Then again, I go the extra mile because I actually like to, not because there's incentives. When I do something I want to do it for the emotional reward, not for the money. So I'm kind of afraid of the money messing with me psychologically.

But you may not notice the issues with only with 3 people. Try comparing 30 vs 10.

[1] http://en.wikipedia.org/wiki/Conway%27s_Law

This poll doesn't really pertain to hiring super-talented people who build JS VMs for Google or rich-UIs for Khan Academy. I'm just talking about productive developers who can do the same thing that usually requires a team of 3-5 average developers.

[1] http://blog.jayfields.com/2009/01/cost-of-net-negative-produ...

[2] http://en.wikipedia.org/wiki/NNPP

I think your definition of 3x more value sounds pretty good. There is no single definition of good/value/great/perfect, it's a balance of many characteristics. But for this question, assume they are equal in both.

1) At a 3x salary you'll need to lose two other employees and sacrifice the redundancy, idea mix etc that you point out. Reducing the HR etc overhead by two employees is far less of a gain.

2) Their higher output is going to be at least partially due to more intensive work habits which may not be sustainable over the long run.

3) If they ever decide to move on you need to hire three replacements and hope they work out and can take over from your superstar without losing too much momentum (all without ever meeting the guy since its unlikely you'll find & hire them within his two week notice period).

I don't know if (2) is necessarily true. Experience, skills, talent, and ability to focus well could easily make this sustainable.

In my experience, (3) happens when they were hired at 1x salary, proved themselves to be 3x as productive, but continue to get just 3% cost-of-living raises. If you compensate them well (and the work-environment remains hospitable), they have very few compensation-related reasons to move. Nevertheless, I agree that they can still leave due to personal reasons.

Many people (including op) look at the 3x in a linear fashion. They think the 3x programmer is working 3 times faster, like a 3ghz CPU to 3 1ghz CPUs, or that they work 3 times longer, or a combination of these. This leads to 3 times more code, documentation, delivery etc. of exactly the same quality

I've never met such a programmer

Usually the <magnitude>x comes from better quality, and less quantity. So the 3x programmer has achieved 3x the functionality as requested, but that code base is likely 2x - 3x simpler, with 3x less bugs, which combined is closer to 6-9x easier to maintain. This code also influences the cost of code built on top of it. So the value increase in some areas is exponential rather than linear. You're unlikely to get this exponential increase in value from 1x developers

Another exponential value characteristic of 3x developers is that they have much better decision making. Maybe that means choosing the right tool for the job (choosing a distributed DB in stead of MySQL when you know you need scalability) or designing a customer facing UI. Recovering from a wrong decision can be very costly

These exponential traits mean that the earlier you are in a project, the more value there is in hiring that 3x developer

I only need one skilled person to solve the curly problems, what I really need is people who do what they say they are going to do without screwing things up.

That said that one person will probably be in an elevated position of responsibility and have an elevated wage respectful of that position. So as a senior role he would have more, but not to skill alone, also responsibility.

I think it would be viable to have all programmers like this at a single company. I've heard anecdotally that some financial firms will pay salary in the 500k range in order to (a) attract world class talent and (b) ensure loyalty.[1]

[1] For example: http://en.wikipedia.org/wiki/Sergey_Aleynikov . According to Wikipedia, making $400k before he quit, and he was involved enough in coding to try to take some of it with him. I actually don't know that this is materially more than he would have been making at Google, I tried checking glassdoor.com but what is listed there doesn't match what I've heard.

And if all the smart people think like I do (and there's enough dumb managers to keep productivity-linked pay from actually happening), that's an even bigger reason.

Sorry if that's not an encouraging answer.

Hiring someone with a 3x salary just because people say he's good or is a Rockstar in the community doesn't work for me.

I experienced several situations where a Rockstar didn't match expectations and unknown guys had to clean Rockstar's mess when they left.

I'm not saying that Rockstars are shit and you should not hire them, what I'm going to say applies to everyone:

Hire this person for a X amount. If he excels at his tasks delivering what you want, and you realize that his performance if above the average, raise his payment and both of you will be happy. Both sides win.

Risk of the person leaving is probably highest on the list.

You see 2x or 3x the normal rate in consulting. You are paying for a temporary boost in expertise.

Salary wise it usually doesn't happen because you don't want that much risk.

Assuming that leaving could be solved for in some form of contract or golden handcuffs, I would consider it. As much as there are drawbacks there are also benefits. Such as a person with 3X the companies overall technology picture in their head. A flatter knowledge structure. More accountability, so it does cut both ways, I would have no issue paying 3X for the tradeoff, so long as they where not my sole employee. In a sole employee situation I would prefer hiring a 2X and a 1X over a single 3X unless it was a temporary project based effort. The harder issue up front would be how to identify the 3X developer. I have a good process for spotting talent, but there is a gradient within that talent. Some are a good deal for market rate, other are what are refer to as the 10Xers and worth far more than they can command. Both are worth hiring and are above the industry but spotting the 10xer up front proves to be a daunting task.

I would pay someone 3X the salary to someone 3X "above average," but that does _NOT mean that s/he would accomplish the work in 1/3 the time at the same quality.

The difference between _brilliant programmers and average programmers is MASSIVE: someone 3X as "good" needs to be paid 3X the salary because they can create something new, imaginative, life-changing.

What is the best team size?

Let's say it's 3. (1 has too much risk and 10 has too much overhead.) Once your productivity needs exceed what 3 average developers can accomplish, if you place importance on team size, it makes sense to pay more for higher performance in order to avoid increasing the team size.

It's like paying to upgrade hardware on the network nodes you have (which exist to mitigate risk), rather than distributing work across more and more network nodes.

I'd rather have something developed in 10 days at a daily rate of $500/day rather than have it take 20 days at $250/day.

However, I wouldn't fire 3 developers to get one "amazing" developer. What happens if he/she gets ill or decides to leave? You're suddenly left with no developers on the team.

I think the short answer is that startups should be hiring for 10X devs, without having to raise their prices.

But, in the real world - you usually get what you pay for.

As an employer, its your goal to get as much done properly as possible for as little as possible

An employee doing 3x the work for 3x the salary is no more effective than 3x workers, as yu explicitly said all else was equal.

Employer snags a 3x for 1x salary and they are winning.

(Why I find self-employment = self-fulfilment)

If you don't have productive environment & plans, even these guys can perform below 3x I guess.

That being said, I wouldn't pay them 3x in straight salary. Equity and profit sharing is a better motivator.

Equity is very similar to the lottery... a few will win hugely large while the masses will mostly receive nothing. Thats how the system works. Ive taken equity and cash at different times over the years. Its not a scientific breakdown but through experimentation the decision to take cash over equity has been reinforced over and over again. Additionally despite my one very fortunate friend, everyone else I know in the industry is far from struggling financially but generally agrees from their experiences that equity is very much just a lottery ticket. (And even if your company does end up being 'the one', you still risk getting screwed out of your shares.) I'd rather have cash.

Or, do you believe it is impossible to know, or at least approximate odds of, any company's chance of success?

Let me put it to you yet another way. Which is more likely to go out of business in the next two years, blockbuster or BP? If you say blockbuster, why?

So now in regard to gaging the success of your company and your investment... I think it pretty easy to forecast the coming trends in the technology space. For example I think its pretty easy to foresee that the internet is going to eventually eat traditional cable providers as its great at eliminating the middleman. There are going to be some big winners in this space and there are lots of early stage companies looking to ride this. But who out of them is going to be the big winner or few winners of this space? - your guess is as good as mine. As an employee I have to pick one but as a VC or investor I can invest in the general space knowing that somebody in my portfolio is probably going to win this market.

This is the reason I take cash - its not that I don't want to invest in equities, its that I want to have the freedom to put that cash into diversified equities.

If I were go back and look for an equity heavy gig, and I don't plan to, I would look at the entire process as an early stage investor. No, I can't invest in 10 or 20 different companies, but I can be selective and increase my odds. Actually, anyone with my current skillset could walk into any startup in the valley and start working the next day. That is something few, if any, angels could do.

With that in mind, I'd say that the industry is already paying 3x developers around 1.5x to 2x and lately it is moving into the 3x range.

First let's address the case where it is a large corporation:

- The risk of the 3x-dev leaving wouldn't be much as it is a large corporation and hence there are mechanisms at place that are ready to fill the 3-devs gap he will make if he left at any particular time.

- It is very economical to pay 3x salary for the 3x-dev as it saves on managerial costs.

- Promoting the 3x-dev to a team leader or a senior position is a good strategy as it will increase retention chances with him and also promotes sharing experiences and knowledge.

* Hiring 3-devs is rather a tedious task and have a significantly higher managerial cost, and the risk of them leaving is even higher as they won't feel that much special, hence no early rewards and the eventuality of them looking for other options out there.

* I daresay there is a statistical certainty that the 3-devs won't be on the exact same knowledge/experience level which will result into communication overhead that includes the assumption that they are good at communicating in the first place.

Office politics will make things even harder if you promote one of those 3-devs - that you hired at the same date - to reward a personal achievement of sorts.

Now in a small company/startup:

- Hiring the 3x-dev involves a high risk if he's not rewarded often with more than just a bonus or a salary increase, like a piece of the company that he helps building.

- Hiring him will cut down on managerial expenses greatly, and will result in a much smoother development process.

- There's no body that of concern to compete with head to head and even if there were, they are supposedly experienced and professional enough to keep things light between them and focus on delivering results.

Hiring the 3-devs have a very high risk if they _are_ rewarded equally - which you hired them at the same date - simply because you are normalizing their efforts and hence creating a class of sorts which can and in fact will backfire. Now if you decided to pay attention to every developer separately, that might create jealously, which can be a positive thing, but it largely depends on the company culture/atmosphere you have and it is quite tricky to maintain.

* There's a managerial overhead that won't be easy to maintain since this is a small company and you yourself have things to do better than maintaining logistics/communication and team spirit.

* The fact that this is a small company by definition will require a diversified skill set, flexibility, and not just willingness to learn but rather knowing how to consume knowledge fast and turn it into an actionable skill, while this is questionable although highly likely, your 3-devs will find it difficult - and quite overwhelming - to take over the many tasks that require different skills in the tight time window and stress that comes with all small companies. But this also depends on how you structure your company culture/political-atmosphere, and yes there will always be company politics whether you like it or not, even if you intentionally and consciously try not to develop them.

All in all, I believe that every environment and work place is quite unique. But from personal experience over multiple quite different industries, quality comes before quantity.

So my short summarized answer is to go with the 3x-dev and give him 2.0x-2.5x if it were a large company with short period rewards, or give him 1.5x-2.0x if it is a small company with emphasis on long term benefits/interests coupled with exciting/teasing short term rewards.

I've done a lot of interviews, and I can tell you, with the exception of a few famous superstar developers, it's practically impossible.

So yeah, the answer is "yes", but the question is equivalent to "Would want to always have perfect health, if you had that option?".

For this experiment, let's say you could compare their past projects (code, product) and they were on equal footing in quality/performance. Or they are consultants you've worked with, have a good measure of their productivity/work-quality, and want to hire 1-super or 3-average of them full-time.

I do not think that this is how it really works. It is more that that one person understands something that these three others do not. For example, the problem can only be solved with a complex regular expression. This guy understands it. The three others do not. That is why you do not see the difference in 99% of the project. It is in the 1% that truly requires his talent, that the 3 guys get stuck forever, while our guy manages to complete the project. If your project does not require any particular, rare understanding, you will not even see that this guy is so good.