3 ms·

Clearly since Ms. Livingston was graduated from college, she has learned a lot about the economy, business, start-ups, making money, getting financial security

by graycat 12y ago

Clearly since Ms. Livingston was graduated from college, she has learned a lot about the economy, business, start-ups, making money, getting financial security, careers, and jobs. And likely and apparently she got a good marriage. Good. History But nearly unbroken history of at least the past 200 years shows that the knowledge frontier she reached keeps moving forward. So, yes, likely Ms. Livingston knows more about start-ups, etc. than her mother did, but, even so, it is not clear that she knows enough about start-ups people in their teens and 20s need to know now to do well to get started on doing well in the rest of their careers. Do well? Sure: Have to select the good start-ups and f'get about the rest. How to do that? Not easily, e.g., the venture capital world tries but, from some data posted by VC Fred Wilson on his blog AVC.com (1) about 2/3rds of the start-ups he funded flop (don't do well, or some such) and (2) on average the return on investment (ROI) of US information technology start-ups is poor, less good than Ms. Livingston helped obtain in her first job in Boston that counted the minutes she was on her coffee breaks. Sure, Ms. Livingston might have majored in computer science and joined Microsoft in 1990. Alas, we come this way only once, and we don't get do-overs or re-dos, at least not until we get a time machine. So, what to tell young people now? Since Ms. Livingston is talking about start-ups mostly in information technology, we should consider such things, and there is less to the significant history there than often meets the eye: (1) In the 1930s, Bell saw clearly that vacuum tubes were too big, expensive, unreliable, and hot for the future of the US long distance phone network. So, a better amplifier was needed. They had a solid state rectifier, so what about a solid state amplifier? They started a project. WW II got in the way, but soon after the war, presto, bingo, fireworks in the sky, a giant step for Bell and a much larger step for mankind -- the transistor. Exercise: Why so important? At the time Bell saw the importance for civilization and, thus, decided not to patent the transistor. So, write an essay that might have been written in, say, 1949, on why the transistor should be so important. (2) Soon enough the Cold War caused the US DoD to want a lot of digital electronics for aerospace, and thus, erupted from Stanford, due in part to Dean Terman, Silicon Valley to supply electronics to the US DoD and later NASA. (3) Fairly soon Silicon Valley saw how to put several transistors on one piece of silicon and make integrated circuits. Gee, could make little devices! An electronic calculator to replace all those mechanical parts? Sure: But, still, need a lot of transistors, maybe more than for just a small, simple, general purpose computer where the rest of the calculator functions are just from software? Yup. So, bingo, simple microprocessors. (4) NSF: MIT had some ideas for interactive computers with a lot of security. So NSF funded their Project MAC which did the operating system Multics with an hierarchical file system and security features capabilities and access control lists. For authentication, MIT did Kerberos. Then Kerberos made use of RSA encryption, also from MIT. Big things still with us; moving right along here. (5) Bell Labs again: They wanted word whacking. DEC had a mini-computer, so Bell borrowed a little from Multics, etc., wrote a simple operating system Unix written in a simple programming language C. Later Bjarne Stroustrup, also at Bell Labs, wrote a pre-processor for C to support software objects -- the pre-processor was called C++. Since Bell was a regulated monopoly, they couldn't sell Unix so essentially gave it away. A group at Berkeley, as I recall funded by the US DoE, did more with UNIX and made their work available as the Berkeley Software Distribution (BSD). Unix became Linux, and C++ and software objects are still with us. (6) Lots of people, not just Bell, were struggling with typing. IBM had their Selectric and, eventually, a correcting Selectric with a little white ribbon that would remove from the page a character struck in error. But, why no actual word whacking? Okay, Apple II, IBM PC, WordStar, etc. Biggie. Really big biggie. (7) We got Microsoft that recapitulated the mainframe history in operating systems, and Intel came along with microprocessors that recapitulated the mainframe history in processors. Now WinTel put on "every desktop" a computer for word whacking, Microsoft Word, and business arithmetic, Excel. Gates on the way to being the richest person in the world. Not too doing too badly, Paul Allen, Charlie Simoni, Nathan Myhrvold, etc. (8) Ah, the US DoD again: It wanted battlefield communications, where even if shoot holes in some of the equipment the rest still works and provides communications. So, we got TCP/IP, e.g., in BSD, that is, we got internets. Soon labs were connected, and we got the Internet. Soon NSF funded it and IBM ran it. With HTTP and HTML for a particle physics newsletter by Tim Berners-Lee at CERN, we got the Web. Companies put their company brochures on the Internet. (9) For the Internet, we needed more in communications capacity. Enter Bell labs again: They'd seen that one coming, too, and had been working on Ga-Al-As (as in the periodic table from a chemistry book) solid state heterojunction lasers and had the solution. Bingo: Send at 40+ billion bits per second (Gbps) on one wavelength, some dozens of wavelength on one fiber, some dozens of fibers in one cable, maybe several cables along a pipeline, electric power line, railroad track, highway, river, ocean shoreline, across an ocean or few, etc. Now, watch movies! (10) Presto, IBM and others learned more about putting magnetic dots on surfaces, and now we have hard disk drives in the 3 1/2" form factor size with a few trillion bytes each. And HP is on the way with a trillion bytes, solid state, on a postage stamp. So, now we can build server farms at Google, Facebook, Twitter, Microsoft, Apple, Amazon, etc., and we can have start-ups like SnapChat, PInterest, Box, etc.

graycat 12y ago
Future So, what lessons for the future might a young person draw from this history? (1) Science. Off and on, some amazing science, especially physics, played a huge role. Maybe that will continue. (2) Information. The desire for information, create it, transmit it, store it, use it, etc., seems nearly unlimited. (3) Logic. Want something done? Well, describe the work in clear steps. For a lot of work done manually in offices over the past 100 years, such a description is now usually fairly routine. Then with such a description, fairly routinely can write software to do the work. So, can automate a huge range of old, manual work of office workers. That's a lot of what for some decades made IBM successful. (4) Social. People are highly social animals. Or to paraphrase E. Fromm, The Art of Loving, "For humans, the fundamental problem in life is doing something effective about feeling alone." In more detail, since humans are also thinking animals, we see that alone we are at risk, that is, vulnerable to, say, the hostile forces of nature (earthquakes, blizzards, tornadoes, floods, wild fires, disease) and society (war, crime, economic depression). Knowing that we are vulnerable, we are worried (have anxiety) and seek security. We feel more vulnerable when alone so want to do something about being alone. From Fromm again, the first recommended solution is a good romantic relationship. Next is a good version of religion -- get all wrapped up. Next is membership in a good group -- get acceptance and approval, a feeling of belonging. Next, not recommended, is what some college students try -- get drunk on alcohol, high on drugs, and go to an orgy. So forget about the worries until recover (but have more worries). So, to do something about the worries, we want security, financial and emotional, don't want to be lonely, do want to be loved, want a romantic relationship, want to belong, etc. More generally we will want to form good families and be in good communities. We will be using computing and the Internet for all they are worth for such things. (5) Economic Security. Likely second only to love, and maybe more important than love, and maybe essentially a prerequisite to love, people want economic security, and for that there is a famous one word answer "more". The drive to use logic, software, computing, the Internet, etc. for "more" will remain powerful for decades, maybe centuries. (6) Information. Now one of the keys to more in economic security, "more", is information, and the drive for that will also continue for decades, etc. For information, we take in available data, process it, and report the resulting information. This processing is necessarily mathematically something, understood or not, powerful or not. Then clearly one approach to more powerful processing and, thus, more powerful and valuable information, is to use mathematics to determine how to do the processing. E.g., how to look for oil? Okay, often oil collects in pockets in the subsurface layers. So, let's map the layers and look for pockets. How to do that? On the surface, have something go "boom". Sound waves go into the ground, and they get reflected off the layers so that there is a convolution. So, to find the layers, take the resulting signal and do a deconvolution -- Enders A. Robinson, 'Multichannel Time Series Analysis with Digital Computer Programs'. The fast way to do deconvolution? Sure, the fast Fourier transform. Once get the oil out, over here have all that oil, from Texas, the Mideast, Venezuela, Canada, etc. -- typically it's all different. Over there know what can sell -- methane, propane, gasoline, Diesel, heating oil, motor oil, etc. So, how to take the available input and sell the output and make the most money? That's a math problem, in particular in optimization. Long the first-cut approach was via linear optimization (programming in the sense of operational planning). At one time, IBM had fun selling mainframes to Houston for just this work. But linear programming is not quite the right stuff. So, want some non-linear optimization. Well, for more details, see the work of Christodoulos A. Floudas in chemical engineering at Princeton. Houston does know about Professor Floudas. There's much more to do. Right: Likely not a single VC in the country says that they want to see some especially valuable software based on some especially powerful mathematics. Hardly a one. And they are not comfortable backing something they understand so poorly. So, right, a lot of confused and unhappy VCs (they so richly deserve it!) but: Presto, bingo, opportunity. Besides, the main raw material into original mathematics is paper, pencils, and coffee, and how expensive are those? Almost inevitably, there will be only a few people going that way with the rest heaping ridicule, etc. Not nearly new: Think of the Mother Goose story The Little Red Hen. Secret: It turns out, no matter how much advanced and/or original mathematics you use, nearly always a lot of the actual computations will boil down to linear algebra and there, numerical linear algebra. So, take linear algebra, elementary, intermediate, advanced, applied, numerical, and related subjects such as linear programming, non-linear programming, multi-variate statistics, ordinary and partial differential equations and their numerical solutions. For more, study the leading generalizations of linear algebra, functional analysis, e.g., Hilbert and Banach spaces. (7) Niches. One of the standard ways to make money is to have close to a monopoly, and one of the standard ways to do that is to have a niche of some kind and where the monopoly is protected by, say, a geographical barrier to entry, an especially good product or service, some crucial, core, defensible technology or know-how, a good customer list, some network effect, etc. So, go for it!

by graycat 12y ago

graycat 12y ago

Future So, what lessons for the future might a young person draw from this history? (1) Science. Off and on, some amazing science, especially physics, played a huge role. Maybe that will continue. (2) Information. The desire for information, create it, transmit it, store it, use it, etc., seems nearly unlimited. (3) Logic. Want something done? Well, describe the work in clear steps. For a lot of work done manually in offices over the past 100 years, such a description is now usually fairly routine. Then with such a description, fairly routinely can write software to do the work. So, can automate a huge range of old, manual work of office workers. That's a lot of what for some decades made IBM successful. (4) Social. People are highly social animals. Or to paraphrase E. Fromm, The Art of Loving, "For humans, the fundamental problem in life is doing something effective about feeling alone." In more detail, since humans are also thinking animals, we see that alone we are at risk, that is, vulnerable to, say, the hostile forces of nature (earthquakes, blizzards, tornadoes, floods, wild fires, disease) and society (war, crime, economic depression). Knowing that we are vulnerable, we are worried (have anxiety) and seek security. We feel more vulnerable when alone so want to do something about being alone. From Fromm again, the first recommended solution is a good romantic relationship. Next is a good version of religion -- get all wrapped up. Next is membership in a good group -- get acceptance and approval, a feeling of belonging. Next, not recommended, is what some college students try -- get drunk on alcohol, high on drugs, and go to an orgy. So forget about the worries until recover (but have more worries). So, to do something about the worries, we want security, financial and emotional, don't want to be lonely, do want to be loved, want a romantic relationship, want to belong, etc. More generally we will want to form good families and be in good communities. We will be using computing and the Internet for all they are worth for such things. (5) Economic Security. Likely second only to love, and maybe more important than love, and maybe essentially a prerequisite to love, people want economic security, and for that there is a famous one word answer "more". The drive to use logic, software, computing, the Internet, etc. for "more" will remain powerful for decades, maybe centuries. (6) Information. Now one of the keys to more in economic security, "more", is information, and the drive for that will also continue for decades, etc. For information, we take in available data, process it, and report the resulting information. This processing is necessarily mathematically something, understood or not, powerful or not. Then clearly one approach to more powerful processing and, thus, more powerful and valuable information, is to use mathematics to determine how to do the processing. E.g., how to look for oil? Okay, often oil collects in pockets in the subsurface layers. So, let's map the layers and look for pockets. How to do that? On the surface, have something go "boom". Sound waves go into the ground, and they get reflected off the layers so that there is a convolution. So, to find the layers, take the resulting signal and do a deconvolution -- Enders A. Robinson, 'Multichannel Time Series Analysis with Digital Computer Programs'. The fast way to do deconvolution? Sure, the fast Fourier transform. Once get the oil out, over here have all that oil, from Texas, the Mideast, Venezuela, Canada, etc. -- typically it's all different. Over there know what can sell -- methane, propane, gasoline, Diesel, heating oil, motor oil, etc. So, how to take the available input and sell the output and make the most money? That's a math problem, in particular in optimization. Long the first-cut approach was via linear optimization (programming in the sense of operational planning). At one time, IBM had fun selling mainframes to Houston for just this work. But linear programming is not quite the right stuff. So, want some non-linear optimization. Well, for more details, see the work of Christodoulos A. Floudas in chemical engineering at Princeton. Houston does know about Professor Floudas. There's much more to do. Right: Likely not a single VC in the country says that they want to see some especially valuable software based on some especially powerful mathematics. Hardly a one. And they are not comfortable backing something they understand so poorly. So, right, a lot of confused and unhappy VCs (they so richly deserve it!) but: Presto, bingo, opportunity. Besides, the main raw material into original mathematics is paper, pencils, and coffee, and how expensive are those? Almost inevitably, there will be only a few people going that way with the rest heaping ridicule, etc. Not nearly new: Think of the Mother Goose story The Little Red Hen. Secret: It turns out, no matter how much advanced and/or original mathematics you use, nearly always a lot of the actual computations will boil down to linear algebra and there, numerical linear algebra. So, take linear algebra, elementary, intermediate, advanced, applied, numerical, and related subjects such as linear programming, non-linear programming, multi-variate statistics, ordinary and partial differential equations and their numerical solutions. For more, study the leading generalizations of linear algebra, functional analysis, e.g., Hilbert and Banach spaces. (7) Niches. One of the standard ways to make money is to have close to a monopoly, and one of the standard ways to do that is to have a niche of some kind and where the monopoly is protected by, say, a geographical barrier to entry, an especially good product or service, some crucial, core, defensible technology or know-how, a good customer list, some network effect, etc. So, go for it!