He's not following in Bill Gates's path at all. 2 years after it was founded, Microsoft had 6 employees. Facebook has more than 10 times that. Microsoft sold DOS to IBM for something like $75k, which was around the same as FaceBook's revenues last year, and significantly less than the total amount of venture money invested in FaceBook. Microsoft didn't go public until 10 years after its founding, and its market cap then is well under what Mark is asking for FaceBook now.
Bullshit. Pay every FaceBook user $100 to switch to TheMoneyBook.com. For $100, I'd do it, and I bet I'd have no problem convincing my friends to follow me, particularly if they also got $100 for it.
Total cost: $500 million, less than TheFaceBook.com has already been offered.
People are idiots.
I've got residential fiberoptics. 5Mb/sec downstream. Verizon sells it for $30/month in the Boston area. For an extra $10/month, you can get 15Mb/sec downstream.
So the answer to the "when" question is getting pretty clear. Now.
I think the big opportunity is really in rackmounts, not desktops. The electricity cost of a typical desktop is really quite low, as mentioned below, and most consumers just don't care. But HVAC is quickly becoming a limiting factor in big server farms; look at Google's paper on their cluster architecture. Electricity may be cheap, but real estate is not, and anything that lets you pack machines more densely without having your server farm go up in flames is a big win for major web businesses.
Disjunctive types are basically like cleaned up C unions, without all the safety holes. For example, the Ocaml statement
type 'a tree = Leaf of 'a | Branch of 'a tree * 'a tree
creates a binary tree where all data is stored in the leaves, while the type
type 'a list = Nil | Link of 'a * 'a list
creates a singly linked list. You can read that in English as "A tree either a leaf containing an obect, or a branch of a left subtree and a right subtree" and "A list is either Nil or a Link containing an object and a list".
(These look a bit more complicated than they have to because they're polymorphic; each is really equivalent to a C++ template class that can be parameterized over any type of contained object. You could define a list of integers only with
type int_list = Nil | Link of integer * int_list
but that would limit the types of objects you can store in the list.)
Pattern matching is a way to destructure a disjunctive type and account for all cases. For example, to compute the length of a list, you'd write this function:
let length the_list = match the_list with Nil -> 0 | Link(_, tail) -> length tail + 1
(Indentation is all messed up there, but I can't figure out how to fix it.)
Basically, instead of thinking of the type system as expressing bits and bytes, think of it as expressing possible values. Those values may themselves be parameterized, like how a Branch of a tree contains two other trees. Then when you pattern-match, you tell the compiler what to do for each possible value that the type could have.
Lawsuits are expensive, even if they're dismissed on prior art grounds. Many companies patent every random bit of technology as a deterrent, so that they can say "If you sue us, we'll find something that you're infringing and sue you back to the stone age." It's like mutually-assured-destruction from the cold war days. Saves on legal bills for everyone.
They might've done better had they not advertised that "they were going to destroy Microsoft", though.
Basic rule: publicity for feedback is good. Publicity for ego is bad.
If they'd had more openness during development, they might have heard the collective yawn that followed their actual introduction and reconsidered their business strategy. Maybe they would have gotten feedback about the price point that people were willing to pay for one, and adjusted their engineering accordingly. Or maybe they'd have canceled the project. But they'd have more information to base their business decisions on.
You assume that you can create demand where none exists. In some very rare cases (eg. Steve Jobs) that's true. But most of the time, if you launch a product when most of your customers say they want the status quo, you're in for a big flop. Better to take the customer feedback and cancel the project.
It's clearly true in every market I can think of, and also follows from general economic principles. The number of farms is going down; the number of phone companies is going down; the number of computer manufacturers is going down; the number of cable companies is going down; number of retailers is going down; etc.
How about the number of social networking sites? The number of portable MP3 players (okay, so Apple seems to have sewn up that market recently)? The number of blogging software products? The number of biotech firms? The number of private sellers on E-bay? The number of hybrid-drive cars (though most are made by existing manufacturers)? The number of temporary-file-storage sites, a la YouSendIt.com? The number of airlines (I can't remember whether that's going up or down right now; it fluctuates every 5 years or so)? The number of microcredit lenders operating in the third world? The number of firms in China?
What you see is consolidation, the process where all the also-rans in a mature industry drop out or get swallowed up by the market leader. But consolidation applies only to mature industries, where the market is saturated and most consumers are satisfied with the existing level of product quality. All around them, new companies get spun off to take advantage of new markets and new technologies. Sometimes even mature industries get eaten up from the bottom, like how U.S. Steel and Bethlehem Steel got replaced by Nucor and other minimills.
If it looks like the economy is settling down to a steady state, it's because you're looking at the stable portions of the economy. There are areas of the U.S. economy that are growing at 1000% or more (eg. IPods). There are areas where every Tom, Dick, and Jane has started their own business and they all hope to strike it rich (eg. social networking). There are undoubtably areas that could support the next Netscape, Google, or Microsoft, but if I could give you an example I'd be out there working on it instead of writing on Slashdot.;-)
Steve Jobs's gift has been the ability to spot those areas. If it had just been the Apple 2, I'd chalk it down to luck, but he subsequently got lucky on the Macintosh, Pixar, the IMac, ITunes, the IPod, and now the Mac Mini. And I suspect that his decision to drop out of college had a lot to do with this, as it forced him to go out and search for opportunities. Most people get tunnel vision and only see what's right in front of them. If you want to make money, you've got to see what other people don't.
When I say competence is a commodity, I mean, for example, that if you design a better PC, every other PC manufacturer can duplicate anything better about that PC in the course of a few months.
So don't design a better PC. Let all the other competent people in that industry design one. Put your talents to work in an industry where competence isn't a commodity.
Go design an IDE for Ocaml or Haskell. Write some AJAX apps that do stuff people think can't be done over the web. Build a mobile computing device that lets you browse the web and doesn't have a screen the size of a postage stap. Write a programming language that can adapt and work around its own bugs instead of crashing and burning.
Anyway, the free market system rewards, more than anything else, not competence or excellence but size. The bigger your company, the better. This is basic economics, although not a part that libertarian idealogues like to emphasize.
And less basic economics then suggests ways that smaller companies routinely and reliably beat big companies. I recommend The Innovator's Dilemma and
Almost all "superior" market failures fail because they are superior in ways that the market finds irrelevant and deficient in showstopping ways. Using your examples:
Betamax could record only an hour, while VHS could record 2 hours for standard play and 6 hours for extended play. A typical movie is about 2 hours. That makes a huge difference. I own both a Betamax and an early-model (c. 1983) VHS, and the VHS was much more convenient. We could stick a whole trilogy of movies onto one EP VHS tape, while it'd take multiple Beta tapes just for one movie.
The Amiga's features were perfect for the early-adopter geeks that frequent slashdot, and utterly irrelevant for the corporate market that IBM targetted. Somebody doing word-processing does not need 409 colors in 640x512, nor do they need sound. They need good productivity programs and lots of corporate support, neither of which was the Amiga's forte. And have you watched an over-35 non-techie use a computer lately? Many of them will open AOL, check email, talk to their kids, close it, open IE, check the weather and their stocks, close it, open Word, get work done, close it, open Outlook, e-mail a report to the field office, close it, etc. They don't need preemptive multitasking now, let alone 25 years ago.
Okay, I haven't used OGG Vorbis, so I have to go on what friends have told me. As I understand it, it gives superior quality and higher compression ratios. Which is nice, except that the audio format market is a commodity now. Almost all mainstream formats are "good enough" in quality, bandwidth is cheap enough to make MP3 downloads reasonably fast, and most players can play all the mainstream formats. In conditions like these, consumers go with the brand they're familiar with, because they have no reason to switch. It doesn't matter how superior you are if the consumer has no need for a superior format.
Understand that the free market works for the betterment of the world. It does not work for the betterment of Slashdot readers. There are many products that are very appealing to the SlashDot market (eg. Linux), and they have substantial market share in markets (eg. servers) where Slashdot readers make up a significant proportion. But there are lots of other people out there that have different criteria for "superiority", so the eventual market winner may not reflect your notions of a superior product.
Slashdot Mirror
He's not following in Bill Gates's path at all. 2 years after it was founded, Microsoft had 6 employees. Facebook has more than 10 times that. Microsoft sold DOS to IBM for something like $75k, which was around the same as FaceBook's revenues last year, and significantly less than the total amount of venture money invested in FaceBook. Microsoft didn't go public until 10 years after its founding, and its market cap then is well under what Mark is asking for FaceBook now.
Bullshit. Pay every FaceBook user $100 to switch to TheMoneyBook.com. For $100, I'd do it, and I bet I'd have no problem convincing my friends to follow me, particularly if they also got $100 for it. Total cost: $500 million, less than TheFaceBook.com has already been offered. People are idiots.
I've got residential fiberoptics. 5Mb/sec downstream. Verizon sells it for $30/month in the Boston area. For an extra $10/month, you can get 15Mb/sec downstream. So the answer to the "when" question is getting pretty clear. Now.
I think the big opportunity is really in rackmounts, not desktops. The electricity cost of a typical desktop is really quite low, as mentioned below, and most consumers just don't care. But HVAC is quickly becoming a limiting factor in big server farms; look at Google's paper on their cluster architecture. Electricity may be cheap, but real estate is not, and anything that lets you pack machines more densely without having your server farm go up in flames is a big win for major web businesses.
(These look a bit more complicated than they have to because they're polymorphic; each is really equivalent to a C++ template class that can be parameterized over any type of contained object. You could define a list of integers only with
but that would limit the types of objects you can store in the list.)Pattern matching is a way to destructure a disjunctive type and account for all cases. For example, to compute the length of a list, you'd write this function:
(Indentation is all messed up there, but I can't figure out how to fix it.)Basically, instead of thinking of the type system as expressing bits and bytes, think of it as expressing possible values. Those values may themselves be parameterized, like how a Branch of a tree contains two other trees. Then when you pattern-match, you tell the compiler what to do for each possible value that the type could have.
Lawsuits are expensive, even if they're dismissed on prior art grounds. Many companies patent every random bit of technology as a deterrent, so that they can say "If you sue us, we'll find something that you're infringing and sue you back to the stone age." It's like mutually-assured-destruction from the cold war days. Saves on legal bills for everyone.
They might've done better had they not advertised that "they were going to destroy Microsoft", though. Basic rule: publicity for feedback is good. Publicity for ego is bad.
If they'd had more openness during development, they might have heard the collective yawn that followed their actual introduction and reconsidered their business strategy. Maybe they would have gotten feedback about the price point that people were willing to pay for one, and adjusted their engineering accordingly. Or maybe they'd have canceled the project. But they'd have more information to base their business decisions on.
You assume that you can create demand where none exists. In some very rare cases (eg. Steve Jobs) that's true. But most of the time, if you launch a product when most of your customers say they want the status quo, you're in for a big flop. Better to take the customer feedback and cancel the project.
How about the number of social networking sites? The number of portable MP3 players (okay, so Apple seems to have sewn up that market recently)? The number of blogging software products? The number of biotech firms? The number of private sellers on E-bay? The number of hybrid-drive cars (though most are made by existing manufacturers)? The number of temporary-file-storage sites, a la YouSendIt.com? The number of airlines (I can't remember whether that's going up or down right now; it fluctuates every 5 years or so)? The number of microcredit lenders operating in the third world? The number of firms in China?
What you see is consolidation, the process where all the also-rans in a mature industry drop out or get swallowed up by the market leader. But consolidation applies only to mature industries, where the market is saturated and most consumers are satisfied with the existing level of product quality. All around them, new companies get spun off to take advantage of new markets and new technologies. Sometimes even mature industries get eaten up from the bottom, like how U.S. Steel and Bethlehem Steel got replaced by Nucor and other minimills.
If it looks like the economy is settling down to a steady state, it's because you're looking at the stable portions of the economy. There are areas of the U.S. economy that are growing at 1000% or more (eg. IPods). There are areas where every Tom, Dick, and Jane has started their own business and they all hope to strike it rich (eg. social networking). There are undoubtably areas that could support the next Netscape, Google, or Microsoft, but if I could give you an example I'd be out there working on it instead of writing on Slashdot. ;-)
Steve Jobs's gift has been the ability to spot those areas. If it had just been the Apple 2, I'd chalk it down to luck, but he subsequently got lucky on the Macintosh, Pixar, the IMac, ITunes, the IPod, and now the Mac Mini. And I suspect that his decision to drop out of college had a lot to do with this, as it forced him to go out and search for opportunities. Most people get tunnel vision and only see what's right in front of them. If you want to make money, you've got to see what other people don't.
So don't design a better PC. Let all the other competent people in that industry design one. Put your talents to work in an industry where competence isn't a commodity.
Go design an IDE for Ocaml or Haskell. Write some AJAX apps that do stuff people think can't be done over the web. Build a mobile computing device that lets you browse the web and doesn't have a screen the size of a postage stap. Write a programming language that can adapt and work around its own bugs instead of crashing and burning.
And less basic economics then suggests ways that smaller companies routinely and reliably beat big companies. I recommend The Innovator's Dilemma and
Understand that the free market works for the betterment of the world. It does not work for the betterment of Slashdot readers. There are many products that are very appealing to the SlashDot market (eg. Linux), and they have substantial market share in markets (eg. servers) where Slashdot readers make up a significant proportion. But there are lots of other people out there that have different criteria for "superiority", so the eventual market winner may not reflect your notions of a superior product.