First of all, the secret ingredient is labor law.

Second of all, here's one prediction for how this market could go: One day someone with a billion dollar company will /present themselves as/ a one-person operation, as part of their PR blitz, and you'll be enraptured by them and resolve this as a vibes-based, spirit-of-the-market yes. The nitty-gritty truth of the matter will not be investigated.

@marvingardens I invite you to view to comments to get an idea of the rigor I aim to bring + my intent to consult traders of this market. I see Manifold as a repeat game, I value the trust traders place in me to resolve markets fairly.

@Putcallparity I read all the comments before making my own. I see that you're basically unsure about the lines to draw, but another question that's just as important is: How will you get reliable information? It's likely enough that someone will make this claim about their company some day. Their claim may lie in a clear-cut zone or an edge case. Media outlets may run with the claim or not. How do you figure out if it's true?

>My discretion will be used to adjudicate how contractor use applies to the “one person” criteria.

Well, it sounds like you're going to make up a resolution based on your feelings, then. Is it a one-person company, or not? Very simple to answer. Pretending to make it complicated, makes this market much less useful as a predictor.

What if it isn't a tech company

As this market gains more traction I’ve been thinking about how to account for contractors or other affiliated humans that aren’t technically employees. From launch I’ve always maintained that I’ll use my judgement (and won’t bet in the market) however I’d like to share my thoughts and proposals as early as possible to avoid any controversy.

—

For resolution, I’m proposing an “above vs. below the API line” rule for counting headcount. Anyone above the API line—integrated into building/operating the product (even if legally a contractor)—counts as an “employee” for this market. Anyone below the API line—arm’s-length work mediated primarily by an app/API (tasks/assignments routed through software)—does not count toward the employee number.

Open to hearing @traders thoughts here!

@Putcallparity Thank you! I appreciate clarifying things as early as possible. You might want to add this to the description.

@Primer will add to the description if there are no substantive strong objections over the next week or so!

@Putcallparity My expectation as a trader was that it'd follow whatever Wikipedia's definition of "employee" is for when you look at a company's Wikipedia page and it says they have 5000 employees. I imagine that's probably roughly "number of people getting w-2s"?

@Putcallparity (Getting out of the market—it still seems to me that there will probably be a “one-person unicorn” in spirit, but perhaps not in technicality—as I’d expect the founder-operator to potentially hire a human to open their mail, vacuum things, or guard their computers.)

@TannerNewell this is interesting, ty for your feedback. I want to be clear that I am not married to my definition. I am seeking feedback (like yours) to formalize (as much as possible) a market that keeps to the spirit of a one person unicorn.

@Putcallparity I think a large amount of trading was already done assuming you actually mean "one person" not "one person in spirit". I'll also note that the "in spirit" becomes IMHO very vague and hard to verify with possibly many edge case (which type of employees would you not count? Secretaries? Compliance officers? Security? Cleaning/facility? Too many options there each could be plausibly argued to be tangential to the "true mission", though they could actually be employed) Whereas getting the information on number of explicit employees is at least in principle possible (some difficulties with contractors notwithstanding).

@Putcallparity This is nonsensical. If the company is paying anyone else to do something, then it's not a one-person company.

@ChurlishGambit (Please clarify your comment—seems nonsensical as written—a one-person company would reasonably be expected to pay someone to, e.g., deliver their products or file their taxes without suddenly becoming a more-than-one-person company.)

@TannerNewell I disagree. That's the sort of CEO exceptionalist myth-making SV is founded on—only certain people "count." A true one-person company, is a one-person company. Doing all of the work tasks, including accounting & janitorial, on their own. No cheating by excluding people because the "visionary founder" deems them non-persons.

@ChurlishGambit I understand your sentiment here.

To clarify, if the company in question has any W2 employees (excluding the founder) this will fail my criteria regardless of what services are provided. Contractors hired by the founder regularly perform any given function will also fail my criteria. The edge case I initially foresaw with contractors had to do with contractors hired by AI agents to perform real world tasks. I’m curious to see where @traders land here.

In response to your comment that “paying anyone else to do anything” would result a failures of the markets criteria, below are some edge cases around services rendered by non-employees. I’m curious to have your (or others) thoughts (ie would each of the following count as a 1 person company).

• Solo founder rents a desk in a coworking space which includes services rendered by staff of the coworking space (e.g. cleaning, security, IT infra)

• Solo founder hires a lawyer for a few hours to review a legal document

• Solo founder pays a notary to notarize a legal document

• Solo found works from home and a technician from the ISP comes to deliver or install equipment

• Solo founder sends legal documents via courier

• Solo founder works from home and their significant other helps clean their office

@Putcallparity I like the exclusion of contractors hired by agents. If a founder doesn't have to interact with the contractors, it probably feels like solo-founding.

Obviously, I'm biased as a yes-holder, but trying to look at the spirit of the question as I understand it, all those edge cases would count as 1 person company to me. If you took it to the extreme otherwise anyone using AWS or other cloud services could be disqualified.

That brings up an interesting set of failure conditions for software companies: If another human has write access to the repo or a role in any of the cloud service accounts, it wouldn't be a 1 person company.

I'm thinking maybe the right line here is "unique humans whose primary source of income is the company." Like, if I hire an independent contractor to do 39 hours a week of work for me (or teenagers working 10 hours a week for me but not working for anybody else), they're really my employee. If 500 AWS employees are spending an hour a month working on stuff for me, they're not really my employees, I'm just buying a product from Amazon that happens to involve humans doing work that benefits me.

In particular, IMO (as a biased NO holder) Uber would not qualify as a single-employee company even if Travis Kalanick were the only W-2 employee and the zillions of Uber driver humans were living below the API.

@musteval I like "unique humans whose primary source of income is the company."

@ChurlishGambit (We might want to point out a distinction between things shaped like a “marketplace” [e.g., Etsy] that probably could be one-person vs. what, Instacart or Uber?)

[Edit: oops, meant for @ZicoVerona]

@ZicoVerona Why? If someone were working part-time for this fictional company, & part-time somewhere else, why would they not "count?" If the CEO had a spouse who put in a bunch of unpaid labor, why would that not "count?"

@TannerNewell No, that's making the same error to an even worse degree. Etsy IS InstaCart & Uber, but for trinkets & bad art.

@Putcallparity "The edge case I initially foresaw with contractors had to do with contractors hired by AI agents to perform real world tasks. I’m curious to see where @traders land here."

So as long as the mythical superhero Founder has a chatbot hire the contractors, they don't count? How dehumanizing. Bad loophole.

• Solo founder rents a desk in a coworking space which includes services rendered by staff of the coworking space (e.g. cleaning, security, IT infra)
  
  Not a one-person company. That's just having contractors mediated by another company, not unlike your chatbot scenario.

  
  

• Solo founder hires a lawyer for a few hours to review a legal document

Not one-person, you've literally hired someone to do some work.

• Solo founder pays a notary to notarize a legal document

This is one that makes sense to be an exception, as you are required by law to have a notary do this, & you can't self-notarize for anything important.

• Solo found works from home and a technician from the ISP comes to deliver or install equipment

If they do the installation, not one-person. You can DIY that, & these days all it takes is plugging it in basically.

• Solo founder sends legal documents via courier

If it's across town, not one-person. If it requires major travel, that's a reasonable exception.

• Solo founder works from home and their significant other helps clean their office

Not one-person. Spouses (especially wives) already are badly under-credited, particularly in the cases of "founders" & also writers. We mustn't perpetuate that sort of thing.

@TannerNewell @ChurlishGambit I don't decide what counts. I'm just trying to elicit how impressed I would be by someone getting to a 1-billion valuation alone, which is what I understood the original question as. For that a spouse doing unpaid labor would invalidate that achievement to me. A promiscuous contractor wouldn't invalidate it to me per se, unless they had write access to the repo or cloud account permissions.

If someone built something like Etsy, YouTube, Uber, Amazon with sole write access to the repo and sole cloud account access, and got to unicorn evaluation that way, I'd be sufficiently impressed. So even "unique humans whose primary source of income is the company." would have those exceptions for me. It's up to the question creator of course.

@ZicoVerona (I like Zico’s metric—because the question is probably getting at a vibes-based single-person impressiveness based on common impressions of what a “company” or “employee” is.)

Whereas @ChurlishGambit is going down a road where the only “real” one-person unicorn company looks like a person on an island who values their island at $1B—because interacting with other firms in the economy would start classifying their workers as employees of the unicorn.

In general, if a founder only talks to AI agents (and maybe users), that would fit 'YES' to me. The resurgence of the solo-unicorn-founder meme is based on the recent AI hype anyway.

@ZicoVerona But if other humans are doing a lot of the work, how is that one-person?

@TannerNewell (I don't know why you insist on putting everything in parentheses, but it doesn't make misrepresenting my position any less rude)

@Putcallparity I can get behind all sides here, everyone has good points. Thanks for this discussion.

I'd be disappointed if this would resolve to a technicality though, like if the company had 200 people working for it, but the founder only communicates via AI.

@ChurlishGambit Other humans outside of solo-founders have always been doing a lot of work (cloud providers, the people who built the laptop the solo-founder uses etc). Again, repo write access and cloud account permissions are cruxes for me.

@ChurlishGambit By these definitions I'd argue that there are basically zero companies worth even $500 operated by just one person. I think the question is getting at something like "does the ceiling for what a single-person company can be worth go from ~$5M to >$1B by 2030?", not "from $0 to >$1B".

@Primer Yeah, I can see some transition phase where AI agents are good enough to manage people but not good enough to do the delegated work themselves. But if those 200 humans were "unique humans whose primary source of income is the company." and had repo write access or could service permissions, I'd not be impressed. However if they were gig workers and the solo-founder had sole repo and cloud access and non of the workers have equity, I'd be sufficiently impressed as I now would be impressed by a solo-unicorn-founder in the news.

@ChurlishGambit (Parentheses are mostly denoting a side-note or clarification [or request for clarification]—and probably not of huge interest to someone scanning the comments.)

It seems like your position (mistakenly) generalizes from a first-order “company with one employee-founder-operator whose other interactions within the company [!] are with AIs” to a second-order “company with one employee-founder-operator who does not purchase [in various senses] any other services from any other human at all—only AIs (and also maybe those AIs can only purchase from AIs…?)”—and… that’s a position, but I don’t think that’s what this market is about.

(I think it’s closer to the first one.)

@TannerNewell @ZicoVerona I think the spirit of the market felt something like

Could a single person set up something like Airbnb alone (including things like all the coding, ads, finance, legal stuff, website, apps, etc. [potentially by using AI for most of it])?

and not

Could a single person set up a scheme where they use an AI to set up a company and let the AI hire contractors who technically aren't directly employed by the founder.

@Primer (Agreed on the spirit of the market—my “definitive” take is—)

It seems to me that:

A "one-person [one-human] company" should definitely be allowed to contract with other firms that employ humans—like cleaning, legal, or accounting services—and there’s an argument to be made from there that directly hiring a human to do that work “might not count” if sufficiently separated from the “real” business. (And I feel confident that the first part is in line with the popular understanding of words like “company” and “employee”.)

A "one-person company" should probably be allowed to be shaped like a middleman or a marketplace—something like Stripe or Etsy—that takes a cut of the transactions that other humans bring to the platform—but something like Uber’s drivers seems much deeper into a gray area.

Gray areas like: “but humans delivered the robots to the factory!”—also seem like they should resolve towards “whatever, it’s still a one-person company”.

[Not least because to do otherwise turns this into a boring goes-to-zero-on-a-technicality question. What’s interesting about the question is, “Can one-person plus a bunch of AIs make a billion-dollar company?”—not, “Will that person also avoid hiring a private chef along the way?”]

@Primer Even the second scenario would impress me sufficiently, because new AI agents would allow people with very limited social skills to make $1B without talking to humans - assuming those humans don't have equity or write access to the repo etc. But I agree that it would be a weaker case.

@ZicoVerona

>Other humans outside of solo-founders have always been doing a lot of work

Yes, that's why the "solo founder" myth is bullshit

@TannerNewell Even though you continue to be extremely rude, I'm going to clarify my position for you one last time.

"One-person" company means, one person doing all the work. If you have contractors, if you hire a company that hires contractors for you, if you have a slop-bot hiring contractors for you, then it's not a "one-person" company.

A "one-person" company is nearly impossible. Businesses, despite the rabid wishes of CEOs, cannot succeed without the efforts of lots of people.

I think the culture in America, particularly in SV, of dismissing labor, is both morally wrong & factually deranged. I can see you drink deeply from the Flavor-Aid of SV myths, & are upset, but at least don't put words in my mouth.

@musteval

>By these definitions I'd argue that there are basically zero companies worth even $500 operated by just one person.

Correct

@ChurlishGambit "A "one-person" company is nearly impossible.". With LLMs steadily improving and becoming more agent-like, this is rapidly moving from "nearly impossible" to "probably not". I think what brings the probability down is: It doesn't make sense. If I alone can get to a billion in 2029, I'll hire a few people and get the billion in 2028 instead and 10 billion in 2029..

By the way: Would Satoshi Nakamoto count, if we were certain Bitcoin wasn't a group effort? (How strict are we on "company"?)

@Primer

>With LLMs steadily improving and becoming more agent-like, this is rapidly moving from "nearly impossible" to "probably not".

If an LLM is hiring contractors for you, then it's not a one-person company.

>If I alone can get to a billion in 2029, I'll hire a few people and get the billion in 2028 instead and 10 billion in 2029..

If you're gonna do that with LLMs, I have real bad news about what's about to happen to the LLM industry. Inflation after the crash might make a billion more like $500 though, so maybe!

>By the way: Would Satoshi Nakamoto count, if we were certain Bitcoin wasn't a group effort? (How strict are we on "company"?)

By the time BTC was worth a billion, weren't lots of people contributing?

@ChurlishGambit

By the time BTC was worth a billion, weren't lots of people contributing?

Yeah, I don't think you can argue that way. If the story around Bitcoin is remotely correct, not a single person was ever hired by the creator of Bitcoin. Satoshi coded something up, put out a white paper, wrote some posts and mined some Bitcoin. If we knew Satoshi was a single person and still held enough Bitcoin, we'd just need to think about the meaning of "company".

@Primer What? People contributing code doesn't count, why? People designing the website doesn't count, why?

@ChurlishGambit Of course those should count. Did other people contribute code or design a website (which website?) before Satoshi disappeared with his wallet? If yes it probably shouldn't count as "single person".

@Primer

>Of course those should count. Did other people contribute code or design a website (which website?) before Satoshi disappeared with his wallet?

Yes lol it sounds like you may not actually know much about the early history of BitCoin

@ChurlishGambit Maybe. But my point is: It's not that far from a one-man unicorn.

@Primer It is VERY far from that. Please go read some real info on the development of BitCoin, I'm begging you.