3 ms·

"If you truly have a good idea, somebody else is working on it too." Standard HN cliche number 298,498,229 and nonsense. I've had several ideas that I can ess

"If you truly have a good idea, somebody else is working on it too." Standard HN cliche number 298,498,229 and nonsense. I've had several ideas that I can essentially 100% guarantee that no one else was working on. My Ph.D. dissertation was officially judged "new, correct, and significant" -- note the "new". Similarly for each paper of original research I've published. Similarly for the crucial, core idea for the company I'm starting. Just what is it about original research that HN just cannot understand?

guptaneil 16y ago
Research is a different beast than creating a marketable product. If you're creating a product that nobody is working on or hasn't tried to make before, you have to stop and wonder, why has nobody done this before? Before Tylenol existed, many people had attempted to tackle the problem of pain management. Tylenol just happened to make a more efficient chemical formula with their research, but the idea of using drugs to control pain was not new. When you talk about your idea, you don't need to reveal how you do something differently, just like Tylenol can talk about their idea without revealing their formula and Google can talk about their idea without revealing their algorithm. I was in St Louis last week for a business pitch competition, and I watched a med school PhD candidate read his dissertation paper and present his research for the elevator pitch competition. He had absolutely no business knowledge or any idea on how to market his research into a profitable product. But because he was able to talk about his idea and research without revealing anything that would allow somebody else in the room to just copy it, he was able to build contacts with several other business-oriented people in the room who liked the idea and were interested in helping him take it to market.
- NY_USA_Hacker 16y ago
  "Research is a different beast than creating a marketable product." Well there are people in research looking for fundamental knowledge, e.g., string theory in physics or the finest details of biochemistry and genetic material in biology. And there are people doing make-work, junk-think, prof-scam, busy work. And in neither case is there much connection with a product or market. But your claim is not true in all cases! I'm creating a "marketable product". That the product is desirable and "marketable" is mostly obvious. The challenge is how to create the thing. I found a way. That way was some original research in applied math. Since I have Ph.D. for such math, I'm able to make a competent attempt at such research and to judge any results for correctness as math and significance for the real problem. There's essentially no doubt that my idea is "new" and that no one else is working on it. Reasons: First, my work has some advanced prerequisites nearly unknown anywhere among full profs in computer science. They didn't take enough of the right grad math courses, and, actually, neither did more than a tiny fraction of math profs. Some of the courses are rarely taught; the combination of courses is especially rare. Second, the crucial part of the work is clearly original; there's no chance it's in the books or journals, and there's essentially no chance that anyone else with the prerequisites would be trying to get the same results. To me, that research I did was the crucial 'idea' for my business. At this point, the rest is routine, a simple Web site, some simple usage of relational database, some fairly simple initial data gathering, and some routine publicity. You are insisting on saying that the research is not business and not the idea and that the idea is the market for a solution for the customer pain. That definition of an 'idea' is new and strange and a big break with centuries of history in technology for business, medicine, and the military and with the whole history of the USPTO. Yes, your new definition of an 'idea' has become very common in Web 2.0 information technology entrepreneurship. The people who invented the formula for Tylenol would agree that that invention of the chemical formula, and possibly how to do the manufacturing, was the crucial 'idea'. Just a better pill to alleviate pain was not the idea. For the market I'm addressing, sure, lots of other people have tried. It's just that for the customer pain I'm addressing, the results from the other guys just suck and the results from my idea, as is quite clear from some of the math, should be much better. Eventually it will dawn on some people in computer science that I must have made good progress in having computer software work effectively with how humans view 'meaning', where they will be correct, and start to say my work is close to earth shaking. But I don't want their characterizations, praise, prizes, approval, admiration, gratitude, tenure, invitations to give lectures, etc.; I want a yacht. Then I want to compose music, starting roughly with the techniques of Richard Strauss, and then pursue mathematical physics, often on my yacht. Yes, other people have tried to solve the same customer pain, but I do not conclude from this fact that they have been working on the same "idea" I am. One pill taken once to cure any cancer is not a significant 'idea' or 'market'. The good idea would be how to make such a pill, and then the 'market' would be quite new. I count aspirin, Tylenol, Advil, Benzocaine, etc. as great ideas for alleviating pain. I don't count the market for alleviating pain as a meaningful 'idea', but you seem to be. Your new definition for an 'idea' actually can have some value in Web 2.0: It is not very clear just what people want from Web 2.0, the 'social graph', etc. so that a guess at what they want and how to provide it, even if all the necessary technology is trivial, might be a valuable 'idea'. There seems to be some progress here at http://www.slideshare.net/padday/the-real-life-social-network-v2 http://www.slideshare.net/padday/the-real-life-social-networ... Still, this definition of an 'idea' is strange, new, very narrow, and in strong conflict with centuries of important history and nearly everything else in 'ideas' today from 'creative' in ads to the USPTO, DoD, NSF, and NIH funded projects, pharmaceutical research, and much more. I have one more idea: Send Jane Fonda and Al Guru to Japan!
  - guptaneil 16y ago
    I think we are arguing the same point but with different terms. To me, the 'idea' is how you would describe the product to a customer in one sentence. The execution is how this idea is made. In your case, this would be your research and unique math algorithm. The business plan is how you plan to market and monetize this product. A successful business requires all three factors. When I said that an idea is being worked on elsewhere, I meant that somebody has at some point tried to solve that pain point for your customers. They are your competition. I stand by this statement. You seem to be arguing that your execution is unique, and I absolutely agree. Your execution has to be unique, or else you would be guaranteed to fail. Based on your description, it seems you are in an ideal position to execute your idea successfully, and I wish you the best of luck. Based on the above definitions, I am suggesting that you can and should talk about your idea without talking about your execution. Share what problem you're solving, but definitely don't tell anybody the results of your math research. Using your example of a cancer pill, you should talk about your pill to anybody who will listen without sharing how the pill is going cure cancer. I disagree that business plan factor are trivial after execution is handled. You could make the perfect product with your algorithm, but if you are unable to figure out a way to monetize or market your product, then it will fail. Of course, the relative difficulty of each factor depends on the industry and product.
    - NY_USA_Hacker 16y ago
      "To me, the 'idea' is how you would describe the product to a customer in one sentence." Yes, that seems to be a better definition of the 'new', Web 2.0 definition of an 'idea' than I was able to formulate. I just don't think it is a very appropriate, significant, or meaningful use of 'idea'. I regard the important ideas as the crucial, difficult, powerful 'secret sauce' and the 'execution' as the routine stuff that comes later, especially after customers arrive. With a sufficiently good idea, say, the first Xerox machine, TWA in its first months as it had the only Lockheed Connies in the world, and more, as relatively 'routine' and this because the work before was so darned good. The US is absolutely awash in ability to do routine execution, from a corner store up to a giant corporation, well. For my project, I've worked out more than the math: The math is now in code ready for production -- tested, timed, documented. It's shockingly fast. I've worked out a server farm 'architecture' that is dirt simple and looks like it will 'scale' beyond what the world could use. Much of the simplicity is because there is a lot of parallelism and redundancy and a lot of data that is read-only for at least a day and maybe a week. Actually, all the big data is read-only in this sense. And, it does appear that in production this big read-only data can be just in virtual memory off either rotating magnetic disks or the newer, faster solid state disks (that love read-only data). Windows Server 2008 (I'm staying with Microsoft) can have 16 TB of virtual memory, and my software does the accesses with very nice, and, thus, fast, locality of reference. The role of the math is to take the available data, 'crunch' it, and give the desired results. The quality of the results is key and where the work of the other guys sucks. There is a 'larger business plan' beyond just the crucial, core math: As a byproduct of the work itself, the Web site gets some data from the users. The users give that data as part of getting the results they want and won't mind giving the data because there is no privacy threat. But from some subtle results in advanced math (the Radon-Nikodym theorem), that data will be the main source of 'base data' for the whole system for everything. So, in particular, compared with the other guys, I have some new data, of unusually high quality, and especially valuable. People won't see why the data is valuable, but it is! In particular, the data will let me do some good ad targeting, and that's where the 'monitization' and 'revenue model' come from. I don't really want to use third party cookies to connect with the ad targeting of the ad networks, and I should be able to do better ad targeting just with my own data and some more of my math. The software promises to be astoundingly efficient: From the timings, data rates, and data sizes, it appears that for the on-line production servers, four racks should be enough for about $4.5 billion a year in revenue, and 2000 square feet of on-line rack space should be sufficient to serve the world. Right: If 4 racks can give $4.5 billion a year in revenue, then 2000 square feet of rack space should give one of the most valuable companies in the world. Of course would want more space for redundancy, backup and recovery, testing, and development. For more reliability and lower latency, might want more than one data center. Could make a little use of some caching on the network. I'm not eager to say what the problem is much before I'm ready to go live. Actually, telling people what the problem is seems to have no upside at all: Plenty of people in venture capital have had descriptions of the work polished to more luster than any gold and really don't give a hoot. The situation is very much as in, "before good ComScore numbers, the project is just an idea, and ideas are easy and plentiful. Execution is everything and doesn't start until after the ComScrore numbers." The flip side of this thinking is an opportunity for me. Net, what they care about are ComScore numbers. But, as threatens to be increasingly the case, for a successful Web site, there will be a window narrow in time between when the ComScore numbers are (1) good enough for the VCs to invest and (2) so good that I won't need or take their check. The advantages of being 100% owner of a 100% private, very profitable Subchapter S corporation, or even a C corporation, are just huge. Actually, I'm terrified of having a Board: (1) They could kill my company just from taking up 25% of my time developing information to report to them. (2) There is no way they could possibly understand anything significant about the future of the business and no way I could explain it to them or keep them happy about it. For the 'algorithms', I would say that they are trivial because the real content is in the math; the only claim of power or quality from the code is the math it implements; actually, given the math, the implementation was not difficult. So what matters is the math, not the algorithms. This is a general situation in newer computing: An algorithm standing alone has little to recommend it, and just 'experience' will be a long, slow path to 'validity'. Instead, for 'validity', quality, whatever, an algorithm needs a prior 'specification' of what it does. Often that specification will have to be some relatively advanced math. Currently this math is a huge bottleneck for promise in computing.

guptaneil 16y ago

Research is a different beast than creating a marketable product. If you're creating a product that nobody is working on or hasn't tried to make before, you have to stop and wonder, why has nobody done this before? Before Tylenol existed, many people had attempted to tackle the problem of pain management. Tylenol just happened to make a more efficient chemical formula with their research, but the idea of using drugs to control pain was not new. When you talk about your idea, you don't need to reveal how you do something differently, just like Tylenol can talk about their idea without revealing their formula and Google can talk about their idea without revealing their algorithm. I was in St Louis last week for a business pitch competition, and I watched a med school PhD candidate read his dissertation paper and present his research for the elevator pitch competition. He had absolutely no business knowledge or any idea on how to market his research into a profitable product. But because he was able to talk about his idea and research without revealing anything that would allow somebody else in the room to just copy it, he was able to build contacts with several other business-oriented people in the room who liked the idea and were interested in helping him take it to market.

NY_USA_Hacker 16y ago
"Research is a different beast than creating a marketable product." Well there are people in research looking for fundamental knowledge, e.g., string theory in physics or the finest details of biochemistry and genetic material in biology. And there are people doing make-work, junk-think, prof-scam, busy work. And in neither case is there much connection with a product or market. But your claim is not true in all cases! I'm creating a "marketable product". That the product is desirable and "marketable" is mostly obvious. The challenge is how to create the thing. I found a way. That way was some original research in applied math. Since I have Ph.D. for such math, I'm able to make a competent attempt at such research and to judge any results for correctness as math and significance for the real problem. There's essentially no doubt that my idea is "new" and that no one else is working on it. Reasons: First, my work has some advanced prerequisites nearly unknown anywhere among full profs in computer science. They didn't take enough of the right grad math courses, and, actually, neither did more than a tiny fraction of math profs. Some of the courses are rarely taught; the combination of courses is especially rare. Second, the crucial part of the work is clearly original; there's no chance it's in the books or journals, and there's essentially no chance that anyone else with the prerequisites would be trying to get the same results. To me, that research I did was the crucial 'idea' for my business. At this point, the rest is routine, a simple Web site, some simple usage of relational database, some fairly simple initial data gathering, and some routine publicity. You are insisting on saying that the research is not business and not the idea and that the idea is the market for a solution for the customer pain. That definition of an 'idea' is new and strange and a big break with centuries of history in technology for business, medicine, and the military and with the whole history of the USPTO. Yes, your new definition of an 'idea' has become very common in Web 2.0 information technology entrepreneurship. The people who invented the formula for Tylenol would agree that that invention of the chemical formula, and possibly how to do the manufacturing, was the crucial 'idea'. Just a better pill to alleviate pain was not the idea. For the market I'm addressing, sure, lots of other people have tried. It's just that for the customer pain I'm addressing, the results from the other guys just suck and the results from my idea, as is quite clear from some of the math, should be much better. Eventually it will dawn on some people in computer science that I must have made good progress in having computer software work effectively with how humans view 'meaning', where they will be correct, and start to say my work is close to earth shaking. But I don't want their characterizations, praise, prizes, approval, admiration, gratitude, tenure, invitations to give lectures, etc.; I want a yacht. Then I want to compose music, starting roughly with the techniques of Richard Strauss, and then pursue mathematical physics, often on my yacht. Yes, other people have tried to solve the same customer pain, but I do not conclude from this fact that they have been working on the same "idea" I am. One pill taken once to cure any cancer is not a significant 'idea' or 'market'. The good idea would be how to make such a pill, and then the 'market' would be quite new. I count aspirin, Tylenol, Advil, Benzocaine, etc. as great ideas for alleviating pain. I don't count the market for alleviating pain as a meaningful 'idea', but you seem to be. Your new definition for an 'idea' actually can have some value in Web 2.0: It is not very clear just what people want from Web 2.0, the 'social graph', etc. so that a guess at what they want and how to provide it, even if all the necessary technology is trivial, might be a valuable 'idea'. There seems to be some progress here at http://www.slideshare.net/padday/the-real-life-social-network-v2 http://www.slideshare.net/padday/the-real-life-social-networ... Still, this definition of an 'idea' is strange, new, very narrow, and in strong conflict with centuries of important history and nearly everything else in 'ideas' today from 'creative' in ads to the USPTO, DoD, NSF, and NIH funded projects, pharmaceutical research, and much more. I have one more idea: Send Jane Fonda and Al Guru to Japan!
- guptaneil 16y ago
  I think we are arguing the same point but with different terms. To me, the 'idea' is how you would describe the product to a customer in one sentence. The execution is how this idea is made. In your case, this would be your research and unique math algorithm. The business plan is how you plan to market and monetize this product. A successful business requires all three factors. When I said that an idea is being worked on elsewhere, I meant that somebody has at some point tried to solve that pain point for your customers. They are your competition. I stand by this statement. You seem to be arguing that your execution is unique, and I absolutely agree. Your execution has to be unique, or else you would be guaranteed to fail. Based on your description, it seems you are in an ideal position to execute your idea successfully, and I wish you the best of luck. Based on the above definitions, I am suggesting that you can and should talk about your idea without talking about your execution. Share what problem you're solving, but definitely don't tell anybody the results of your math research. Using your example of a cancer pill, you should talk about your pill to anybody who will listen without sharing how the pill is going cure cancer. I disagree that business plan factor are trivial after execution is handled. You could make the perfect product with your algorithm, but if you are unable to figure out a way to monetize or market your product, then it will fail. Of course, the relative difficulty of each factor depends on the industry and product.
  - NY_USA_Hacker 16y ago
    "To me, the 'idea' is how you would describe the product to a customer in one sentence." Yes, that seems to be a better definition of the 'new', Web 2.0 definition of an 'idea' than I was able to formulate. I just don't think it is a very appropriate, significant, or meaningful use of 'idea'. I regard the important ideas as the crucial, difficult, powerful 'secret sauce' and the 'execution' as the routine stuff that comes later, especially after customers arrive. With a sufficiently good idea, say, the first Xerox machine, TWA in its first months as it had the only Lockheed Connies in the world, and more, as relatively 'routine' and this because the work before was so darned good. The US is absolutely awash in ability to do routine execution, from a corner store up to a giant corporation, well. For my project, I've worked out more than the math: The math is now in code ready for production -- tested, timed, documented. It's shockingly fast. I've worked out a server farm 'architecture' that is dirt simple and looks like it will 'scale' beyond what the world could use. Much of the simplicity is because there is a lot of parallelism and redundancy and a lot of data that is read-only for at least a day and maybe a week. Actually, all the big data is read-only in this sense. And, it does appear that in production this big read-only data can be just in virtual memory off either rotating magnetic disks or the newer, faster solid state disks (that love read-only data). Windows Server 2008 (I'm staying with Microsoft) can have 16 TB of virtual memory, and my software does the accesses with very nice, and, thus, fast, locality of reference. The role of the math is to take the available data, 'crunch' it, and give the desired results. The quality of the results is key and where the work of the other guys sucks. There is a 'larger business plan' beyond just the crucial, core math: As a byproduct of the work itself, the Web site gets some data from the users. The users give that data as part of getting the results they want and won't mind giving the data because there is no privacy threat. But from some subtle results in advanced math (the Radon-Nikodym theorem), that data will be the main source of 'base data' for the whole system for everything. So, in particular, compared with the other guys, I have some new data, of unusually high quality, and especially valuable. People won't see why the data is valuable, but it is! In particular, the data will let me do some good ad targeting, and that's where the 'monitization' and 'revenue model' come from. I don't really want to use third party cookies to connect with the ad targeting of the ad networks, and I should be able to do better ad targeting just with my own data and some more of my math. The software promises to be astoundingly efficient: From the timings, data rates, and data sizes, it appears that for the on-line production servers, four racks should be enough for about $4.5 billion a year in revenue, and 2000 square feet of on-line rack space should be sufficient to serve the world. Right: If 4 racks can give $4.5 billion a year in revenue, then 2000 square feet of rack space should give one of the most valuable companies in the world. Of course would want more space for redundancy, backup and recovery, testing, and development. For more reliability and lower latency, might want more than one data center. Could make a little use of some caching on the network. I'm not eager to say what the problem is much before I'm ready to go live. Actually, telling people what the problem is seems to have no upside at all: Plenty of people in venture capital have had descriptions of the work polished to more luster than any gold and really don't give a hoot. The situation is very much as in, "before good ComScore numbers, the project is just an idea, and ideas are easy and plentiful. Execution is everything and doesn't start until after the ComScrore numbers." The flip side of this thinking is an opportunity for me. Net, what they care about are ComScore numbers. But, as threatens to be increasingly the case, for a successful Web site, there will be a window narrow in time between when the ComScore numbers are (1) good enough for the VCs to invest and (2) so good that I won't need or take their check. The advantages of being 100% owner of a 100% private, very profitable Subchapter S corporation, or even a C corporation, are just huge. Actually, I'm terrified of having a Board: (1) They could kill my company just from taking up 25% of my time developing information to report to them. (2) There is no way they could possibly understand anything significant about the future of the business and no way I could explain it to them or keep them happy about it. For the 'algorithms', I would say that they are trivial because the real content is in the math; the only claim of power or quality from the code is the math it implements; actually, given the math, the implementation was not difficult. So what matters is the math, not the algorithms. This is a general situation in newer computing: An algorithm standing alone has little to recommend it, and just 'experience' will be a long, slow path to 'validity'. Instead, for 'validity', quality, whatever, an algorithm needs a prior 'specification' of what it does. Often that specification will have to be some relatively advanced math. Currently this math is a huge bottleneck for promise in computing.