing and /tools/llm-api-pricing for API cost modeling alongside posts comparing frontier models.