Interesting fact - valve no longer works this way because it didn't work.

I've always been curious about claims that such an approach inevitably doesn't work. I feel like my intuition tells me it could with the right people and variation on the concept. But it's probably difficult. Even the best managers I've seen usually have blind spots on how the team is running.

Tons of very proficient people are simply not interested in responsibility that comes with flat structure.

Because of that such company will either stay small (probably leaving ambitious people with unfulfilled ambitions) or change along the way, because you simply will not find enough people to make it grow.

Many ambitious people thrive in a flat environment, though. That's one reason some people like startups.

It's been a few years since I pondered this, but I used to spend a lot of time thinking on this subject. I think for it to be successful you need a lot of the type of person that prefers to work in such an environment, and to make effort to keep it that way. That might mean staying small staffed, but I'm not totally convinced it even needs that.

A good boss is a BS shield, that enhances productivity. I enjoyed startups when I was younger, but now I like to be focused. Fighting fires for a living can result in broad experience, but also less deep work, which can “stunt growth”. At least that’s my opinion/experience. They were also the most fulfilling, that completely disappeared 8 years of my life.

I think not all startup experiences are the same. I had a startup experience that had lots of deep, focused work and not putting out fires. That sort of holy grail is one thing that had me thinking about this topic and how to buck conventional wisdom about what type of environment is inevitable, or what can't work.

But where is the room for ambition in a large company with flat staffing? If you, completely independently, build something adding $2 million of revenue at e.g. Google, you've increased the company revenue by 0.0007%. Even hundreds of millions of dollars remain but a decimal point. The company is simply so large that even if you achieve amazing things, they're going to be largely irrelevant. So the only real avenue for ambition is climbing the ladder of hierarchy. But when that ladder doesn't exist?

One could argue that social economic systems, at scale, are functionally similar to a massive corporation operating on a flat environment. And you end up with a similar issue there. "Company" ambition doesn't really have a viable path forward, and so the optimal solution is to just coast by - and direct your ambition towards your own personal stuff. Or even set the ambition aside, and just enjoy life. At least until the consequences of everybody operating under a similar system of demotivation kicks in.

Some interesting approaches that companies are taking around this are described in Matt Parker's book, "A Radical Enterprise" for example, companies being made up of many microenterprises.

Once you can break away from many traditional imaginations of what a business is and how it should run there are some interesting possibilities.

Very challenging and easy to get wrong though I'd guess.

https://itrevolution.com/articles/radical-enterprise-pioneer...

Isn’t ~50% of people working at companies with less than 10 people?

Somehow that works, so maybe that suggests ~50% of employees at a business could be in self-organising units of less than 10 people?

My intuition is basically its like networking. Using broadcast on a shared medium works fine for small groups, but in big groups it overwhelms. Similarly, in small groups flat structure is fine, but eventually in large groups you have to spend 110% of your time managing relationships and stakeholders, which simply isn't practical.

And that is true even if the group is the best people possible. The only way would be to have no disagreement, but then you would probably be creatively dead.

To put another way, i suspect communication scales quasi linearly with managers, but at least quadratically without.

It works for attorneys and accountants. It’s an approach with ups and downs. The biggest down being it’s harder to scale.

Interestingly though attorneys and accountants charge by the hour to customers. So their individual value to the group is easy to quantify and measure.

And while they may have support staff, they typically "work alone" in the sense that there's one partner per account, so very little oversight or management is needed.

The primary purpose of this "collective" is to create a "brand" of sorts. Ernst and Young has more prestige than say Franklin and Franklin. This draws in higher paying customers.

I feel like there are sufficient differences in this kind of business to a typical software business that the comparison is not especially useful.

Which is not to say that a coop can't work, just that the comparison to accountants and lawyers doesn't work.

Attorneys: associates, junior partners, senior partners, managing partners, named partners. Does not look like flat structure to me.

Like I said, growth is hard.

My brother is an attorney in a small practice of 6 attorneys. They are all partners, share ownership, and rotate who the managing partner annually.

Even assuming those "right people" exist, you'd need a hiring process that could pick them out, and they'd need to want to work for you.

And hope they never change. Say something happens in one of their lives and they no longer contribute anything. Getting rid of them would be exactly what the OP post is criticizing.

I think you're probably right. I've seen it work at small companies. What it doesn't do is scale to large companies, which is the heuristic most growing companies use for any decision. You'd have to say "we'll grow no larger than our flat structure can sustain, even if it costs us billions," and who is going to do that? Not saying it's a good thing, in fact the opposite — optimizing for scale is the source of many of our woes as an industry, in my grumpy opinion.

Technically there is nothing preventing you from stacking 10 eggs on top of each other, If you balance it just right, they won't move assuming no outside forces.

Realistically, you may as well talk about it as something that can't be done. If anything, it's probably easier to stack eggs than to get more than 4 people to work collaboratively without the current incentives and rules making it all hold together.

Valve was never actually flat, they always had a corporate wing setting payscales and directing the company. It's a common misconception, partly brought about by their recruitment marketing.

That's some great marketing. I never applied to work there, I just assumed "now that's how you run a company!"

It didn't really make sense to me. How did you imagine any decisions working?

For technical stuff it works okay, because people can make decisions just by doing whatever it is, or by talking to the team and agreeing on something. It's a lot harder to get people to do that for bonuses and pay.

Right, or "should we spend millions on Steam Deck which will pay off in 10 years" vs "should we all just get a big pay rise now with the money?"

The fact that even Valve, which is one of the companies closest to the idea of having endless free money, and thus the most forgiving to a structure like that, didn't make it work, must surely at least give us pause for thought!

At Valve the corporate wing made (and still makes) those decisions, trying to have ten people agree on who deserves to be paid what and whether it should be invested instead hasn't been tried in any company I know of. In contrast having managers who are collectively answerable to those ten employees through the share ownership structure has been tried, and it works out okay.

> having managers who are collectively answerable to those ten employees through the share ownership structure has been tried

How does that work? As in the employees are owners but the managers aren't?

In this example we're talking about 11 more or less equal shareholders, one of whom is a manager.

Sorry. I genuinely don't understand. Is it that the ten people report to the manager, but the manager also reports to the ten people?

What happens if half of the ten think the manager is good and the other half think he's bad?

But are those 10 professionally qualified to check the manager's work quality during those 10 years? (Or do they buy some audit service from another division?)

It is usually pretty obvious when a manager is clearing blockers, communicating well and not punting on hard decisions; and since you're giving them 10 years, it would also be easy to see the outcomes of their riskier choices.

Oh wow, I had no idea. When did they stop?

I think it was always so, ie cliques of power happened almost right away and a de facto power hierarchy manifested organically. This is what I've surmised from ex Valve people but until we have a book or documentary that reveals what really went on we may only speculate.

Ah, thanks for the info.

Do any sources state that Valve transitioned to a traditional hierarchy? There's been reports of cliques and internal opaqueness going back a decade, but the only concrete difference I've heard versus pre-2012 is a shift towards microtransactions and hardware design.

To be fair, there are worker-owned co-ops that function like companies, and there are those that are heavy on the ideological experience. I once did some consulting for one of the latter ones and it was worse than any bureaucracy I've ever encountered. Everyone has a say, everyone must be heard on everything, decisions take forever and there's a lot of overhead.

On the one side, you'll have a normal company where the workers own the company (well, most employed engineers own some shares in FB, Google etc, too), on the other hand you have anarcho-syndicalist groups that are super hard into the process and THE PLENUM is paramount.

In a public big tech company your ownership is immaterial in terms of your influence on the direction of the company.

Presumably in a smaller private company with 100% insider ownership, it could be arranged that employees have more influence. For example, the employees could vote on board seats or raise and discuss issues at a General Meeting of the company’s shareholders.

This would presumably not be the case in VC-funded startups where the founders and VCs would choose the board and retain the lion’s share of equity and all nominal control.

In practice, in most companies that are not agency-style partnerships the insider ownership doesn’t really matter in terms of influence from what I’ve seen. In agency partnerships there is a dual class issue of partners and regular employees.

A private company with a BDFL who retains 50% + 1 control but has the minority stake dispersed among the rest of the employees seems like an interesting approach. The “B” comes into play in how the majority owner supports decision making. The “D” acts as a forcing function and guardrail to limit the effect of internal politics.

>HN's tendency to GPT whenever exposed to something it doesn't know about is driving me nuts today. :-)

today? checks created date.

I mean, lol, this place was basically blasting anyone who dared say GPT Bing was going to be less than the revolution of our lifetimes. Right before it went off the rails.

For a similar example of the difference between ownership and management, while it's not a co-op, Vanguard Group (the investment firm) is structured so that the company is owned by the funds it manages and thereby by its own investors. The internal structure is still basically the same as any other big corporation, the difference is how the top-down management behavior is incentivized.

>> to GPT

Nice verb, is this an original coinage?

love it, I'm going to use it when someone talks without knowledge of what they're talking about.

I've said this elsewhere, but GPT's probabilistic filler, emitted without concern or awareness of whether it's true or not matches the academic definition of "bullshit." :)