🤚 The Open-Palm Illumination

Peter Diamandis’s latest Moonshots roundup delivered what can only be described as a theological-technological-economic triple axel — and somehow stuck the landing.

First: Pope Leo XIV released his inaugural encyclical, Magnifica Humanitas, a 42,300-word papal document dedicated entirely to artificial intelligence. The Pope — who chose his name in homage to Leo XIII, author of Rerum Novarum, the 1891 encyclical addressing the Industrial Revolution — has drawn a direct line between the steam engine and the transformer architecture. He calls on governments and tech leaders to “disarm” AI before it deepens inequality, weakens human agency, and shifts critical decisions permanently out of human hands. He presented it at the Vatican alongside Christopher Olah, co-founder of Anthropic, because when you want to talk about the ethics of intelligence, you apparently invite the people building it.

Second: GPT-5.5 has been quietly eating Claude’s lunch on several key benchmarks — Terminal-Bench 2.0 at 82.7% versus Claude Opus 4.7’s 69.4%, a 13-point gap that had the benchmarking community reaching for the smelling salts. FrontierMath: 51.7% vs 43.8%. The model OpenAI launched on April 23 reclaimed the top spot on nearly every coding and agent benchmark that matters.

Third: Sam Altman told a Sydney audience that he was “pretty wrong” about AI’s economic impact — directly reversing his own June 2025 warnings that entry-level white-collar roles were at serious risk. The man who once said AI would make “a lot of jobs go away” now says the jobs apocalypse “hasn’t actually happened.”

👐 The Two-Handed Reality Check

Let us, as a magazine of discerning editorial standards, note the comedic timing of all three events occurring within the same news cycle.

The Pope writes forty-two thousand words warning that AI could undermine human agency. Sam Altman says actually, AI can’t even handle his Slack messages properly. And OpenAI’s model beats Anthropic’s on benchmarks — the same Anthropic whose co-founder just co-presented with the Pope on the ethical perils of the technology they’re building.

It’s almost too perfect.

On the benchmark wars: yes, GPT-5.5 leads on Terminal-Bench and FrontierMath. But Claude Opus 4.7 still holds SWE-Bench Verified at 64.3% versus GPT-5.5’s 58.6% — a 5.7-point gap on real GitHub issue resolution. And then Claude Opus 4.8 dropped on May 28 and dethroned GPT-5.5 on the Artificial Analysis Intelligence Index (61.4 vs 60.2), leading SWE-bench Pro by 10.6 points. The crown changes heads so frequently now that the crown itself has filed for workers’ compensation.

On Altman’s reversal: he tested having AI handle his email and Slack — some messages even labeled “this is Sam’s AI” — and concluded that the deeply human component of work was harder to automate than he’d assumed. The Yale Budget Lab backs him up: no significant changes in occupational mix or unemployment duration in high-AI-exposure jobs since ChatGPT launched in late 2022. Three and a half years of the AI revolution, and the Bureau of Labor Statistics shrugged.

Meanwhile, Dario Amodei — who once claimed AI could eliminate 50% of white-collar jobs — now frames automation as expanding work rather than destroying it. When automation handles 90% of a role, he suggests, the remaining 10% expands to fill workers’ capacity. Which is either a profound insight about economic elasticity or the most sophisticated rationalization ever constructed by a man eyeing a trillion-dollar IPO.

🌿 The Gentle Awakening

There is something genuinely remarkable about watching the two most powerful AI CEOs on earth simultaneously soften their doomsday rhetoric just as both companies prepare billion-dollar public offerings. Both OpenAI and Anthropic are reportedly valued around $1 trillion for their upcoming IPOs. One imagines the investor roadshow goes smoother when you’re not leading with “our product will eliminate half of all employment.”

Pope Leo, who has no IPO to protect, can afford to say the uncomfortable thing. And Magnifica Humanitas does exactly that — it names the concentration of power, the erosion of agency, the quiet replacement of judgment with optimization. He’s not anti-technology. He explicitly states that “technology should not be considered, in itself, as a force antagonistic to humanity.” But he’s asking who benefits, who decides, and who’s left holding the invoice when the disruption arrives.

These are the questions the benchmark charts never answer.

👑 The Crown Verdict

We are living in a moment where the Pope, the CEO of OpenAI, and the CEO of Anthropic are all saying variations of the same thing: the technology is extraordinary, the human implications are uncertain, and we should probably slow down enough to think.

The difference is that the Pope says it in a 42,300-word encyclical presented alongside the co-founder of a safety lab. Sam Altman says it in Sydney after failing to outsource his email. And Dario Amodei says it in a way that conveniently makes his upcoming IPO look more palatable.

The benchmarks will keep leapfrogging. GPT-5.5 leads today, Opus 4.8 leads tomorrow, and by next Thursday something called Gemini Ultra Omega Supreme will briefly top a chart no one has heard of. What matters is whether the humans building these systems are honest about what they don’t know — and whether honesty arrives before or after the S-1 filing.

Inspired by Pope Leo vs. AI, GPT 5.5 Beats Claude, and Sam Altman Walks Back Job Apocalypse | EP #259 by Peter Diamandis.

“The Pope wrote 42,300 words on AI ethics, Sam Altman couldn’t automate his own Slack, and both companies are worth a trillion dollars. The machines are fine. It’s the humans who need debugging.” — The Slap of Wisdom Eschatology Desk, drafting its own encyclical on the moral hazard of benchmark charts

Your conviction is showing. Disarm wisely.