.
Profile URL | https://www.quora.com/profile/Bill-Paseman |
Question | Answer | Date |
---|---|---|
What is the background of axonometric projection, isometric, dimetric and trimetric projection? | Axonometric projection - Wikipedia discusses the history | 11/12/2017 |
What is the best automated trading software using interactive brokers? | This may be too "hands on" for you, but someone recommended Quantopian | Algorithmic Investing | Algorithmic Trading to me. | 12/12/2014 |
What is the best high school project you have made? How did you make it? | I supported my daughters’ projects in non-invasive blood analysis. We first went to a university and asked for titles of work that they were impressed with. (Theory being that this was the primary audience for the projects). They then used that as the basis for further investigation. | 4/27/2017 |
What is the best mathematical book? | I don’t know about best, but by popularity, the most reprinted math book of all time is probably Euclid's Elements - Wikipedia. One abbreviated version is Byrne: Six Books of Euclid: Werner Oechslin: 9783836559386: Amazon.com: Books . This book is“very Taschenbuch”, i.e. if you are color blind, the book loses about 1/2 its explanatory power. The second most reprinted book is Leonhard Euler - Wikipedia’s Elements of Algebra - Wikipedia If you want to do for Computer Science, MIT has an online text. “Mathematics for computer Science” https://courses.csail.mit.edu/6.... If you want to read about it (vs do it) The Mathematical Experience - Wikipedia Key Point: My high school calculus teacher (Maarten Kalisvaart) would often admonish the class with “reading a math book is not like reading a comic book” and punctuated this with his favorite proof quote “This proof is totally obvious to the casual observer after about 3 weeks of hard study”. So it my take days to “understand” a few pages. E.g. I remember going over the proof of d sinx(x)/dx in a calculus text when I was 16. Took days and even after I was able to recite the proof, I -still- did not have the intuitive feeling that the result was “correct” (despite the fact that derivatives of x**n and e**x felt “natural” to me.) I met up with Dr. Kalisvaart for lunch a few years ago and asked him what books he felt provided the most insight into math. He said Euler‘s notebooks. Euler pretty much wrote down his thoughts every day, and so if you want to see the flow, you can read copies yourself. This approach was a lot more “human” than that employed by my favorite mathematician Carl Friedrich Gauss - Wikipedia . He would come up with an idea and rework it until no trace of the original inspiration remained. Kind of like reading a text handed down by an (inhuman) god. Demonstrably correct but pedagogically barren. Another Point: Alan Perlis - Wikipedia once opened up a guest talk at MIT with the line “I assume that there are more people in the room who learned calculus than invented it.” So see if you can invent the math yourself instead of reading about it. E.g. One trick I used in Freshman physics (mechanics) was to come in remembering just one equation, then I would derive all the rest of the stuff I needed myself instead of memorizing the formulas (this didn’t always work for me, see the sin(x) example above). The one equation? F = d(mv)/dt. Seems trivial and needlessly anal, doesn’t it? Check out this post (www.SisterZone.com) and maybe you’ll be convinced of the value of the approach. | 12/1/2017 |
What is the best online bank (fintech) to use if you want to withdraw money around the world without fees and at the best exchange rate? | I have not done a detailed comparison, but if you keep your balance at a high enough level, Cards issued by Schwab investor services in the US work at ATMs in Europe and places I’ve visited in Asia. | 6/20/2016 |
What is the best places or programs to do on the weekend in Silicon Valley? Thinking in entrepreneurship, startups etc. | Checkout Find your people - Meetup | 8/3/2016 |
What is the best poetry book that I should read? | If you want to start by sipping, not drinking deeply, You might try a slim book of Robert Frost’s poetry (Frost: Poems (Everyman's Library Pocket Poets Series): Robert Frost, John Hollander: 9780679455141: Amazon.com: Books). Here’s my favorite one. Some say the world will end in fire, Some say in ice. From what I’ve tasted of desire I hold with those who favor fire. But if it had to perish twice, I think I know enough of hate To say that for destruction ice Is also great And would suffice. This came from a girlfriend 35 years ago, a comparative lit major who graduated magna cum laude from Berkeley and hand wrote comments in the margin for me. | 2/20/2018 |
What is the best quantitative research topic? | Abstractly? I’d watch Powers of 10. Then ask the question, how does the concept of space and time “quantity” change at these different scales? E.g. At the cosmological end, is there a maximum distance? Will we ever know? Note that space curves at this scale and so distance is measured differently. What is the earliest/latest time we will every be able to observe? At the quantum end, is there a minimum distance? What is the smallest thing we can measure? Is space quantized? How about minimum time interval? Is time quantized? | 6/4/2018 |
What is the best way for me to learn professional "computer science" stuff? (see description) | You can use Coursera, or MIT’s OCW or Udacity or a variety of other open courseware curricula. | 1/7/2017 |
What is the best way to find small-scale ideas for a first time entrepreneur that could be bootstrapped without angel or VC funding? | If you want to sell a product or service, I did this by starting a consulting practice to gather requirements, create a prototype and get initial funding (from the consulting fees). Once I got traction, I got VC funding. Teknekron was expert at this. Checkout the Harvey Wagner references here: P&A: Pitch Checklist | 4/6/2018 |
What is the best way to propose a side project to your boss? | Not sure what state/country you are in, but generally any work you do, especially work using company time or resources, legally belongs to the company that gives you a W-2. So if you quit, do it yourself, are successful and your current company is paying attention, they will then choose to sue and take the company away after you have done all the work and proved it. One way to handle this is to talk to your boss directly. Say that you have an idea you’d like to pursue under the umbrella of the company. Don’t show a detailed plan, show a broad outline and ask if he is interested in talking further. If he seems interested, say you want to head up the effort. If he says no, ask why. Another way is to become a consultant. You then get a 1099 and in your employment contract, state that projects using your own time/resources belong to you. Use the idea you have now to get you into that consulting gig. Then develop another on your own time. | 3/18/2018 |
What is the best way to study application of neural networks in finance? | Selshyam Selvaraju’s answer, where he recommends a book written in 2005, is great considering he answered the question in 2013. The 2012 ImageNet competition pretty much changed the AI world when University of Toronto’s “Deep Convolutional neural Network” entry blew the competition away. After that, the field was called DeepLearning, not neural networks. The 30 year old technology finally started showing results due to model scale (60M model parameters), processor speed (e.g. GPU processing) and data size (3–4*10**4 data classes). However, whereas deeplearning has made great strides in image and voice recognition, in (published) financial projects, not so much. These overviews as of 2016 are not particularly overwhelming, Even now, people are pretty much just trying sh*t out. https://bfi.uchicago.edu/sites/d... is a recent effort applied to mortgages. These guys (CFA Society San Francisco) did technical analysis by feeding a a rolling time series of stock prices with a year lookback (252 days) into a shallow convolutional network built on top of Keras (Keras Documentation). Results were not great. Hopefully studying these examples will at least get you started. | 7/6/2017 |
What is the biggest pain point for your business? | Getting and supporting customers. | 1/8/2017 |
What is the concept and definition of resources? | vague question. here is a vague answer. In computer science, a resource is a shared object subject to contention. Hence the need for an Operating System, which serves to mediate access to shared resources. Each mediation mechanism usually has it own model. E.g.
| 7/31/2017 |
What is the data structure used to perform recursion? | Traditionally, one uses a stack. Tail recursion can be done with static variables (What is tail-recursion? has a nice discussion of this). In a distributed messaging architecture, one can (sometimes) extend tail recursion by passing "continuations" (complete buckets of state) between processes (or processors) along with a "goto" address. | 1/4/2015 |
What is the difference between a file system and a database? | in the Pick operating system - Wikipedia, nothing. | 2/8/2017 |
What is the difference between Deep Boltzmann Machine, Deep Belief Networks and Deep Auto-Encoders? Is there is tutorial which explains the difference between procedures for pre-training and fine-tuning the above networks? (pseudo code would be of additional help) | A Deep Learning Tutorial: From Perceptrons to Deep Networks | 4/13/2015 |
What is the dumbest coding question you have been asked? | Not so much dumb as clueless. Part of completing my chemical engineering degree involved solving some piping flow equations using an interpretive language called APL (in 1975). Like Lisp or Python, you’d type stuff into a console, hit carriage return and and expression would be reprinted, evaluated and an answer returned. I worked by myself on the problem set and on the day before the stuff was due (Sunday), two pretty girls (April and Terry) came by and asked if I could help. They said that the computer had changed their password and they could not get back into their account. Sure enough, their default password did not work. Somehow they had managed to change it. So. Given no other information than that, what could they have changed it to? I put myself in their shoes, thought a moment and guessed the answer correctly the first time. Can you guess? They had changed their password to a carriage return. I got into their account, turned triumphantly to them and they said “Ok,….unh…. thanks”. No applause, no praise. It was clearly clueless of me to expect any other response since the solution assumed total ignorance on their part anyway. Shortly afterwards, a two guy “jock” team came by and said that the terminal was unresponsive. I sat down and looked at the terminal. Unlike the girls, these guys had a printout. They had been quite busy. Pages and pages of code, commands, epithets. I clearly couldn’t read through it all. So I applied the same reasoning. What could two clueless guys have done that made the terminal act that way? Again I thought a bit and came up with a single keystroke answer. Can you guess? At some point in all those pages of mess, they had typed an opening quote. They then spent hours typing in a string filled with code, commands, carriage returns, epithets, etc. and the interpreter sat there and waited dutifully for a closing quote. I typed the closing quote, hit carriage return and all the crap they had typed in was re-typed out by the interpreter, interpreted and “ERROR” was returned. In front of everyone in the machine room, who turned to watch the reams of paper spill out. The jocks reaction was different from the girls. They were pissed that I could solve their coding problem by typing in just one character, that the machine drew attention to their plight and that they then looked like idiots. They didn’t talk to me for the rest of the semester. Again, no applause, no praise. And again It was clearly clueless of me to expect any other response since the solution assumed total ignorance on their part anyway. So, all this is prologue. To answer the question. The dumbest coding question I was ever asked was to myself. “What can I do that fixes the problem and invokes appreciation?” The clueless one here was me. Given the above situations, having complete and accurate models of how these people thought AND actually being able to think a few steps ahead, I expected to be able to perform a magic trick for these guys and get some sort of positive response. However if I had just thought one more step ahead in both situations, I could have predicted PERFECTLY the complete system response. From then on, I worked hard to run my simulations to completion. | 8/13/2017 |
What is the earliest form of AI you can think of? | From Gulliver’s travels - 1726 The Engine is a fictional device described in Gulliver's Travels by Jonathan Swift in 1726. It is possibly the earliest known reference to a device in any way resembling a modern computer. It is a device that generates permutations of word sets. It is found at the Academy of Projectors in Lagadoand is described thus by Swift:
Stanisław Lem in Summa Technologiae (1964) and McCorduck (2004) connect the machine with the Ars Magna of Ramon Llull (1275), a mechanical device for combining ideas to create new ones. | 9/13/2017 |
What is the example of a programming skill or a computer science skill that I can apply for a small business entrepreneurship in developing countries? | If the goal is to help small business entrepreneurs... Program a stable currency, credit, debit and/or banking system. Filipino Smart phone users use prepaid cards as debit cards to enable commercial transactions. PayNearMe allows the 65 million US citizens who are members of the cash economy to order products by phone (e.g. Greyhound bus tickets) and go to 7-11's to pay for them (essentially expanding 7-11's inventory of products offered). The ticket is printed on the receipt after payment. | 7/11/2015 |
What is the first thing to learn about mathematics? | The value of Formalism (philosophy of mathematics) - Wikipedia Gödel, Escher, Bach - Wikipedia has a nice lay description of formalism, however the upshot is that various areas of math (e.g. geometry, algebra, calculus, computing machines (post, turing, von neumann)) can be viewed as games governed by unbreakable rules and if you can ‘map’ practical problems to these games and if you can construct a machine (e.g. a computer program) to express and solve them, you can answer some very interesting (and sometimes profitable) questions. | 12/27/2017 |
What is the future of programming? Will there be any easier language available for the masses? | Re english. If it is used, it will probably be a restricted subset (like COBOL). There is one (coarse) example of a broadly used, comprehensive English programming language now: Legalese. Contracts and laws are programs written in a broad English subset, meant to interpreted by lawyers. In the US, all make use of a shared library (constitution, Bill of rights and case law). However, the language (and subject matter) is so imprecise that you need a series of trained interpreters (appellate courts and eventually the supreme court) to determine if the programs are "correct". I don't believe there will ever be an automatic law interpreter; and if there is, it may run into the same problems as Murphy in Robocop II (Program example below). Here, although the directives were clear, the result was a program that was not adequately tested and so did not function well in its intended use. The film pointed out a truism about programming in general. Compared to Testing, Programming is relatively easy. Coming up with test sets that run the program through all combinations of conditions is difficult, especially if you want the tests to run in a finite time. Even a language as simple as excel shares this problem. I've seen many business plans which use Excel with crap assumptions or careless mistakes that make their way to the desk of a VC. They were not adequately tested and so embarrassed the authors at the one instant that they absolutely needed it to work. | 11/14/2014 |
What is the hardest obstacle you had to overcome when you decided to start your business? How did you do it? | Two: Finding a market and Hiring. Finding a market: I started a consultancy and started looking for a problem that was so bad that the client would pay me to develop a solution and let me keep the software. It took two years to find a market that also met my size requirements. Hiring: The first set of partners/co-workers is critical. I got around this by bringing everyone in as a contractor and then converted them to employees after working with them a few weeks. I went through 3 VPs of marketing that way before I found one that was a match. | 3/9/2016 |
What is the importance of reliability in research? | Well, if by that you mean reproduceability, it depends. As a purist, I’d like to point to cold fusion and poly water as examples of research at its best. No one could reproduce the results so the associated theories were discarded. As a pragmatist, I saw an estimate somewhere that 40% of published results could not be reproduced. Whether that is the result of mistake, fraud, or poor procedure, I don’t know. I do know that no one likes it when mistakes are pointed out. E.g. this science fair project (Improving Non-Invasive Blood Analysis) pointed out a fundamental flaw in a Samsung home health care project published in a refereed journal and not a ripple resulted. | 2/8/2017 |
What is the most beautiful proof in math? | Pythagoras These 4 triangles can be arranged in a big square shape in the two ways shown below. Clearly, the sides of each big square are of equal length (a + b) and so must have equal area. Also, taken together, the 4 triangles take up the same area in each big square. So if we subtract out the triangles, the area of colored square "C" must equal the combined areas of the colored squares "A" and "B". That is to say “area of square C” = “area of square A” + “area of square B”. But "C" is a square with sides of length "c" and "B" is a square with sides of length "b" and "A" is a square with sides of length "a", so c2 = a2 + b2. Gauss Carl Gauss is one of the smartest mathematicians who every lived. There is a famous story about him listed here. It says "in elementary school his teacher tried to occupy pupils by making them add up the integers from 1 to 100. The young Gauss produced the correct answer within seconds by a flash of mathematical insight, to the astonishment of all. Gauss had realized that pairwise addition of terms from opposite ends of the list yielded identical intermediate sums: 1 + 100 = 101, 2 + 99 = 101, 3 + 98 = 101, and so on, for a total sum of 50 × 101 = 5050 (see arithmetic series and summation)". Now, suppose you copied the above staircase and flipped one over, like we do below. Now, fit the staircases together. This makes a rectangle that is 5 tall and 6 (not 5!) wide. The number of blocks in a rectangle is easy to figure, it's just the width times the height or 6 x 5 = 30. But to get the number of blocks in just one staircase, you have to divide the area by 2. So the sum of the digits from 1 to 5 is | 6/9/2017 |
What is the most efficient use of a penny? | electroplating. | 12/12/2014 |
What is the most profound PhD dissertation you have seen on the topic of artificial intelligence? | Terry winograd’s | 5/15/2018 |
What is the next, next big thing? I mean after AI/AR or autonomous vehicles? | Tough to tell. This may help. Why Software Is Eating the World Create a timeline. I have one bet on Waldos. Waldo (short story) - Wikipedia | 3/10/2017 |
What is the perfect book for beginner book readers? | 11/8/2017 | |
What is the point of life if you cannot change the world? | Lots of answers to that one. You can try. You can change it one person at a time (with an act of kindness or an act of hate). Find another point. There is no point, was no point, has never been a point. You get to pick your own point. It kind of goes on like that. For what it’s worth, I’m of the opinion that if, in your opinion, a question does not have an answer, the issue lies with the question, not the answer. | 2/20/2018 |
What is the relationship between the factors affecting the running time of an algorithm? | Computational complexity, code size, and data size. O(n) is better than polynomial time is better then exponential time. Keeping the code in cache is better than swapping it in. Keeping data in registers is better than keeping it in cache is better than keeping it in ram is better than keeping it on the network is better than keeping it on disk is better than keeping it on the internet. | 2/19/2018 |
What is the solution of a? | You can cross check this stuff by using google as a calculator. x = (-b +- sqrt(b*b-4*a*c))/2a 9+sqrt(18*18-4)/2 = 17.94427191 First Root Check 17.94427191*17.94427191 -18*17.94427191 +1 = 1.5063506e-11 First Answer 17.94427191**.33333 = 2.61800879229 First Root Cross Check 2.61800879229 * 2.61800879229 * 2.61800879229 + 1/(2.61800879229 * 2.61800879229 * 2.61800879229) = 17.9994835167 9-sqrt(18*18-4)/2 = 0.05572809 Second Root Check 0.05572809*0.05572809-18*0.05572809+1 = 1.5048185e-11 Second Answer 0.05572809**0.33333 = 0.38196968739 Second Root Cross Check 0.38196968739*0.38196968739*0.38196968739+1/(0.38196968739*0.38196968739*0.38196968739) = 17.999483518 | 3/3/2015 |
What is the task of philosophy? | 1) A lot of philosophy concentrates on answering one question: "What do I do next". 2) Ayn Rand actually touched on this in her essay: Philosophy: Who Needs It? after describin several branches of philosophy, she makes the following comment: Now some of you might say, as many people do: "Aw, I never think in such abstract terms — I want to deal with concrete, particular, real-life problems — what do I need philosophy for?" My answer is: In order to be able to deal with concrete, particular, real-life problems — i.e., in order to be able to live on earth. You might claim — as most people do — that you have never been influenced by philosophy. I will ask you to check that claim. Have you ever thought or said the following? "Don't be so sure — nobody can be certain of anything." You got that notion from David Hume (and many, many others), even though you might never have heard of him. Or: "This may be good in theory, but it doesn't work in practice." You got that from Plato. Or: "That was a rotten thing to do, but it's only human, nobody is perfect in this world." You got that from Augustine. Or: "It may be true for you, but it's not true for me." You got it from William James. Or: "I couldn't help it! Nobody can help anything he does." You got it from Hegel. Or: "I can't prove it, but I feel that it's true." You got it from Kant. Or: "It's logical, but logic has nothing to do with reality." You got it from Kant. Or: "It's evil, because it's selfish." You got it from Kant. Have you heard the modern activists say: "Act first, think afterward"? They got it from John Dewey. Some people might answer: "Sure, I've said those things at different times, but I don't have to believe that stuff all of the time. It may have been true yesterday, but it's not true today." They got it from Hegel. They might say: "Consistency is the hobgoblin of little minds." They got it from a very little mind, Emerson. They might say: "But can't one compromise and borrow different ideas from different philosophies according to the expediency of the moment?" They got it from Richard Nixon — who got it from William James. Now ask yourself: if you are not interested in abstract ideas, why do you (and all men) feel compelled to use them? The fact is that abstract ideas are conceptual integrations which subsume an incalculable number of concretes — and that without abstract ideas you would not be able to deal with concrete, particular, real-life problems. You would be in the position of a newborn infant, to whom every object is a unique, unprecedented phenomenon. The difference between his mental state and yours lies in the number of conceptual integrations your mind has performed. You have no choice about the necessity to integrate your observations, your experiences, your knowledge into abstract ideas, i.e., into principles. Your only choice is whether these principles are true or false, whether they represent your conscious, rational conviction — or a grab-bag of notions snatched at random, whose sources, validity, context and consequences you do not know, notions which, more often than not, you would drop like a hot potato if you knew. But the principles you accept (consciously or subconsciously) may clash with or contradict one another; they, too, have to be integrated. What integrates them? Philosophy. A philosophic system is an integrated view of existence. As a human being, you have no choice about the fact that you need a philosophy. Your only choice is whether you define your philosophy by a conscious, rational, disciplined process of thought and scrupulously logical deliberation — or let your subconscious accumulate a junk heap of unwarranted conclusions, false generalizations, undefined contradictions, undigested slogans, unidentified wishes, doubts and fears, thrown together by chance, but integrated by your subconscious into a kind of mongrel philosophy and fused into a single, solid weight: self-doubt, like a ball and chain in the place where your mind's wings should have grown. | 12/3/2014 |
What is your biggest pain as a self-directed investor? | In the financial markets: Unbiased reporting. Financial News reporting is indistinguishable from political reporting during the 2016 presidential campaign. Time is given to what will be watched and what sells ads (Trump bloviating), not what the product (the reading audience) needs. E.g. most funds fail. Most financial advice is bad. Still time and again, talking head time is given to people that ought just shut the hell up. As Calvin remarked once in Calvin and Hobbes. Its amazing that no matter what happens during the day, news always takes just 1/2 hour to report. In Angel investing, finding someone who treats my money as preciously as I treated my initial investment when I was an entrepreneur. | 11/19/2017 |
What is your definition to the "startup culture"? | For me, a “business culture” is created by group of people who dedicate themselves to achieving a shared business goal. Individually, they commit by contributing their time, focus, resilience, and energy. Jointly, they posses enough trust, mutual respect and common vision to believe that the others will execute toward that goal in their absence. | 6/14/2016 |
What is your favorite electronic invention and why? | The first active electrical component: the Vacuum tube - Wikipedia (forerunner of the transistor). | 5/8/2018 |
What is your favorite invention you created? | It was the early 80’s and workstations were just appearing. Daisy sold a Computer-Aided-Engineering workstation with a set of Design Automation Tools: a drawing editor, a design compiler and a simulator. Unfortunately, the system operated via a command line and the newer workstations (Apollo/Sun) had window based operating systems. Our operating system group of 6 or so had been working on a window based system for a year and we were being slaughtered in demos. The president of the company asked me to create a window system that we could use in the simulator, and I made an internally demoable version in 3 days and a customer facing demo in a week. The user could bring up a schematic and create windows with “marching waveforms” that associated with wires on the display. Lots of limitations, but it helped us make our quarters until the OS group could deliver a working window system. | 9/27/2016 |
What is your idea of fair? | A deal where you are willing to take either side. | 10/25/2017 |
What is your unpopular opinion of Mathematics? | Math is a series of kludges based on faith in inverses. It probably started with Shepherds counting sheep and noting that the properties of addition worked pretty well for the job. Shepherds then noted that they could use an inverse operation (subtraction) to also do accounting. But what happens if you have 4 sheep and give away 5? Some shepherd probably got in that fix with a creditor and so invented negative numbers. Corralling sheep in a pen gave our shepherd friend the idea of multiplication and squares. So there must be a square root. But what is the square root of two sheep? or worse, the square root of -1 sheep? (I won’t go into the mess of division, otherwise you’ll get Clarice Starling’s nightmares). And just as inverses can create new domains willy-nilly, they can do the same for functions. For sure, all these created new, valuable areas of mathematics, but the inspiration was based on faith that inverses would ‘work’, that inverse operations on numbers would produce another number that was still in the original domain. As shown, they may not. | 10/24/2017 |
What kind of cheap equipment does it take for a teenager to start a lawn-care bussiness? | I started by going door to door, offering to do garden work (including mowing lawns) and asking if I could use the homeowner's lawnmower. As a teenager, many people understood my lack of equipment. By being sincere and ready to work , many lent me theirs. It's a tough way to start, but it can be done. | 3/30/2016 |
What kind of computer would have been used for desktop scientific computing circa 1992? | IBM, Dell, Compaq, Micron or Gateway. 386 based. 486 came in 1994-1995 | 1/2/2015 |
What kind of jobs should I do if I get depressed from staying in the same spot for longer than an hour? | You might consider stepping back and asking “why”? E.g. For me, I can stay in one spot if I’m doing it for a reason that is under my control, or is of limited duration, etc. What is your long term goal and what job will help train you for that? As suggested by Ashutosh Kumar below, being a waiter is a fine goal if you want to to eventually start your own restaurant. | 8/13/2016 |
What math should I take after Single Variable Calculus? | Skim Gödel, Escher, Bach - Wikipedia , Especially the horse/apple chapter. If you like it, consider studying formal foundations of math and basic computer science. All the other stuff you mention will be covered in college anyway. | 12/30/2016 |
What nowadays technology would have the largest impact on mankind if it was discovered 2000 years ago? | 10/18/2017 | |
What online courses are available for the Lean Startup methodology? | How to Start a Startup | 11/14/2014 |
What percentage of new CS graduates are able to find jobs? | I don’t know the breakdown, but You probably should have majored in computer science says there are 10x more positions than graduates, so if you don’t get an offer before you leave, you might consider expanding your network and then getting a consulting gig. | 3/12/2018 |
What problems did you expect to be remedied by 2015? | Hell, I'd settle at this point for problems to be recognized, not remedied. Going back just a few years, I expected
| 9/13/2015 |
What programming language(s) do you recommend learning in order to best understand artificial intelligence/machine learning? | Python | 12/14/2017 |
What should be the suitable name for dept/academic work automation project? | Interesting. When I named my last company (Calico), I had a list of names that I showed my potential customers and let them add /pick since they were the ones I was getting the money from. Approach seemed to work OK. First guess, try a prefix plus the word ‘flow’ or “ation” in it, the people you are trying to impress will hate it and suggest something different. Then pick one of those. | 11/20/2017 |
What should I do after Google's Machine Learning Recipes? | Why not figure out a way to do warehouse optimization? Ask your boss where the major cost is in your work and figure out a way to reduce it using a machine learning technique. | 2/12/2017 |
What should I know before starting a trucking company? | Work for one for a while, ask questions and write a everything down. | 3/8/2016 |
What should I read to learn more about AI safety? | Joseph Weizenbaum - The Tech wrote on aspects of this in the 70’s. | 1/2/2017 |
What single object, if sent back 2000 years in time, would have the most profound effect on civilization? | Rats, Lice and History (Social Science Classics Series): Hans Zinsser: 9781412806725: Amazon.com: Books stated that the single factor that determined whether a siege would succeed is how much wine the attacking force brought. They knew that if too many people gathered together, that they would get cholera, and that if they drank wine instead of water, that they would not get sick. So basic, actionable hygiene points would probably have a big effect. E.g. Boil your water before you drink it. Have women wipe themselves from front to back instead of back to front. Converting poo to fertilizer, etc. All that, if accepted, could have a profound effect on civilization. But how would this actually turn out? Instead of traveling 2000 years in time, you can travel 6700 miles in space (NYC to Zaire). What happened when peace corps volunteers showed stuff like this to the local population, and proved it by having 1/2 the village try it and the other half keep the old ways? The population kept it up as long as the volunteers were there. When the volunteers left, they stopped. When asked why, they said “Hey, it was YOUR idea. It was important to YOU. So when YOU left, WE stopped”. Few wanted to adopt habit changes in the present in exchange for a theoretical benefit in the future (despite evidence that they themselves gathered and measured). Stupid you say? What percent of the American population still drink, smoke, do drugs, overeat and don’t exercise despite ‘evidence’ that the opposite is in their best interest? These attitudes are re-enforced by the local culture. I.e. dedicated drug users who congregate, dedicated smokers who are isolated outside, obese people at buffets, etc, -strengthen- their common behavior/beliefs together since most would rather commune with like-minded people than endure a theoretical unasked-for ‘improvement’ forced on them externally by an outsider. Apparently, Culture trumps ‘evidence’. (pun intended, also see To err is human; so is the failure to admit it). (Not you? Your decisions are evidence based? So tell me, how many pushups did you do today?) This argues for the ‘right’ person getting the object, otherwise you are likely to get the same reaction to this ‘new’ technology that Dr. Zaius showed in ‘Planet of the Apes’, or receive the same reception to your object that Galileo got for his work. One can argue that new objects were not the initial force that changed our world. First, our perception of the world changed, and then new objects enabled us to operate within that new perception. (See James Burke’s The Day the Universe Changed: How Galileo's Telescope Changed The Truth and Other Events in History That Dramatically Altered Our Understanding of the World (Back Bay Books): James Burke: 9780316117043: Amazon.com: Books) That said, something with an immediate effect (e.g. a machine gun with several crates of ammo) could probably change perceptions pretty quickly. | 8/3/2017 |
What statistical study can I do in a thesis about the impact of artificial intelligence on the labour market? | Economically, I’d model AI as slave labor without human rights. Then I’d look look at the impact of labor outsourcing in the textile industry in particular and when NAFTA was implemented. I’d include a product pricing model as well as a labor model. E.g. in cases where people made pennies an hour (e.g. India), I suspect end user prices went down in two stages. One: A sharp drop which was enough to displace US labor. Two: a gradual decline as all offshore workers (competing AI technologies) competed against each other. I’d also look at unintended but predictable consequences. E.g. in textiles, the fashion cycle became more rapid and custom. This all has been heavily studied and there are lots of stats. | 3/16/2018 |
What subfields of Machine Learning are the closest to those with background in Fluid Mechanics? | I assume you are interested in contributing? If you look at the first few lessons in Andrew Ng’s course (see Bill Paseman's answer to What are the most important deep learning algorithms? In which order should I learn them?), he solves a prediction problem using standard linear regression. I.e. for a set of Ys (predicted house prices) and a coefficient matrix A (Square footage, home age, house location) he fits a hyperplane (intercept, slope, …) using the standard formula x = (A.T * A).I * A.T * y (This is python numpy notation). Now, as you apply this approach to a large amount of data (millions of ys) with thousands of features, A gets very large and inverting A.T * A is problematic. So Ng introduces the method of gradient descent. Here, he adopts an initial (wrong) solution, takes the partials and uses them to incrementally travel ‘down’ the hyper-dimensional surface to a local (or if you are lucky, global) minimum. This approach is used recursively in neural nets. If you are finished with standard undergraduate Fluid mechanics texts (e.g. Transport Phenomena, Revised 2nd Edition: R. Byron Bird, Warren E. Stewart, Edwin N. Lightfoot: 9780470115398: Amazon.com: Books) The equations should look pretty familiar and the figuring out ways to solve them quickly is where a lot of time is spent. Is that what you were looking for? | 2/3/2017 |
What tax benefits does California offer high tech companies? | Look into Qualified Small Business Stock (QSBS).Rules change, but in my time it excluded the first $10 M gain from state tax.Check with your accountant. | 6/10/2015 |
What technology should we expect cities to realistically adopt within the next 5-10 years (2021–2026)? | Stuff that increases revenue, reduces cost and is under municipal control. Look at San Francisco. Cameras mounted at license plate level on buses that monitor whether a car has been longer than 2 hours in a 2 hour zone. This will get totally automated with image recognition improvements to overcome people's current explanations (E.g. "I left and came back." "Then why were your wheels in exactly the same place?")) Variable rate parking meters (now in existence) will include license plate level cameras to detect if someone is stopped there (even for less the 6 minutes) and issue fines fines if the meter isn't paid. Note that Fastpasses are, in essence, a municipal automatic payment system. So, completely automated tollbooths which force even visitors to buy "Fastpasses" or be fined (automatically of course). An option to tie fastpasses in with the field effect detecting parking meters to force the motorist to pay for parking using the fastpass system. Admittance to municipal facilities (e.g. parks) will require Fastpasses to pay. A system to make fastpass work in any city. Fastpass for paytoilets. Further out, but following the same theme. Before there is Image recognition in both surveillance cameras and drone overflights to direct police activity to predicted hotspots, there will a automatic inspection of restaurants for code violations and automatic generation of fines. Incorporation of government A/C controls to smooth power spikes in hot weather. Simulations as to whether particular tax initiatives will actually pass at the ballot box. Anyway, you get the idea. In response to the ‘RFID license plates’ mentioned in comments. The question specified a ‘5 year time frame’ and ‘cities’. The bus mounted cameras are not speculation. SF already uses them to issue tickets and collect revenue. Legally, since the buses belong to SF, they can modify them anyway they want. Since SF also control the municipal courts, they can enforce ticket collection. License plates are issued by the state, not the city. This means that SF mechanism has fewer legal barriers to realization than something that requires state co-operation. Again, the driver will be revenue | 5/16/2016 |
What technology would you design to upload data collected from 10 plants/sites? | (I believe) the problem is under-specified. “provide sites “ - do you mean “Accomodate sites “? How many sensors are there per plant? “Not a PC”. Ok, than what are the sensors connected to? What is the sensors data format and precision? How often do the sensors update? Is the data secure? What communication mechanism(s) are in place? Wired? Wireless? Bluetooth, Wifi? How much noise is there? My questions go on quite a while like that. | 3/2/2017 |
What things do you enjoy most about starting and having your own companies? | I worked a year for ARCO fresh out of school. Big Company. I saved them more than my annual salary in my first week of work. Did better than anyone else in the division the first year. I got the top of the range in salary review - 8%. Later, after I started my own consulting company, I made $250K in 1992. I then learned how to do a product company, 8 years later I made 100’s of millions. I owe it all to ARCO. Good people there, all thinking about themselves and their futures. They drove home the fact that no one (not even my mother) cared more about my future than me. | 8/12/2016 |
What types of architecture projects does the world need in 2018 going forward? | In line with the other answers, Figure out ways to eliminate usage of concrete. | 12/7/2017 |
What types of ideas draw the attention of angel investors? | Elaborating on some of the answers already given, here is what I consider a perfect pitch (P&A: Pitch Checklist).
| 11/13/2017 |
What vendor would you choose to help create and operate a internet accessible international clinical trial database like https://databank.vhl.org/? | Home Page - VHL Alliance is hosted by Home - NORD (National Organization for Rare Disorders). NORD's FAQ lists a maintenance cost of $500/month. Another possibility is REDCap which provides free software but no hosting. Redcap in particular makes a big deal about being able to export de-identified data. Not sure how this plays in Germany though. Comparing STRIDE-RDM, REDCap and Medrio shows how Stanford views three different approaches. Note that neither NORD nor RedCap cover Data Cleansing Costs (patient data validation); biospecimen banking or serial blood collection. A rough rule-of-thumb for patient data validation is $100/record if a radiologist needs to look at it. | 12/27/2015 |
What was the key takeaway for you from Tim Cook's graduation speech at MIT, if you saw it in person or online? | I did not see it, but was texted this by an audience member. [12:38 PM, 6/9/2017] some highlights from an apple ad [12:38 PM, 6/9/2017] i'm sorry [12:39 PM, 6/9/2017] MIT commencement speaker [12:39 PM, 6/9/2017] - Joke about experimenting once with a Windows PC, "and obviously that didn't work". - "I finally felt aligned with a company (Apple) that brought together challenging, cutting-edge work with a higher purpose." - "Aligned with ... technology that didn't exist yet could reinvent tomorrow's world." - "It means an iphone that allows a blind person to run a marathon" - "It means an Apple watch that catches a heart condition before it becomes a heart attack" - "It means an iPad that helps a child with autism connect with his or her world" - "Someone questioned Apple's investment and focus on the environment..." - "I pointed out that Apple does many things, like accessibility features for those with diabilities that don't rely on an ROI. We do these things because they're the right thing to do." | 6/11/2017 |
What will be the next big tech trend after Ai? | Figuring out how to turn the damn thing off….. | 6/13/2016 |
What will my grade be if it is currently 93% and I got an 85/100 points on a project (projects are worth 25%)? | If 93 is 75% of your grade and 85 is 25% then .75*93 +.25*85 = 91 A quicker way to calculate this is to realize that your project grade is 8 points below your average, but counts 25% of the total, so the total must drop by 25% of 8 or 2 points. | 12/18/2017 |
What working model should I make for my science project? | What age are you, what are your interests and what budget do you have? Based on Comments: How about a Double pendulum? A double pendulum quickly shows the limits of physics’ predictive power even for simple devices. Its Chaotic behavior arises from the inability to determine its initial condition in the real world. What is your math background? Do you know how to program? A basic Question is how long do you need to observe a double pendulum’s behavior until you can determine the path it will take from that point forward? You could try to do this via machine learning or other methods. | 11/23/2016 |
What would be a science project ideas that use pulleys? | Look at old sailing ships and figure how much more force they would have required if they simply hoisted sails instead of using the block and tackle. Then relate that to the crew complement and figure out how many more sailors they would have needed if they hadn’t used that technology. A more prosaic example is how many people you would have needed to wash windows using a simple pulley instead of the compound pulley setup used in the early 20th century. You can proably get more ideas by looking a rigging manuals. | 11/6/2017 |
What would be the best science project in a science fair? | Saving the World: Measuring which plant species consume the most CO2. (and so are best at fighting global warming.) See also Tropical forests now emit more carbon than they soak up and Bill Paseman's answer to How can a carbon capture and storage model be made for a high school science fair project? What are some tips? Use deepchem/deepchem to repurpose a drug for Papillary Kidney Carcinoma a listed in The Cancer Genome Atlas Home Page Improving Non-Invasive Blood Analysis is an example of fact checking an existing paper. No awards, but it lead to an Intel semi-finalist paper: http://mazziotti.uchicago.edu/jo... If you get into an area, rummage around a bit and then pursue something of interest, it might lead somewhere. | 3/10/2018 |
What would be the best science project in a science fair? | If the goal is to win, independent research into a novel topic that shows new results. | 1/7/2017 |
What would happen if Google uses its data to predict the stock market? | As discussed, it is useful to know which data. However
I’ve got more. Is this what you are looking for? | 7/8/2016 |
What would make a great science fair middle/high school project that is also affordable? | Consider using this as a starting point: The Serious Physics Behind a Double Pendulum Fidget Spinner You could use it as an introduction to chaos theory. | 11/28/2017 |
What would you do if you had a lot of ideas but had no knowledge on how to produce them? | If you are truly passionate about the idea, use it as a driver to learn how to realize it. E.g. When I was 20 (1974), I wanted to get a computer to compose music and play it in real time. I learned programming, became the TA at an extracurricular arts class that had access to a graphics terminal and convinced a student to hook up his homebrew digital synthesizer to it. In the process I learned sales, digital hardware, programming and (some) music composition. | 7/7/2016 |
What would you like to know about starting a startup? | What do I want to know? How to have guaranteed success with a big payout and as little work and investment as possible. | 6/13/2016 |
What's a good first artificial Intelligence project for a group of students to gain first experiences and insights? | AI cover a lot of ground. It kind of depends on your interest. However, given no prior… Nowadays AI is viewed as being synonymous with Machine Learning, which is viewed as being synonymous with DeepLearning. If you want to get your feet wet in that, try Tutorials | TensorFlow If you want a general grounding in Machine Learning, try the first few homeworks in Machine Learning | Coursera If you are past that, you might enjoy Competitions | Kaggle or The Marketplace For Algorithmic Trading Systems | Quantiacs | 7/3/2017 |
What's it like to be the founder of a high-visibility startup that's failing? | Remember the public perception of the trump campaign before the election? | 11/23/2016 |
What's the best book I can read to get me started with computer science? | MIT’s “Structure and Interpretation of Computer Programs” by Gerry Sussman and Hal Abelson http://web.mit.edu/alexmv/6.037/... | 11/6/2017 |
What's the best orientation should we look at to find startup ideas? | To Ken Ho’s point, in the past, Sequoia has often looked for entrepreneurs who are solving their own problem. That said, solving an problem valued by the customer and you can make money from is also important….. | 7/7/2017 |
What's the best personal computer for data science? | It depends on what kind of data you are analyzing. A Single Tumor Genome can take on the order of 400 GB. The best way to deal with datasets of this size is via AWS, in which case the PC is less of an issue. | 3/25/2017 |
What's the best source for news about biotech companies and clinical trials? | For clinical trials on human, try Home - ClinicalTrials.gov | 2/15/2017 |
What's the best way to narrow down which startup idea to pursue when you have multiple? Context: have written down/elaborated on dozens of ideas, done some kind of analysis on around 10, and currently confined it to 3 or 4 top choices. | 4/14/2018 | |
What's wrong with transhumanism or uploading everything known/imagined to a huge construct of neural networks? Why is this idea so disturbing? | I can’t speak to the attitudes/beliefs/opinions of others, but I don’t find the idea disturbing, I just don’t think it’ll work (assuming transhuman means we keep what we got and get something extra). As I discuss in Bill Paseman's answer to Life originated from chemicals, Abiogenesis, and evolution did the rest of the work. All of our thoughts and emotions are simply electrical charges and chemical reactions in the brain. Are we just bags of chemicals, moving around, and nothing more? , our models of the brain are low fidelity and I suspect that the way they work is different from what we “know/imagine”. By way of example, suppose I put snot in a bottle, freeze it and claim that this makes me immortal because at some later point in the future, it will be unfrozen, have dna extracted and cloned with memories restored from the information recorded in my high school yearbook plus a 5 1/4″ floppy. Lot of holes in that theory, despite how loudly and clearly I proclaim it. | 7/2/2017 |
What's your motivator to work on your side projects? | Staying alive as long as possible, and enjoying my time on earth as long as I have it. | 2/20/2018 |
What’s the next big step in AI? | A partially supervised learning algorithm with an automatic motivation system to label input. For input, it would have both machine vision and hearing and sensors (temperature, pressure) so it can cross couple the input, motivation subsystem and the partially supervised learning algorithm. It will be able to run full time at human frame rates in the real world. It will have reasonable actuators and voice generation. It will be modeled after a baby, able to develop a mental model of the world simply by observation, directed by motivations like hunger, interest and approval from its mother. It will learn language mostly by observation. The mother does not continually interact with it but primarily provides social cues as to acceptable and unacceptable behavior (E.g. you cannot take apart a live cat to see how it works, regardless of your level of curiousity). | 11/29/2017 |
When hype surrounding deep learning stops, what specialist skills will still be in demand? | The hype will stop when it stops delivering value. It can't be applied to everything, however where it can be applied with good results (e.g. picture processing), demand will continue. | 3/21/2016 |
When measuring the electrical conductivity of a few standard solutions, why is it important to start with the most dilute solution and work your way up in concentration? | No real clue. My speculation is that the instrument you are using might have a narrow calibration range. In which case once it is calibrated for a dilute solution, it is likely (in percentage terms) to be reasonably accurate for small concentrations. If then exposed to large concentrations, it is likely to “peg” the meter to the right, but in percentage terms, still give a reasonably accurate reading. If, on the other hand, you calibrate for a strong solution, the weak measurement would likely have a lot of noise. Think of it this way. Suppose you want to measure the height of a marble, a baseball, a house and the empire state building all with the same instrument. If you calibrate it to measure the height of the Empire State building first, percentage error is in feet. If you measure the height of the marble first, percentage error is in millimeters. Anyway, without more details, this is just a guess. | 5/3/2018 |
When the enthusiasm and commitment with which you started a personal project start dying down, how do you rekindle? | Some repetition here, but see Bill Paseman's answer to How do I stay motivated and on track as a 'lone wolf entrepreneur'? and | 12/7/2017 |
When will deep learning get good at extrapolation? | Not clear what you mean by "get good at". Re your title, deep learning is one of many machine learning methods, it is much better at smoothing (looking at the past) and filtering (interpreting the present) than prediction. Re your quote: There are several machine learning models that identify outliers when they occur. (Thinking outside the sample) It is difficult but they are used for detecting rare events such as kidney cancer pulled from samples in a normal population. Standard control systems theory is used to adjust set points when inputs get out of range. | 5/2/2016 |
When will you leave your career and start your own business? | Been there, done that, got the T-shirt. Check out the Harvey Wagner references in P&A: Pitch Checklist. Basically, I started my own consultancy to explore potential markets; this allowed me to control the IP in exchange for cheaper pricing. I saved my money. Once I found a market I turned the service into a product, incorporated and started selling it. I grew the company (Calico) to 300 people and took it public. | 3/4/2017 |
When you fight valiantly and lose, what do you win? | I came from a lower middle class family, who saved to send me to college. The goal was to graduate with an BSChe in 3 years to save them some money. After three years, I blew my last course "History of Germany from Bismark to the present". So I regrouped and earned a BSChe, MChe and BSEE in 4 years. I did well in my first Silicon Valley job, and then proceeded to startup my own in 1985, which failed after $18M investment and 5 years work. It took another 5 years, but I corrected each mistake and was able to raise more money from two of the three original investors and take the company to IPO (I then left the company in what I felt were capable hands, but unfortunately the company foundered). And I've done more s--t like that. What helped me "overcome unbeatable odds?". Well, when I was 17, my friend next door was in a car accident. The other car took his head off. In my first job at ARCO, my cubicle mate was made a paraplegic and I saw an 18 year old guy lose a leg on a rig when we were laying casing. They did better than the the guy who hired me. He died in a helicopter crash at sea. This gave me at least a little perspective when I came to the valley. Guys would see sad faces appear on their computer screens when their code crashed and then they would throw fits. I'd think "well if this is their idea of bad day; I'll do just fine here". So for me, a rough rule of thumb is that if you can post a lament on Quora, you probably didn't "spend everything" and you'll do just fine. | 7/9/2015 |
When you graduated from MIT, did you become successful right away? Did you get a high salary job offer right after graduating from MIT? | There is an underlying point here that is probably worth making. When I was (very) young, I thought that I would accomplish some great act in my life, and then I would be a celebrity, dispensing advice from on high to acolytes hanging on my every word. If that theory is true, apparently MIT graduation is not that act. Instead, people then expected me to solve tough problems under pressure on a continuing basis for an extended period. That was true when I was salaried and when I created my own companies (Customers had the expectation at that point). I took my first job in Silicon valley at -below- market rates because I believed that we would go public. Turned out to be a good decision, and I made my first million 4 years later on the IPO. However, it could have easily gone the other way, and did for some who worked elsewhere or sold too early or too late. My boss back then, an Israeli, once remarked on my MIT education. I said “you realize that I did not make all A’s”. He said “Yeah, but you did make it through”. Kind of like Hell week. If you make it through, in the eyes of others, it qualifies you for tougher missions. | 2/6/2018 |
Where can I find embedded system mini project with source code? | Check out SparkFun Electronics or Make: Magazine | 10/26/2016 |
Where can I find mentors for my Google Science Fair Project? | Your local universities. The more competitive the better. See also Bill Paseman's answer to What are some great science fair ideas that could qualify for ISEF? | 6/2/2018 |
Where can I get a training data set about tweets for movie reviews? | Full of noise: Tweets from behind the camera. | 3/1/2017 |
Where can I get someone to fund my project apart from online crowdfunding? | If it is needed enough, a future customer. | 10/27/2014 |
Where is the best place to start writing a computer vision library? | 4/25/2018 | |
Where should I start learning about Deep Learning? | Nick has some good implementation resources. Here are some deep learning resources | 3/9/2017 |
Where should I start making a 'drag & drop' neural network editor? | 1/6/2017 | |
Which books can I read? | I have only read a fraction of these. But if you can follow Eliot’s premise and get through them in 4–5 years (an are able to digest it), you have presumably saved yourself $500K on a Harvard education. If you are interested in political philosophy, you might substitute “Duties of Man” ( http://www.csun.edu/~jaa7021/wes...) for Mazzini’s “Byron and Goethe”. His basic idea is that instead of creating a nation based on rights (like America), you create a nation based on duties. Unfortunately Fascism corrupted the original premise, but given that we (Americans) are now defined by what divides us vs what unites us, and quote our rights as a justification, I kind of wondered how it would have worked out if we had had an explicit “Bill of duties” set out by the founding fathers (vs. someone like Mussolini or Hitler). | 10/16/2017 |
Which crafts or Engineering disciplines need better tools for cheap and fast prototyping? | Aerospace | 8/25/2016 |
Which discipline of engineering is the best to go into for someone who is interested in starting a business after graduation? | Those who were wildly successful usually had something working while in school. Dell build computers out of his dorm room. Zuckerberg stole the Winklevoss' idea and implemented it for a few schools before graduating. Joe Liemandt (Trilogy) leveraged his father's contacts to build a configuration business. Gates had a functioning business. Many didn't finish their degrees, giving credence to the position that the idea is more important than the education. So; do you have an idea? Is it getting traction? If so, whatever degree supports that idea. If not, whatever field gets your blood running. If nothing excites you, EECS is not a bad field to get into since it often requires little more than an idea to start a software business. You could then set up as a consultant/contractor and use that company to get the idea. | 3/30/2016 |