Hacker News new | past | comments | ask | show | jobs | submit login
OpenAI removes Sam Altman's ownership of its Startup Fund (reuters.com)
340 points by mfiguiere 7 months ago | hide | past | favorite | 45 comments



I'm not a lawyer, but I've been involved in a fair bit of venture fund formation.

A fund is not an incorporated entity, but rather a partnership of two persons. In a limited partnership, one of those persons is the general partner and does all the work and assumes the related liability, and the rest are "limited partners" who just commit capital.

Banks and tax authorities only open accounts for actually-formed partnerships, not hypothetical partnerships that might happen in the future. But, you need a bank account for the LP investors to send their money to.

So, in my experience it is super common to have a temporary Initial LP, which is usually a related person (one of the founders of the firm acting in their personal capacity, for example), to allow the Initial LP and the GP to form the fund and open the accounts. Then the Initial LP withdraws in the actual closing event where the investors sign onto the partnership agreement and commit their money.

I have never seen an "initial GP" though. I believe the explanation, but can't really figure out why one would be necessary. Maybe they wanted to incorporate a GP Co, but didn't get around to it?

Love to hear any theories why this might be useful..?


None of this makes sense in a Master-Feeder model, as the master fund is a partnership between the feeder funds, one US, one offshore for foreign investors

The general partner is not an owner they are a manager. The general partner doesnt have to be a human either it can be another entity or multiple entities

all of this can be hotswapped

Even in a mini-master models or single US entity model, LPs are the owners, GP is just a managing member. “ownership” in the context of this article is ambiguous, but the SEC filings would add the clarity necessary for this discussion, probably “manager” or “controlling person”


I suppose you caveated your statement well , by starting your writeup with "master-feeder" setups. However in a vanilla fund, your description does not apply.

In a vanilla fund, the GP is another partner and has "ownership" just like LPs do. Its just traditionally very insignificant.

The GP is simply a special partner as outlined in the LPA. They vote etc just like other LPs in major matters. They make minor decisions for the partnership. This is why usually that means the GP is the managing member.

Ive described the traditional VC fund.

No need to make it more complex with a master-feeder structure. Or a secondary, etc


The gp gets carry


Where can you go to learn more about these entity / LP / LLP setups that finance orgs use?


I went to law school to do that, but without that you can start by focusing on an area of law called "business associations/corporations." There are resources online to assist law students studying for exams in that class so you can use those to get a grounding in the concepts of agency law and partnerships.

From there you can start to specialize in more specific structures by examining the filings of public companies or scouring the web to figure out how a particular fund you are interested in is structured.

To let you know, it is further complicated in US law by each state having different corporate rules. If you go with Delaware or New York it's probably your best bet for public, but closely held or private companies could be anywhere. As a starting place, most states adopt the Revised Uniform Partnership Act, or RUPA - that is available online through law.cornell.edu


be rich enough to be a limited partner, means you're rich enough to have your own specialized lawyer to review the contract (instead of "I know a guy" "my cousin" "prepaid legal" "general counsel"), rich enough to afford an MBA, rich enough to do a dry run in setting up your own hedge/vc/pe fund

otherwise, I don't know exactly. but I've personally been reading wiki pages on fund structures since I was 18 and not eligible for anything, just dissecting the quiet class systems I noticed. scouring SEC documents, and validating my assumptions with hedge fund lawyers and being an LP when I could afford it later


Rewriting this to just say it could be simply to shift liability for your initial GP to the new one; not sure why you would choose this structure in that case but it's possible?

It could also be that the initial GP had individual contractual or ethical obligations that were counter to the best interest of other investors so could not continue acting in control of the investment vehicle, that'd be a reason to change

But no, no idea why you would plan to have an initial GP


You might be simplifying, but just to clarify an LP can have many general partners and many limited partners.



There's... a lot of other stuff that turns up on Google, including GPT chatbots associated with "Vespers Inc" that claim to help people with online filing, name overlaps with people who had genuine court cases filed against them in California several years ago, a defunct and obviously spammy Vespers Inc LinkedIn AI and Cybersecurity profile, and a Github profile of a "Governmental Forensic Fraud Investigator that shares the online filings' interests in Pirate Stock(s) and er.. other overlaps. Oh, and a LinkedIn profile of a Jacob Vespers, supposedly an Orange County Station Chief involved in "Persona Creation" for the CIA!

I'm inclined to go with this interpretation: https://news.ycombinator.com/item?id=39875809

But I will say that if it turns out that Sam's latest scheme for world domination has been foiled by him dogfooding his own product, then it's very very funny


Hi, author of "GP Hallucination" here.

I think the linked interpretation is very, very reasonable and it's a big reason why I tried to deemphasize all the "stuff" you've highlighted in my piece.

I think most of us would agree the person in question - if they're real - is exhibiting signs of a mental health crisis.

That said, I wouldn't have written the piece if I felt the aforementioned interpretation was the primary explanation to everything I'm seeing in the disclosures. I proactively tried as I went through the disclosures.

The issue I'm having is - regardless of what this person is able to concoct with ChatGPT or other methods - they shouldn't have been able to insert themselves and their Vespers "creation" into OpenAI Startup Fund I GP LLC's CA filings even if they wanted to.

One could argue they either needed to "hack" into the CA filing to insert themselves as Manager/CEO or someone at OpenAI allowed it happen (knowingly or unknowingly is tbd).

It also doesn't help OpenAI won't explain 1) how it happened, 2) why it took so long to catch it, and 3) why it wasn't reported to regulators once discovered:

"[OpenAI] declined to elaborate on how exactly fabricated documents came to be filed with the state of California."

I recognize this sounds a bit conspiratorial, but there are elements (I didn't cover) to this person's filings involving OAI that looks very thoughtful/intentional and are hard to dismiss as random/coincidental/hallucinated actions of someone that likely needs help.

And OAI's "response" didn't help address the actual issues I'm seeing.


Whilst I guess it's not entirely impossible that someone would become obsessed with OpenAI, GPT, online filing, fraud and espionage after OpenAI had accidentally attributed ownership of its venture capital fund to a shell company registered in his existing multiple aliases on the same day, I think Ockham's razor suggests the story fits better the other way round.

(I'm not sure what the security processes for online filings are, but it wouldn't surprise me if someone sufficiently motivated and ambivalent about it being a felony could provide the necessary fake documentation, social engineering and/or brute force hacking to get their registered business given a notional role in someone else's unrelated business on a nonbinding document, at least until the other business noted and corrected the record. I know GPT "hallucinates" real names, but picking the existing aliases of the sort of person most likely to go down those rabbit holes seems quite the coincidence. And it'd have been odder if OpenAI's response was anything other than a generic spokesperson denial)


I just want to comment that the image of the Santa Ana property looks a lot like the main character’s apt in the movie Momento, adding some umami to the whole vibe.


Why have you not covered the elements that make it look very thoughtful/intentional?


Using fictitious names on legal documents is a crime, isn't it? So if those names are truly not real people, then isn't it criminal to put them on notarized documents?


It is a crime, although it doesn't seem to be public knowledge who actually did it. The "Jrbiltmore" account is the most likely suspect, but AFAIK that's just speculation at this point.


I feel like this is more interesting than the original post. Though I am not sure what I am supposed to make out of this: is it 1. shady operations using fake identities with their signatures or 2. an early example of OAI trying to use AI for making money through investments? I lean towards the first one. I mean maybe there is an explanation which makes this seem more reasonable. Do people create boards with fake directors and owners for companies?


The Substack was badly written, and it took me a long time to figure out what their point was. And I might be wrong! I got lost in the purple prose.

Regardless of how shady OpenAI's aims may have been, I genuinely think the root cause of this is that someone at OpenAI foolishly used ChatGPT to automate boring tax paperwork. ChatGPT decided "John Q Vesper" or whatever was a statistically plausible name for the CEO, and this dumb mistake wasn't caught by a human because nobody wants to read tax paperwork if they think a magical talking robot is capable of what seems like a routine task. I am assuming OpenAI didn't intend to tell a ridiculous and easily falsifiable lie in its tax filings (especially if that lie contradicted their public explanations about Altman's management of the fund!). OpenAI probably wanted the paperwork to say "Sam Altman."

FWIW the IRS is generally forgiving of good-faith tax errors, but I would suspect "we spent so long lying about our chatbot that we ourselves forgot that it's dumber than a pigeon and doesn't actually understand human language" doesn't count. Considering how many two-bit lawyers got in significant legal trouble for relying on ChatGPT hallucinations, it would be outrageous if OpenAI manages to get off scot-free here.


Fun fact: I consider myself a pretty bad writer and initially started the newsletter to try and get better at it. The results have been mixed on that front so apologies if you got confused by the "purple prose". I'm trying.

Anyway, my point was:

1. The disclosures point to "OpenAI Startup Fund I" being under the control of a fake company and CEO.

2. The fake company and CEO in question looks like the creation of an AI hallucination.

3. If a fake company and CEO were intentionally used by OpenAI, this isn't a minor infraction and there are highly consequential ramifications to consider given Altman's prior "ownership" of the fund, reported concerns it was being used to circumvent OpenAI Board oversight/governance, and participation in the fund was allegedly a means to invest in OpenAI.

Much of the piece is admittedly a helter-skelter collection of disclosures and explanations focused on supporting #1 and I don't really get into my rationale for #3.

Overall, it makes for bad long-form reading, but it did compel BI to do the needed journalistic follow-up work, made OAI acknowledge the filing was "illegitimate", and pushed the story forward.

We now know, according to OpenAI:

1. The company and CEO doesn't exist to their knowledge. 2. The document indicating this fake company and CEO controlled "Fund I" isn't legitimate and completely fabricated. 3. They're not willing to elaborate/explain how or why these "fabricated" entities got filed by them.

It's ugly spaghetti prose, but I'd say the write-up worked as intended.


Hilarious!


The AGI is now running OAI. Hide your paperclips.


Nah, it's just the mysterious new board member, Hugh Mann.


No need to hide your paperclips. The AGI will make more.


As long as we stop it before it creates hypnodrones, we might just be ok.


The revenge of clippy


Since I’m not an expert in corporate law, anyone know how much of what’s described in the above is legal?


At the very least, any false disclosures in the governmental filings would be illegal, and the bizarre mixture of overlapping names and addresses for the listed directors (as discussed extensively in the article) suggests some degree of dishonesty.

I too am not an expert on corporate law, but very few governments make it legal to lie on these types of forms.


It's like a wilderness of mirrors.


"OpenAI has said Altman does not have financial interest in the fund despite the ownership."

Can someone explain how this could be true?


Sometimes I wonder exactly how much financial engineering a "non profit" entity is supposed to need...


You have it backwards. The non profit is part of the financial engineering, not the other way around.


It simply means he didn't put his own money into the fund.


Might it therefore imply that he seeks to put others but not his own money into the fund?

Or does it imply that he has no intention of profiting from the fund?


why is it so outlandish that it could be the latter?


It could mean there were other agreements, where he didn't put any money into the fund and all profits were contractually obligated elsewhere?


Given that apparently people don't understand the difference between LPs and GPs this unfortunately makes sense


Given Elons lawsuit, how would this impact that is a wonder?

Is it a way to solidly say that Sam had no real say with authority so anything he and Elon agreed was not binding to the company and formally acknowledging that by this action?


Is the fund actually owned by OpenAI the company or OpenAI the non-profit? I thought it was a separate entity that basically had the name and Sam's ownership attached, and was owned and operated separately.


Would have been a great April Fools'.



Not a lawyer, what does this mean? Is this yet another attempt by OpenAI to remove Sam Altman?


How did he do that deal? That's really rich! He should buy the trademark OpenAI and lease it back to the non-profit now to boot! Gotta love a mastermind capitalist like Altman at work! We can all learn a thing or two from him.


You can't register a brand trademark unless you're (already) using it; this is one key difference between copyright and trademarks.


i hope they're forced to convert into a normal for-profit model

give this man the shares he deserves, as well as any seed investors, stop all the petty drama and whatnot and let them be a company, it just makes sense. there was no way to know where the tech was going when they started




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: