OpenAI pursues public benefit structure to fend off hostile takeovers | Hacker News Reader

OpenAI pursues public benefit structure to fend off hostile takeovers

ft.com

134 pointsJumpCrisscross68 comments

Poison pills are much more effective at fending off takeovers. I think the public benefit form mainly helps with workforce and customer alignment to dampen disruptions from transformative and invasive applications.

Peter Gassner converted Veeva (pharmaceuticals software) to public benefit a few years ago. Veeva is probably 10X smaller and mostly captured by big pharma customers who don't want to compete on software infrastructure and want one place to stay in (FDA) compliance. So in that case the stakeholders have a pretty clear and distinct voice and mission.

In the absence of dominant stakeholders, it's unclear who would govern OpenAI as a public benefit corporation. At a minimum, it's a moral hazard and an invitation to bureaucracy and politics.

I'd be interested in OpenAI developing super clear and transparent customer feedback channels. That could improve AI, give incentive for people to participate in model improvement (instead of reflexively protecting their data), and be a forcing function against governance gaming. Amazon/Bezos has strong experience in managing by metrics, so it's a shame OpenAI has committed itself to Azure/Microsoft.

Another alternative is the M-form, where subsidiaries are coordinated but live or die based on their own performance rather than cross-subsidies (recently in Google becoming Alphabet). But the history there isn't particularly good for innovation, and even profit lags except in industries with stable markets (Berkshire-style).

I agree with most of what you’ve said, but here:

> Amazon/Bezos has strong experience in managing by metrics, so it's a shame OpenAI has committed itself to Azure/Microsoft.

If they are a customer/investor, as opposed to an internal leadership team, it seems preferable to have someone with comparable resources who is more permissive/hands-off.

Not necessarily to fend off hostile takeovers, Anthropic adopted that structure too and it might be that they're just admitting that it's the right option

> The PBC model is relatively new. Delaware, where most companies are incorporated, only adopted PBC legislation in 2013 and then changed provisions in 2020 to make the structure more enticing. Of the thousands of publicly traded companies in the US, fewer than 20 are PBCs. [...]

> However, it has proved popular with AI companies. Both Anthropic, the maker of the Claude AI tool, and Musk’s xAI are PBCs. One person close to xAI said this meant its probability of being sued was reduced if it did not act in accordance with the shareholders’ interests.

[...]

> “It is intentionally a way for incumbent management and directors to entrench themselves,” he said. “If you can convey the idea to people that you are a good enterprise, a morally safe enterprise and that comes with very little constraints, that has to be tempting to entrepreneurs.”

That really goes against the original spirit of the PBC, which was the exact opposite: you should be able to fend off (not get sued by) evil profit-seeking shareholders, while pursuing the long-term vision.

But I guess these things goes both ways, and the path to hell is paved with good intentions.

We just can't have any nice things.

As the article says, the PBC structure has never been tested in court, therefore this is all speculation at this point, so it seems a bit premature to despair.

Yes, a big benefit of a PBC is that you're not only beholden to increasing shareholder value - that's obviously inherent to the PBC model, and I think it could be used for both "good and bad". At the same time, PBCs need to say what their "public benefit" is, and (my understanding is that) they have fiduciary responsibility to others besides just shareholders. This, I imagine they could also be sued by people saying they aren't fulfilling their public benefit mission. Point being, PBCs have pros and cons, and it will take some court cases to find out where the line truly lies.

I can’t imagine thinking agency risk would be reduced by the agent(s) declaring good intentions.

Public Benefit Corporations (especially with the B-Corp certificate) is just a marketing strategy for gullible people uncomfortable with capitalism

But like with all entities, there is nothing inherent about them that would make the individuals within them act differently than a simple C-Corporation.

It all comes down to the bylaws.

What does it even mean to fend off hostile takeovers when Microsoft owns 49% of the company and the remaining 51% is split between a small number of investors (Musk, Thiel, Infosys, AWS, and a half dozen VC funds). All it would take would be one of those other early investors to join up with Microsoft and they would have a majority ownership stake.

Did infosys actually get equity when they donated to it as a nonprofit?

I had thought that was most of the reason Elon is lashing out, because he donated to the non profit and received no equity.

> A key benefit of this PBC structure is its potential to thwart an unwanted acquisition or an activist’s demands, according to multiple people familiar with the company’s thinking. This means an existing investor such as Microsoft or another party could be frustrated if they mounted an effort to acquire OpenAI.

An astonishing justification proffered for OpenAI's attempt to remove itself from being controlled by a non-profit entity. A PBC might be better than a regular c-corp, but it is not better than a non-profit. OpenAI is pursuing this arrangement in order to grant Sam Altman more control and enable fundraising; the PBC thing is a way to fob off those concerned by exactly the wrong things (i.e. that Sam Altman might be incorrectly removed from power by external stakeholders, rather than, uh, being correctly removed from power by internal stakeholders).

> ChatGPT maker will retain a not-for-profit entity, which would be independent and have a stake in the PBC. This not-for-profit would have access to research and technology but solely focus on pursuing OpenAI’s mission of benefiting humanity.

And up to 60% tax deductions for the for-profit entity and executives that carry forward for 5 years when unused and exceeds current year Modified Adjusted Gross Income

Can you explain how that works? Why would either the for-profit entity or the executives get a tax deduction because one of the stakeholders is a nonprofit?

It's not automatic, they'll donate something of value to it.

Here is the default reality:

A) they can do this to any "60%" non-profit. 60% refers to the category of maximum tax deduction

B) they won't be in control of the assets after they donate

C) but they'll be loosing money they otherwise would have that exceeds the tax burden

But all of these are surmountable by already having your own non-profit with the same board members, or aligned board members:

A) OpenAI is a 60% non-profit that they control and so

B) they'll still be in control of the assets they donate as if they never moved it

C) you can donate illiquid things that would have never been able to be converted to dollars. for example, those PPUs? or perhaps just some cloud compute credits? shares, or membership interests, of an organization as long as a market value can be pointed to. A bunch of GPUs?

An illustrative example would be how the new for-profit OpenAI entity sold some shares for $6bn to represent a 157bn valuation, this means that 3% of the shares were exchanged for dollars. And all 97% of the shares are said to be worth the exact same instantly, and indefinitely into the future. You could donate 1% of the shares to the non-profit OpenAI and that's a $1.57bn tax deduction against whatever income you currently have that year. and if you don't have enough income to offset then it keeps rolling forward.

in a 60% non-profit, assets can only offset up to 30% of your tax bill, and cash donations can be used to offset an additional 30%. Alternatively, cash alone can offset up to 60%. Since donating cash is suboptimal, do 30% illiquid, appreciated, assets, and 30% cash. The cash can also be found by borrowed funds but this is not seen as optimal, it can be seen as strategic though.

Once again, all the donated assets are in a non-profit you still control, while obtaining the tax benefits.

How is this useful? many ways. Non profit has financial ammunition and firepower. It can pump investments you also own by purchasing, or getting involved with. The regulations curbing this are quite flexible. Its a pretty high percentage ownership threshold between you and the nonprofit for it to violate self-dealing regulations, and even when violated you have like 1 - 3 years to get under those thresholds. But even that's just for shareholdings. It can print revenue for things you like, buy more GPUs, burn more energy in compute from your organization, make external investors enamored.

It will also be doing its stated mission, research. Honestly, the stated mission is exactly what I would do if I was also interested in doing all of the above.

The longest preparation is creating the non-profit and getting that approved. They already have that. so the rest is just pressing play.

This is great, and I'd like to see more of it. Honestly we should be asking why we tolerate any type of corporation except one that's public benefit. Why are we as a society giving owners of capital any special, privileged shield from liability if they aren't using it for the betterment of society?

And it wasn't ever thus, either. Dodge v. Ford Motor Co. really screwed things up.

After reading your comment, I did a little searching and found a great essay on the shareholder primacy standard in Delaware: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2916960

Contrary to what a lot of other commenters are saying, the shareholder primacy rule is relatively recent (basically since the 80s), and before that there was broad recognition that corporate boards could balance the interests of various stakeholders. This essay also argues that Dodge v Ford Motor (which was in the Michigan Supreme Court) actually had little practical impact - it was cases like Unocal and Revlon in the Delaware courts that had a much bigger impact on actual board behavior in the US.

It's better compared to the baseline of a for-profit corporation, especially given the stakes -- but it's strictly worse compared to the intended purpose of the existing structure that was recently hollowed out. The nonprofit entity that was originally supposed to be in control actually won't be.

I'd also question whether there is actually going to be a meaningful check on any of OpenAI's actions, compared to what there would be for a for-profit corporation. I'd bet no.

I agree that in this case, PBC is a downgrade in terms of strictness of mission. In general though, PBC is better than standard corporate goals of strictly maximizing shareholder value, employees/customers/the public be damned.

The nonprofit will hold stock in the BPC. I think the effect of this is that the nonprofit board, which is independent of Altman, will be restricted from selling their shares. However the nonprofit will eventually convert from a public charity to a private foundation and will be required to give away a percentage of their assets every year. I'm not sure how that will work with a BPC.

Who gets to decide? Thats the challenge.

The amazing thing is that the pursuit of personal enrichment, when regulated in moderation, tends to create a system that is in the public’s benefit. So we allow companies that are self interested because that, at scale, actually supports the public benefit more than requiring all companies to do specific things mandated as for the public benefit.

> tends

Tends is doing a lot of heavy lifting here.

The model of "Let's have the worst people do whatever they want, for the most selfish possible reasons, and hope for the best" has also produced a horrific amount of suffering. Crazy amounts of socialized externalized losses get hand-waved away.

And yet all the most prosperous countries in the world are the ones that allow individuals to seek unbounded profits, and all the darkest holes of history are in situations that attempt to tell people how to live

Every single society has told people how to live since the Code of Hammurabi. The darkest times in history have happened when people with power dehumanized large groups of people, and that's not particular to any economic philosophy.

Waking up on the wrong side of a hundred monopoly rents enforced by the property rights of rich people doesn't feel like freedom to me.

Yes, fully socialist countries seem to be doing even worse, but so many of the good parts of my employment contract came from the New Deal and its wild success shows in everything from the infrastructure to the demographic charts that are still defined by echoes of the baby boom! It's no wonder the businessmen got together and tried to coup FDR and have spent the last 50 years trying to wind back these policies. We don't need a Glorious Revolution but we do need another Roosevelt.

Why the most prosperous country in the world has homeless, opioid epidemic and is pretty consistent in having the same kind of problems? Can you really call it the most prosperous then?

Yes I can. Because none of the others would allow my family the standard of life we have here.

The US does a great job at hyping itself, but let's be realistic: it's 20th in human development index[1]. "Sure," you might say, "regular people don't do as well here, but you can get ahead here if you work hard! That's the American Dream™."

Nope, it's also 27th in social mobility[2].

[1] https://en.wikipedia.org/wiki/List_of_countries_by_Human_Dev...

[2] https://en.wikipedia.org/wiki/Global_Social_Mobility_Index

I was part of a student exchange with Sweden. This is just an anecdotal experience but if that is what 4 on the mobility list looks like then they got something horribly wrong. I enjoyed many aspects of the country, but prosperous and highly mobile is not what I observed in any way.

> all the darkest holes of history are in situations that attempt to tell people how to live

I sure hear a lot of faux-market fascists telling me how to live these days.

> The model of "Let have the worst people do whatever they want, for the most selfish possible reasons, and hope for the best" has also produced a horrific amount of suffering.

This is a profoundly unserious framing. Suffering is the default state of man, and our current system has done far more to eradicate said suffering than any other.

The historical alternative system was voluntarily abandoned by all of its impoverished practitioners in favor of the current system.

The current "alternative" to our system is just our system with high taxes.

Just take the L.

> Suffering is the default state of man

What an awful and depressing view to hold. I hope one day you find a small piece of happiness in your life.

Things used to be worse, and our default state is “suffering” so we should give up, stop here and lick the boots of the current, obviously flawed system? That doesn’t even make sense as an argument.

> Suffering is the default state of man, and our current system has done far more to eradicate said suffering than any other.

This seems like gaslighting. Of all the takes, "You were meant to suffer, so if I made you suffer and then I did stuff to reduce the suffering you should be grateful" is certainly one of them.

Chopping up the future and selling it to rich people is a brilliant way to obtain decentralized skin-in-the-game capital allocation decisions. But then the future arrives and it belongs to rich people whose only goal is to squeeeeeezzzzzeee.

Capitalism is good at avoiding boondoggles but it's a sucker for asset pumps. We don't have to pay heavy taxes to support a gigantic network of high-speed trains to nowhere but we do have to pay monopoly rents on land, two-sided markets, Metcalfe-law networks, and so on.

First of all, it is extremely difficult to say whether certain actions are good for society. I think there are some things that you can say are clearly good or bad for society, but what about all the stuff in the middle?

Would the Coca-Cola company be allowed to exist in this world? Are professional sports team good for society?

Sure, it's hard to say exactly what's good for society, it's a broad aim. However, being a PBC also insulates you from extremely short-sighted shareholders. It's different than being a charity.

Coca-Cola could easily be a PBC, so could pro sports teams. It just gives you the freedom to shut down short-sighted activist shareholders who can't see past next quarter. It doesn't require you to be Mother Teresa, it just doesn't let anyone punish you if you decide that's what you want to be.

Requiring benefit is a huge step from where we are. Perhaps it would be easier periodically determine the n most harmful and dissolve them. Predation is known to make an ecosystem healthier.

This would sidestep the problem of:

> but what about all the stuff in the middle

> Requiring benefit is a huge step from where we are.

The phrase "requiring benefit" is doing a huge amount of heavy lifting in this statement by assuming benefit is something that can easily be agreed upon, or that it's not already taken into account.

Since it will obviously be difficult to agree on this, what if we just vote? I'd propose a system where each of us gets a certain amount of tokens, and then we use our tokens to vote on which products are beneficial. Those companies that make beneficial products will receive the most tokens. They can distribute some of them to their employees and shareholders. The companies that don't provide enough benefit will naturally go under as they fail to accumulate enough tokens to attract employees. We can also set up systems to ensure that everyone receives a minimum number of tokens, and also so that certain people don't accumulate so many tokens as to give them too much control over society.

I believe this system will work well for rewarding companies which produce benefits and punishing those who don't and look forward to it being implemented.

Something along those lines is what I would propose, although once the tokens were used to decide which companies get dissolved (and maybe which ones get tax breaks) I'd zero-out all balances rather than creating a secondary economy around their circulation. By that point they'd have done their job. But I'm not really enough of a game theorist/economist/evolutionary biologist to be coming up with the details. My point is just that we should try something like this.

What do you think the top 10 worst companies are? How likely do you think it is that your list matches mine or anyone else’s?

Exact match? Unlikely. But suppose that with each tax return, everybody had the option of listing 5 corporations which they believed were doing more harm than good. It would be simple statistics to determine if the signal was strong enough to warrant action. As for mine, well I'd have to do a bit of research but they'd likely come be in defense, oil, or adtech.

Perhaps we also could list 5 which we think are having a positive impact. The assets of the bad ones could be auctioned off to the good ones, the money made from that auction could go towards severance for anyone whose job was dissolved.

As I recall, Kickstarter was a very early Public Benefit Corporation. If nothing else, it probably gives a card to activist employees who think the corporation is hurting the public in some particular way.

Because Coca Cola is more than gaseous sugar water.

As far as I understand Public Benefit is as meaningless as non profit. The names are extremely misleading. A public benefit corporation doesn’t necessarily have to provide benefits for the public and non profits can be extremely profitable for the right people.

Because the privileged shield from liability encourages economic growth and raises the standard of living for everyone. Without that shield, potential investment losses would be infinite and investors would hoard their capital instead of using it to fund new ventures. Our entire economy would be smaller, slower, and less innovative.

Yes occasionally corporate owners and managers do bad things, and escape serious financial consequences due to the liability shield. So what. That is an acceptable consequence considering all of the other benefits.

Unlimited corporations (anonymous and personhood but not limited liability) could also be tolerated.

My life has been more improved by people acting for their own profit and being open about it than people ostensibly acting for the “public benefit”

I can't read the article, nor am I deeply familiar with US financial structures, but are OpenAI doing the same as Bluesky, a Public Benefit LLC? Another HN comment mentions OpenAI adopting a "PBC model" but I'm not sure if that's the same or not. If not, what are the material differences?

It’s a Public Benefit Corporation (sub-type of C-corp) not an LLC.

>> obliged to balance the best interests of shareholders, a public benefit, and stakeholders, such as employees and society

This seems like a very murky structure. Probably fine in good times but how would it operate in practice when things are going badly?

> Chief executive Sam Altman will also receive equity for the first time in the for-profit company, which could be worth $150 billion after the restructuring

A net worth of $150 billion is quite an achievement.

> A net worth of $150 billion is quite an achievement.

And not too long ago he was touting he was in it for the good of humanity, and "proved" it by only taking a $60k salary with no equity.

A true paragon of honesty and straight-dealing.

Indeed. Plus it's rather late in the company lifespan for them to be handing out large swathes of equity. A weird decision for a company nominally controlled by a non-profit.

It's saying the company itself is worth $150B, not Sam Altman's net worth

$150B is just a (somewhat arbitrary) estimate of the current value of OpenAI.

Poorly phrased. $150 billion is the total company valuation, not the share that Sam would receive.

If anybody is as confused as I was, the quote in this comment is from a separate article from weeks back...

https://www.reuters.com/technology/artificial-intelligence/o...

Don't play their game where they pretend that their net worth is proportional to their achievements.

Is the article linked on archive different? This text is not in the article I read.

Reads like a play to dilute the power of executives that departed

Can't wait to see the contortions Altman pulls to retain a Zuckerberg-tier grip on the company, they'll put Cirque do Soleil to shame.

It's not that hard we just have the original not-for profit and then a non-profit llc, and then 503c, all of that gets grouped under a Hybrid profit holding, which funds the PBS but has no ownership relationship. Meanwhile a sister L3C invests in the original NFP without profit motives, but retains call stock options WITH voting rights (only if the option is NOT executed). The for profit LLC is just a holder of the stock and IP in Ireland, nothing suspicious about it, do not concern yourself with this entity.

Source: I made it up

There's also the subtitle on the article:

> ChatGPT maker considers largely untested company model to protect chief executive Sam Altman from outside interference

Or maybe the motivation is some other technical maneuver.

Or maybe the motivation is PR/optics.

Is there anyone who trusts OpenAI to be honest about their intentions, at this point?

It's actually incredible how much headlines there are regarding OpenAI--how much cap they're raising, the ins and outs of their corporate structure, how much that psychopath will be compensated--yet very little on what they've actually shipped.

Eh, or to avoid public backlash for longer and minimize the odds of legislation targeting them.

Here's a Reuters article about it for those that can't access the Financial Times article: https://www.reuters.com/technology/artificial-intelligence/o...

Is it hostile if the call is coming in from the house?

Reply on news.ycombinator.com