4 ms·

> Sure, you need some knowledge, but you'd better believe other people have that knowledge as well. For any chunk of knowledge, except for very unlikely ties,

by graycat 11y ago

> Sure, you need some knowledge, but you'd better believe other people have that knowledge as well. For any chunk of knowledge, except for very unlikely ties, there was a first person who had that knowledge, and for that person, then, the claim is false. It really is possible to do research that is genuinely new. Indeed, the usual criteria for publication are "new, correct, and significant". Criteria for a Ph.D. dissertation are sometimes "an original contribution to knowledge worthy of publication". Each each paper published in a peer-reviewed journal of original research is supposed to be an example. When the US NSF or NIH give research grants, there is supposed to be research that will be publishable in a peer-reviewed journal of original research, that is, new, knowledge that did not exist before. In computing, we write software. It takes in data, manipulates it, and puts out data. For my startup, I wanted some software that would take in some available data, do some powerful manipulations, and put out some valuable data. So to get some powerful manipulations, difficult to duplicate or equal, and to know right away that the manipulations were correct and likely powerful, I wrote out some original applied math. I have to doubt that anyone else in the world has written out anything really the same. Then I wrote software to perform the data manipulations specified by the math. The software and the math are my proprietary intellectual property, and I intend to use trade secret protection. Tell the world the details? Heck no. Am I able to do such original math derivations, "original contribution to knowledge", "new, correct, significant"? Yup, as demonstrated by my Ph.D. dissertation and published, peer-reviewed papers.

lordnacho 11y ago
>>For any chunk of knowledge, except for very unlikely ties, there was a first person who had that knowledge, and for that person, then, the claim is false. For practical purposes, if you want to make anything resembling a business out of your knowledge, you are going to need time. And chances are people in the field would have been aware of what interesting things there are in that field, and would have some idea of what you've come up with. They may not have some very specific things like particular implementations, but they will nonetheless have a good idea of what's up and down in the field. Obviously you are the only person writing your code, but should you assume that someone with a similar background could not come up with something quite similar if he wanted to research it?
- graycat 11y ago
  > And chances are people in the field would have been aware of what interesting things there are in that field, and would have some idea of what you've come up with. So, do some research that others are not likely to do. Then carry that research all the way to a successful business. My research is some math, based on theorems and proofs, based on some advanced prerequisites. Sure, at least in principle, there are some university mathematicians who could duplicate what I derived. It's been a while since I was a teenager. Since then I've spent a lot of time around low end business, and high end research mathematics, business, startups, and computing. Sure, a lot of that experience has just junk quality information, but the experience does let me have some understanding of the people and cultures. Based on that understanding, there are at least three bottlenecks that help protect me: (1) To duplicate my research, the university research mathematicians would mostly need a relatively clean formulation of the the math part of the real business problem as a math problem. That formulation requires also understanding some of my business model, which is not mathematics, in particular, what data to get and how to get it and, then, what the heck data to give the users/customers that they will like, a lot, and why. E.g., first cut a lot of people will say that I will need ballpark 10^27 bytes of data. Nope! Just why not is a bit clever; although the basic reason is some solid math, it's a bit novel; also need to understand and keep in mind the business model; only a small fraction of research mathematicians have all the math prerequisites; and, net, only a small fraction of people in pure math would think of it. Some of what I'm doing most people who are close to such things will say I can't do that. For things close to what I'm doing, they are right. For exactly what I'm doing, they are wrong. I found a narrow path through the woods that works for just what I'm doing. If my path were broad instead of narrow, a lot of books would have to be revised. (2) The research I did is relatively applied, i.e., would be regarded by the research mathematics community as a niche; then such mathematicians would have to do the work of focusing on the niche; that would take them away from their main stream of research, get them few or no academic promotion points, and is something they wouldn't want to do. (3) Even among good university research mathematicians, some of the crucial math prerequisites are not well understood. For a human, there is little so obscure as some math theorems and proofs where the human is missing the important prerequisites. Outside of university research math departments, the chances of anyone even being able to duplicate my work is much smaller. Then there is the issue of business: The university research mathematicians are a very long way culturally from a startup in information technology; they don't have nearly enough background in the basic computing; e.g., they don't really know in any good detail how a Web server works, or TCP/IP, know an appropriate programming language, understand server farm system management, etc. Then in business, they don't understand information technology startups. E.g., they haven't been reading Hacker News, VentureBeat, AVC.com, Mark Andreessen, etc. They couldn't, and even if they could likely wouldn't, do the work alone (too big of a distraction from their university research career) and with just the math wouldn't be able to attract a team without some significant bucks. Further, for a VC, a professor with a math paper and a dime won't cover a 10 cent cup of coffee. My effort is yet another example of something classic in innovation -- field crossing. So, for the fields, there are some advanced math prerequisites, some understanding of some niche topics in pure and applied math, the particular, niche math research I did, the computing, and, to tie it all together, the whole business model I have in mind. So, early on, no one will bother trying to compete with me. Later I will have some barriers to entry from the data I've gathered, some viral and network effects, some happy users who won't t want to try something else or pay switching costs, and will have some publicity, and a brand name. Or, until I'm quite successful, people who could do the math couldn't easily and wouldn't do the rest; the people who could and would do the rest won't do the math; and no one with a suitable checkbook would cough up enough money to put a team together that could do all the required field crossing. Later? Mostly too late unless someone wants to buy me out and I'm willing to sell. > Obviously you are the only person writing your code, but should you assume that someone with a similar background could not come up with something quite similar if he wanted to research it? The "code" is not the point. The crucial, core point is the math. Given the math, nearly all the code is quite routine, and the rest is not much more difficult. The real, core technical challenge of the code is the math in the design and specifications document for the code. Saying that I did some math research doesn't give away much; it just puts one in the QA section of a research library with racks and racks of books and journals where one can easily get totally lost for many lifetimes. And, for the rest of the news, my work is very likely not in the that library at all. Can I do original math research? Done it often enough in the past!

by graycat 11y ago

lordnacho 11y ago

>>For any chunk of knowledge, except for very unlikely ties, there was a first person who had that knowledge, and for that person, then, the claim is false. For practical purposes, if you want to make anything resembling a business out of your knowledge, you are going to need time. And chances are people in the field would have been aware of what interesting things there are in that field, and would have some idea of what you've come up with. They may not have some very specific things like particular implementations, but they will nonetheless have a good idea of what's up and down in the field. Obviously you are the only person writing your code, but should you assume that someone with a similar background could not come up with something quite similar if he wanted to research it?

graycat 11y ago
> And chances are people in the field would have been aware of what interesting things there are in that field, and would have some idea of what you've come up with. So, do some research that others are not likely to do. Then carry that research all the way to a successful business. My research is some math, based on theorems and proofs, based on some advanced prerequisites. Sure, at least in principle, there are some university mathematicians who could duplicate what I derived. It's been a while since I was a teenager. Since then I've spent a lot of time around low end business, and high end research mathematics, business, startups, and computing. Sure, a lot of that experience has just junk quality information, but the experience does let me have some understanding of the people and cultures. Based on that understanding, there are at least three bottlenecks that help protect me: (1) To duplicate my research, the university research mathematicians would mostly need a relatively clean formulation of the the math part of the real business problem as a math problem. That formulation requires also understanding some of my business model, which is not mathematics, in particular, what data to get and how to get it and, then, what the heck data to give the users/customers that they will like, a lot, and why. E.g., first cut a lot of people will say that I will need ballpark 10^27 bytes of data. Nope! Just why not is a bit clever; although the basic reason is some solid math, it's a bit novel; also need to understand and keep in mind the business model; only a small fraction of research mathematicians have all the math prerequisites; and, net, only a small fraction of people in pure math would think of it. Some of what I'm doing most people who are close to such things will say I can't do that. For things close to what I'm doing, they are right. For exactly what I'm doing, they are wrong. I found a narrow path through the woods that works for just what I'm doing. If my path were broad instead of narrow, a lot of books would have to be revised. (2) The research I did is relatively applied, i.e., would be regarded by the research mathematics community as a niche; then such mathematicians would have to do the work of focusing on the niche; that would take them away from their main stream of research, get them few or no academic promotion points, and is something they wouldn't want to do. (3) Even among good university research mathematicians, some of the crucial math prerequisites are not well understood. For a human, there is little so obscure as some math theorems and proofs where the human is missing the important prerequisites. Outside of university research math departments, the chances of anyone even being able to duplicate my work is much smaller. Then there is the issue of business: The university research mathematicians are a very long way culturally from a startup in information technology; they don't have nearly enough background in the basic computing; e.g., they don't really know in any good detail how a Web server works, or TCP/IP, know an appropriate programming language, understand server farm system management, etc. Then in business, they don't understand information technology startups. E.g., they haven't been reading Hacker News, VentureBeat, AVC.com, Mark Andreessen, etc. They couldn't, and even if they could likely wouldn't, do the work alone (too big of a distraction from their university research career) and with just the math wouldn't be able to attract a team without some significant bucks. Further, for a VC, a professor with a math paper and a dime won't cover a 10 cent cup of coffee. My effort is yet another example of something classic in innovation -- field crossing. So, for the fields, there are some advanced math prerequisites, some understanding of some niche topics in pure and applied math, the particular, niche math research I did, the computing, and, to tie it all together, the whole business model I have in mind. So, early on, no one will bother trying to compete with me. Later I will have some barriers to entry from the data I've gathered, some viral and network effects, some happy users who won't t want to try something else or pay switching costs, and will have some publicity, and a brand name. Or, until I'm quite successful, people who could do the math couldn't easily and wouldn't do the rest; the people who could and would do the rest won't do the math; and no one with a suitable checkbook would cough up enough money to put a team together that could do all the required field crossing. Later? Mostly too late unless someone wants to buy me out and I'm willing to sell. > Obviously you are the only person writing your code, but should you assume that someone with a similar background could not come up with something quite similar if he wanted to research it? The "code" is not the point. The crucial, core point is the math. Given the math, nearly all the code is quite routine, and the rest is not much more difficult. The real, core technical challenge of the code is the math in the design and specifications document for the code. Saying that I did some math research doesn't give away much; it just puts one in the QA section of a research library with racks and racks of books and journals where one can easily get totally lost for many lifetimes. And, for the rest of the news, my work is very likely not in the that library at all. Can I do original math research? Done it often enough in the past!