AI's real job market impact is still murky and contested
The 'what is happening to jobs' thread is one of the more honest conversations on HN today. The author notes that coding agents like Claude Code and OpenAI Codex only started working well in late November 2024, meaning most organizations experienced the real productivity shift in January 2025 at the earliest. That makes any study published before mid-2025 essentially pre-impact data.
The thread's most interesting signal is the 'solo builder' comment: one person says they've done enough work in the last 18 months to consider themselves equivalent to a small team, but their organization doesn't know how to evaluate or reward that. Organizational inertia is the key variable. Fortune 500 companies with internal AI bans are still common. The productivity gains are real but distributed so unevenly that aggregate statistics look flat.
The honest framing from multiple commenters: we are in a period where AI has clearly changed what one person can build, but hasn't yet changed how most companies hire or structure teams. That gap is where the disruption will actually happen, and it hasn't fully arrived yet.
So what?
Founders building for the enterprise should expect long lags between AI capability improvements and organizational adoption. The opportunity isn't just building better AI tools, it's building tools that help organizations change the processes around those tools. If you're hiring, the solo builder who's 5x productive is real, and your leveling rubrics probably don't capture them yet.