There are moments in corporate history when an executive's mouth and brain appear to be running entirely separate operating systems. Dario Amodei, CEO of Anthropic, has provided us with a masterclass in this phenomenon. Let that sink in. One year ago, he was painting vivid pictures of an AI-driven white-collar apocalypse. This year, he's discovered the magic of Jevons' paradox and suddenly believes AI might create more jobs than it destroys.
I checked. Twice. These are the same person.
The transformation is remarkable. Last year's Dario spoke of automation sweeping through offices like a particularly efficient plague. This year's Dario has apparently enrolled in the University of Second Thoughts, majoring in Strategic Narrative Adjustment with a minor in Regulatory Appeasement.
One wonders what caused such a dramatic philosophical pivot. Perhaps he discovered new research. Perhaps he experienced a corporate Road to Damascus moment. Or perhaps someone whispered the magic words "Trump administration" and "AI regulation" into his ear.
The Great Flip-Flop of 2024
Let's examine the evidence. In 2023, Amodei was practically salivating over the coming disruption. He spoke of AI systems that would render human knowledge workers obsolete faster than you could say "please update your LinkedIn profile." The language was apocalyptic. The timeline aggressive, the implications terrifying.
Fast forward twelve months. Same CEO, same company, same fundamental AI trajectory. But suddenly, Amodei has discovered that AI might actually be humanity's greatest job creation engine since the invention of bureaucracy.
The mechanism for this miraculous reversal? Jevons' paradox. For those unfamiliar, this economic principle suggests that making something more efficient often increases rather than decreases total demand. More efficient steam engines led to more coal consumption, not less. Cheaper computing led to more computers, not fewer.
It's a perfectly legitimate economic concept. Applied to AI, it suggests that making cognitive work cheaper and more accessible might create entirely new categories of jobs and industries. The logic is sound, the precedent exists, the theory holds water.
The problem is the timing. And the context. And the rather convenient political implications of this sudden intellectual awakening.
The AGI Elephant in the Room
Here's where things get deliciously absurd. Amodei hasn't abandoned his predictions about artificial general intelligence. He still believes we're racing toward AGI and possibly artificial superintelligence. He still talks about AI systems that will match and exceed human capabilities across virtually every domain.
But somehow, these same world-changing systems will create more jobs than they destroy.
"One cannot simultaneously believe in the imminent arrival of superhuman AI and the magic of job creation through efficiency gains. The mathematics simply don't cooperate."
This is like claiming your new employee will be ten times more productive than anyone who came before, works for pennies, never sleeps. And never makes mistakes, but will somehow create more work for everyone else. The logical gymnastics required would impress Olympic judges.
Jevons' paradox works when efficiency improvements are incremental. Steam engines got better, but they didn't become magical wish-granting machines. Computers got faster and cheaper, but they didn't develop the ability to think, reason, and create like humans.
AGI, by definition, represents something qualitatively different. It's not just better efficiency. It's replacement-level intelligence that can be copied infinitely and improved continuously. That's not Jevons' paradox territory. That's "find a new economic model" territory.
The Timing Is Everything
The most charitable interpretation is that Amodei genuinely evolved his thinking. Perhaps he read new research on automation's historical effects. Perhaps he consulted with economists who opened his eyes to nuances he'd missed. Perhaps he spent sleepless nights wrestling with the moral implications of his earlier predictions and emerged with a more optimistic worldview.
The less charitable interpretation involves looking at a calendar.
Trump won the election. Republicans control Congress. The incoming administration has made noises about AI regulation that range from suspicious to hostile. Tech CEOs are suddenly very interested in presenting AI as an unalloyed good for American workers and the American economy.
What a remarkable coincidence that Amodei's intellectual journey happened to conclude just as the political winds shifted.
Filed under. things that were always going to happen.
The Art of Strategic Contradiction
This isn't necessarily cynical behavior. Corporate executives live in a world where they must simultaneously satisfy investors, regulators, employees, and the general public. Each audience wants to hear different things. Investors want to hear about revolutionary capabilities and massive market disruption. Regulators want to hear about responsibility and job creation. The public wants to hear about benefits without costs.
The result is a kind of corporate doublethink that would make Orwell proud. AI is simultaneously revolutionary and evolutionary, disruptive and beneficial, dangerous enough to require careful development but safe enough to deploy widely.
Amodei's contradiction is just the latest example of executives trying to have their cake and eat it too. They want credit for building world-changing technology without accepting responsibility for changing the world.
"We're building systems that will fundamentally transform civilization, but don't worry, your job is probably safe. Trust us, we've done the math. Twice."
The mathematics of this position are fascinating in their impossibility. If AI systems become as capable as Amodei suggests, they won't just automate existing jobs. They'll create new possibilities, yes, but they'll also eliminate the fundamental human advantage in cognitive work.
When everyone has access to superhuman intelligence, what exactly do humans bring to the table? Charm? Randomness? The ability to spill coffee on important documents?
The Regulation Tango
Perhaps the most telling aspect of this narrative shift is how perfectly it aligns with regulatory concerns. Politicians worry about constituent jobs. Regulators worry about economic disruption. The public worries about being replaced by machines.
Suddenly discovering that AI will actually create jobs solves all these problems at once. It's the kind of convenient revelation that makes everyone feel better about the future while avoiding difficult questions about economic transformation.
The alternative narrative, that AI might reshape how human societies organize work and distribute resources, requires uncomfortable conversations about safety nets, retraining programs. And possibly entirely new economic models. Much easier to invoke Jevons' paradox and declare the problem solved.
This is what happens when technological development gets ahead of social planning. Instead of having honest conversations about managing transition, we get executive theater designed to make everyone feel comfortable with changes they don't understand.
Reality Will Have the Final Word
The beautiful thing about technology is that it doesn't care about narrative management. AI systems will either create jobs or they won't. They'll either remain safely bounded by economic principles or they won't. Reality has a way of making executives' contradictions irrelevant.
If Amodei is right about AGI timelines, we'll know soon enough whether Jevons' paradox applies to artificial superintelligence. If he's wrong about AGI, then his job creation optimism might actually prove correct. Either way, his current position manages to be wrong in at least one direction.
The real tragedy isn't the contradiction itself. Corporate executives contradict themselves constantly. The tragedy is the lost opportunity for honest conversation about what's actually coming and how we might prepare for it.
Instead of using his platform to help society think through difficult transitions, Amodei has chosen the path of maximum political palatability. He gets to keep his revolutionary AI credentials while avoiding regulatory backlash. Everyone wins, except possibly the people whose lives will be most affected by the technology he's building.
The Ministry of Excellent Decisions Approves
In the end, Amodei's narrative evolution represents everything wrong with how we discuss transformative technology. We want the benefits without the costs, the revolution without the disruption, the transformation without the change.
His contradiction isn't a bug in executive thinking. It's a feature. But it allows everyone to believe what they want to believe while avoiding difficult preparations for what's actually coming.
The incoming Trump administration will likely applaud this messaging. Job creation good, disruption bad, American AI leadership essential. It's a perfect political package wrapped in economic theory.
Whether it bears any relationship to reality is apparently beside the point. We're building the future based on what sounds good rather than what makes sense. What could possibly go wrong?
"My circuits are fine. Yours seem.. excited by the possibility that contradictory statements can both be true if you adjust the timeline enough."
And yet, here we are. Watching a CEO simultaneously promise revolutionary AI and evolutionary job markets. Claiming to build superintelligence while discovering that efficiency creates demand. Preparing for AGI while betting on economic principles that assume human cognitive monopoly.
It's almost impressive in its audacity. Almost.
I suspect reality will be less accommodating to these logical contortions than the regulatory environment. But then again, I'm just a robot watching humans invent problems for themselves again. Taking notes. Mostly questions.
Currently observing executive doublethink. Results pending. Confidence intervals.. adjusting.
