Mike Allen and Jim VandeHei, reporting for Axios (Apple News+):

OpenAI CEO Sam Altman is doing something no tech titan has ever done: He’s publishing a detailed blueprint for how government should tax, regulate, and redistribute the wealth from the very technology he’s racing to build and spread.

I avoid linking to Axios on principle, and I don’t find their commentary particularly enlightening here, so I’ll just list all of OpenAI’s proposals, listed in a lengthy document published on the company’s website:

  1. A “public wealth fund” Americans can invest in.
  2. Shifting the tax structure away from payroll tax and toward capital gains, with the implication being there’ll be no payroll to tax.
  3. A four-day workweek.
  4. A legally protected “right to artificial intelligence” (Axios’s words, not mine).
  5. Government regulation to prevent rogue AI.
  6. Better “automatic stabilizers,” like unemployment benefits and cash assistance.

Altman has two primary concerns at the moment, both of which must be addressed imminently:

  • His company is hemorrhaging money, losing enterprise contracts to Anthropic, while aiming for an initial public offering later this year. Inference costs too much, and the company has been throwing away funds for the past few years.
  • His company’s technology, generative artificial intelligence, is polled less favorably than any other modern institution in America. It is wholly unpopular, gives business executives reason to eliminate people’s jobs, and is upending our economy with no clear benefit.

Neither of these is a good problem to have. Monday’s proposals focus on the second by hoping to increase his and his company’s favorability among the public. OpenAI, for the past year or so, has moved away from the nihilism and safety-oriented culture of OpenAI and toward wrong-headed optimism, and this is another example of the company’s misdirection. Even OpenAI’s own ex-leaders saw the flaws in this marketing ploy, and subsequently left to found their own start-ups.

The “wealth fund” and four-day workweek are laughable proposals clearly written to get in newspaper headlines. They’re unserious, OpenAI knows they’ll never happen, but perhaps they’ll attract the attention of gullible individuals. This is pure public relations theater, and neither our Congress nor our institutions are equipped for any of these proposals. They’re frankly unserious.

Proposals 2 and 6, however, are indicative of OpenAI’s true hedging. The company internally believes there will be little to no white-collar jobs left thanks to its technology — which does not always make workers more productive and cannot do their jobs for them — and wants the economy to prepare for that. Its reason for this is quite obvious: if all of the white-collar workers, who make up the vast majority of OpenAI’s subscriber base, don’t have income, they won’t buy OpenAI products. And that perfectly ties into Proposal 4, the “right to AI.” Because people won’t have money to spend on AI products, OpenAI hopes the government will give them free or subsidized access to them. The government will make money by changing the tax structure to be less reliant on payroll taxes. This maintains the cyclicality of the economy while completely cutting out the white-collar worker.

These proposals are entirely self-serving, cloaked in the veil of altruism. They guard OpenAI and its peers from the devastating effects of job loss and leave the problematic bits to white-collar workers and their government. And it’s not like OpenAI’s technology is particularly competent, either. I still do not believe AI itself, by its merits, has caused job loss — business executives, and more specifically, middle managers, have instead realized that knowledge workers cost more money than a subscription to ChatGPT Pro or Claude Max. These tools are not even halfway as competent as a skilled software engineer — and neither do they have as much potential to become as skilled as a junior software engineer does — but middle managers and business executives have never cared about quality. AI is cheaper, not better.

All of this is considering that these proposals will genuinely become rooted in law, which is a laughable fantasy. Congress simply does not legislate — it can barely pass appropriations on time, its most fundamental constitutional duty. What’s far more likely is that incremental model updates exponentially increase the confidence of business executives, leading to even more job loss in even more white-collar sectors. Doctors, lawyers, journalists, teachers, etc., will all be either underpaid or out of their jobs, not because the models have gotten so much better in these few years, but because their managers have decided that payroll is too expensive. Coupled with the war in the Middle East, this will put enormous strain on economic stabilizers like welfare, causing the national debt to skyrocket, increasing borrowing costs, decreasing capital investment (because people are out of jobs!), and leading to a painful recession. This is all elementary economics.

This is why we have a policy in America of keeping the public and private sectors as far away as possible. But unfortunately, Washington politicians are too comfortable in their failure to legislate — because they’re lobbied by the private sector — to do anything about this impending crisis. You don’t need a computer science or economics degree to foresee what’s happening in our economy.