Tuesday, November 21, 2023

Does OpenAI Prove that East is East and West is West?

Rudyard Kipling quote: Oh East is East and West is West and never...

Take a look at OpenAI's announcement of Altman's sacking.  Read it closely.  The whole crisis is about tax exempt status and whether profit making and doing good can co-exist, don't you see.  And it's not by coincidence that Microsoft -- a major profit-making partner in the OpenAI LP -- is mentioned prominently in a saga about the governance of OpenAI, Inc., the exempt organization that is supposed to be in charge of it all.  

The governance crisis at OpenAI is forcing me to reconsider my estrangement from the nonprofit doctrine of my earlier years.  Then, I wholeheartedly agreed with the "never the twain shall meet" mantra.  It’s an initial article of nonprofit catechism that charity and profiteering don't mix; that private inurement and private benefit are invariably, inevitably inconsistent with charity.  Over the years, even the Service has softened because charities exist in a ruthless capitalistic market.  Charitalists and capitalists are just gonna have to get along together.  I still thought the Service was being miserly in its nearly impossible barriers to non-profit joint ventures with for-profits.  And I have written that explicit profit sharing with capitalists in a whole [hospital or artificial intelligence] joint venture should not be so presumptively improper as to be nearly impossible to achieve.  I still think that’s true.  But doggonit, OpenAI is making the assertion harder and harder to defend.  Maybe it’s true after all that good works and profit-making can never really coexist on a tax-exempt platform.  I thought I was enlightened.

What OpenAI is doing regarding OpenAI’s governance and capped profit merger is about as revolutionary to tax law as its inventions are to artificial intelligence and the real world.  I initially thought that the “capped profit” joint venture was a bridge too far for tax exemption.  Because it gave private investors explicit rights in the charity’s profits.  It’s a literal violation of 501(c)(3), whether we call it private inurement, private benefit, or excess benefit.  It’s basic law that tax exempt profits must be plowed back into the charitable purpose and not diverted to private consumption.  On the other hand, exempt orgs are allowed and most of the time required to pay market rates for inputs – supplies, labor and even capital.  If the market, in an arms-length transaction to achieve OpenAI’s charitable mission, demands the sort of joint venture represented by OpenAI LP, then so be it.

The market safeguards against elevating profit over charity break down, though, when the charitable folks striking the deal are on both sides of the deal. Initially, the charitalists were on one side as the fiduciaries of OpenAI and the capitalists (Microsoft) were on the other.  Charitalists seek charity, profit takers seek profit.  If both sides bargained in good faith, we could assume that the joint venture was what the market dictated to achieve the charitable purpose.  Tax exemption remains appropriate.

But soon after the initial deal, the charitalists migrated to the capitalist side.  They took stock in the for-profit partner, Microsoft, by which they might share in the profits being shared with Microsoft via the joint venture.  The details of the charitalists’ personal investment in the for-profit partner are murky but the fundamental economics are clear.  The charitalists are on the verge of indirect private inurement and excess benefit.  Except the Service reserved that section of the excess benefit regulations that govern the insider profit sharing that the charitalists are now enjoying.  In any event, the OpenAI board members (the erstwhile charitalists) are now on both sides of the transaction defining and operating the joint venture; we can’t just assume the terms or operation will make charity of singular importance.  See, this I why charity can’t ever have anything nice.

That Altman is now working for Microsoft raises a whole bunch of other issues.  If he gets a profit-sharing deal – he’d be a fool not to – will his profit-sharing result in private inurement and excess benefit?  He’s probably gonna be a disqualified person for 5 years.  But even before Altman’s firing, board members and employees (all the techy guys who want to save the world – the alleged “charitalists”), had announced plans to cash in like plain old capitalists.  The employees took stock in OpenAI LP (probably through Microsoft, but I don’t have the details) and announced that they were going to offer that stock to the public.  WSJ speculated that the stock sale would bring in $90 billion.  Considering their potentially astronomical profit from the stock sale, the OpenAI charitalists now have as much interest in elevating profit over charity as any of the other profit-seeking partner. 

There are all sorts of speculation regarding the power struggle.  Some say that Altman’s ouster is a result of the conflict between his profit-making ideas and the board’s attempt to keep the charitable mission paramount.  Reportedly, Altman wanted to go to market with new stuff faster; the board thought doing so would sacrifice the charitable mission of releasing new technology only when safe and equitable to do so.  Just to state the conflict a different way, delaying the release of technology could cost somebody immense short-term profit. And since Altman and the rest of the OpenAI board members hold stock in OpenAI LP, they too lose money by delay.  Because delay could reduce the sale price of the stock they are holding right along with the other capitalists involved in the joint venture.  The board’s firing of Altman is thus sort of like the pot calling the kettle black, even if the Board has not yet gone all the way over to the capitalist dark side.

I initially thought the capped profit entity violated 501(c)(3).  But when I came to understand AI’s  incredible potential for both good and bad, I understood that original OpenAI had a legitimate charitable purpose for which it deserved tax exemption.  AI is like a nuclear bomb that we do not want monopolized neither by despots nor capitalists.  The only way to achieve that is through huge influx of public capital (directly or via the tax code).  So I reasoned that the capped profit deal was justifiable, but only if the details were negotiated with charitalists on one side and capitalists on the other. The charitalists have plenty of bargaining strength because they own the technology and have access to tax exemption and tax deductions.  The capitalists have the necessary capital.  But when the charitalists morph into capitalists the protections of equal loyal bargaining power evaporate.  Tax exemption is no longer appropriate.


darryll k. jones


| Permalink


Post a comment