5 ms·

>Such gaps are resume fraud Wait, what? I mean, you can argue that they are really bad and maybe you are right; I disagree, but maybe I am wrong. But how is

by lsc 13y ago

>Such gaps are resume fraud Wait, what? I mean, you can argue that they are really bad and maybe you are right; I disagree, but maybe I am wrong. But how is a blank spot fraud? Hell, I leave off most of my short jobs, I've done a fair bit of contracting, so there's a lot of that, and I've run businesses on the side for most of my life, so there is a fair bit of overlap, but even so, I don't see blank spots as dishonest (even if they may look bad) >Net, it's much easier for me to be the CEO of my own company, and likely get a few million in venture funding Hey, good luck. But until it's in the bag, guessing what is in the mind of VC is... not something I'd bet money on. Of course, I have gotten exactly zero dollars of VC, so what do I know? But my impression was that they were even more irrational than your average hiring manager, which is saying something. Of course, you don't always need VC to run a company. >A good Ph.D. in a STEM field sets one apart in a way that means that they might be able to get a job stocking shelves at Wal-Mart. Sorry 'bout that. Huh. See, last time I tried to get a job of that type, I got turned down because I had serious experience (even though my level of official education was appropriate.) - It was actually easier for me to get another well-paid SysAdmin job than to get a job at wal-mart or the coffee shop, even at the depth of the dot-com crash in 2001. (It was... actually pretty funny when I applied at the coffee shop. It was very clear that the hiring manager thought I was not remotely cool enough to work there.) Interestingly, most of the PhDs I've worked with were physics majors.

graycat 13y ago
It has appeared that the physics community has done much better getting good non-academic jobs and careers for their Ph.D. holders than most other STEM fields.
- analog31 13y ago
  Granted, I got my physics degree 20 years ago. At that time, there was actually a lot of frustration in the physics community about the lack of preparation for industry jobs. So I'd hesitate to speculate on why (or even if) the situation has improved. But I will anyway. ;-) I think that physics students aren't always "monogamously" in love with physics, so much as we're intellectual omnivores / opportunists. I was in love with math, electronics, and programming. But I ended up at a college with no engineering school, and the people who seemed to have the most fun playing with toys were my physics profs. But as a result, my ability to roll circuits and sling code with the best of 'em was still marketable in 1993. Today, I see a trend towards hiring people with more formal skills in their job areas, such as CS and EE degrees for programming and electronics design. And I see more emphasis on tools (IDE, CAD, etc.) than on knowledge of fundamentals. As an aside, it's worth noting that the public image of physics is the Theoretician, but most physics students are Experimentalists.
- lsc 13y ago
  Now, I'm seeing things from the production programming/systems administration side. But my impression was that physics people use large clusters of Linux boxes in college, and thus have practical experience when dealing with that sort of thing in industry. From my perspective, I'd much rather have a physics major who has large-scale systems administration experience and linux experience than a CS major who used java and windows. (But again, I'm not working on "hard CS" - I'm keeping clusters running, with bondo and bailing wire, if need be.) That, and they tend to be bright sorts who are able to break down problems into something that can be reduced to code. My impression is that the standards as far as intelligence goes are higher in the physics department than in the CS department, but that might just be that where I work, I don't get to see the best output of the CS department.
graycat 13y ago
> Wait, what? I mean, you can argue that they are really bad and maybe you are right; I disagree, but maybe I am wrong. I need to get some code written and typed too quickly, but basically I'm correct. So, need to fill in dates on a resume. So, if have years in grad school and some more years in a Ph.D.-only slot as a prof, then how to account for that time on the resume? If put in the truth, then won't get hired. If just leave a gap, then won't get hired. If fill in the gap with the jobs before and/or after the gap or something else, then that's resume fraud. Likely if just omit the Ph.D. in the section on "education", then that's resume fraud much the same as claiming a degree don't actually have. > But until it's in the bag, guessing what is in the mind of VC is... not something I'd bet money on. It took me a while, but eventually I got a reasonably good view of what VCs want. For how I learned to do projects for the US DoD and in grad school, the VCs know essentially nothing: E.g., at one time I was programming in the group that did the orbit determination for the Navy's version of the later USAF GPS (my work was in processing data for passive sonar), and I heard the stories for how the Navy's system was developed. Heavily the plan was a dressed up version of something on the back of an envelope. The work got funded, and has been a great success, from a proposal just on paper. Good for the Navy and US national security. But VCs essentially won't invest in projects described just on paper if only because they don't even want to read the paper -- it may be possible to get some VC seed funding just from paper, but that seems to be recent. NSF and DARPA will read project designs on paper, but the VCs won't. What the VCs want for a Series A round is essentially in one word, traction. Or traction significant and growing rapidly. At one point Menlo Ventures wrote me that the level of traction they want to see is at least 100,000 uniques a month, that is, for a Web site project 100,000 unique visitors. Okay, consider 100,000 uniques, on average: Considering what my site does, 100,000 uniques would mean about 500,000 visits; each visit sees 6 Web pages, with 4 ads per page; and I get paid, in effect, $2 per 1000 ads displayed (drawing from the Mary Meeker data at KPCB). Then my monthly revenue would be 500,000 * 6 * 4 * 2 / 1000 = 24,000 dollars at which time no thanks seed or series A and welcome some new computers and a new car! If the site keeps growing, and once it gets to 100,000 uniques it should, start hiring and grow just organically. > Of course, you don't always need VC to run a company. Yup. My first server might be an 8 core AMD processor at 4.0 GHz with maybe 32 GB of ECC main memory and 12 TB of hard disk in a mid-tower case for about $1500 and software, Windows Server and SQL Server, from the Microsoft BizSpark program. From my software timings, that server might be enough for the $24,000 a month in revenue at which time I can get a $100 wire rack from Sam's Club, put maybe 12 such servers on it. My software is written for such scalability now -- e.g., so far I'm using SQL Server essentially only as a key-value store so that each SQL Server table (there are only a few tables in the SQL Server schema) can be in its own running instance of SQL Server with no communications among the instances; for the various servers, there can be a lot of parallelism via sharding, etc. E.g., I wrote my own session state store; the timings indicate that on a single core running at 4.0 GHz I should be able to do the work needed to send 2000 Web pages a second; such a store is easy to parallelize (partition the collection of possible keys and have one server for each partition); and a few racks of such servers could do enough for session state to serve the world and build a big company. For a seed round, that now can be as high as a few million dollars and no longer limited just to, say, $200,000, the VCs want to see the running software and "play with it". So, soon, maybe later today, more likely before next Sunday, my software should be ready to be played with over the Internet. This software development has taken a while as I learned enough about essentially system administration and Microsoft's .NET Framework to get good with such infrastructure. For the .NET Framework, I found, downloaded, indexed, and studied about 5000 Web pages from Microsoft's MSDN Web site. Yes, my code is awash in comments with tree names on my file system of such Web pages. So, when I'm reviewing code and have a question, e.g., a TCP/IP timeout parameter, one keystroke in my favorite editor pops up the relevant MSDN Web page. In the end, I have found the .NET Framework to be nice, but the 5000 pages took a while. I actually studied all those 5000 pages; at 10 pages a day, that's 500 days just for the MSDN Web pages, and in the end I spent more than that. Heck, I did the research for my Ph.D. dissertation independently in my first summer in graduate school, about 12 weeks, did a second pass later in about 2 months, wrote and ran some illustrative software in about 2 months, and did the final writing and typing in about six weeks. Those 5000 Web pages from MSDN mean that at times I have been known to scream to the point of a sore throat at what I regard as some really poor technical writing at MSDN. In the end, once understood, the .NET Framework is nice and not very difficult to use, but going from the first day with .NET, through 5000 Web pages and maybe a cubic foot of books to some good facility can be a bit much. My whole project is now a surprisingly large number of lines of code (maybe I will add them up tomorrow), but the work unique to my project was easy once the system administration independent of my project was done. E.g., SQL Server installation was an unanesthetized root-canal procedure (once wiped out my boot partition). My software installation got wiped out by a virus in an Akamai download manager I used to get a PDF file from an Asus Web site. There were more such problems, all independent of my work. For my project, I created and derived the basic core applied math in a few days and typed it into Knuth's TeX in a few more days; the software unique to my project has been easy for me; but the routine software part has been grim. The last thing I did, as of last night, was finish the code to have the code of a Web page communicate with a server, in my architecture, a Compute Server, for the core work. So, it's a simple TCP/IP socket application. .NET streams over TCP/IP might have been easier; maybe I rewrote some of that functionality. I avoided such streams because the MSDN documentation seemed too sparse to give me good confidence in what I really might encounter and have to consider, and I didn't want to write and run test cases to clarify just what such streams would and wouldn't do. For what I did write, based just on the basics of TCP/IP sockets, "the workhorse of the Internet", that I've had experience with for a long time, there was little question about how to write the software or what it would do. So, it was easier just to stand on basic TCP/IP than to make sense of Microsoft's documentation of the streams they wrote to use TCP/IP. I put the data to be sent to/from the server code for a Web page and one of my compute servers in an instance of a class, serialize the instance to a byte array, and send the byte array via TCP/IP. > It was actually easier for me to get another well-paid SysAdmin job than to get a job at wal-mart or the coffee shop, even at the depth of the dot-com crash in 2001. Wow, but I can believe it.

by lsc 13y ago

graycat 13y ago

It has appeared that the physics community has done much better getting good non-academic jobs and careers for their Ph.D. holders than most other STEM fields.

analog31 13y ago
Granted, I got my physics degree 20 years ago. At that time, there was actually a lot of frustration in the physics community about the lack of preparation for industry jobs. So I'd hesitate to speculate on why (or even if) the situation has improved. But I will anyway. ;-) I think that physics students aren't always "monogamously" in love with physics, so much as we're intellectual omnivores / opportunists. I was in love with math, electronics, and programming. But I ended up at a college with no engineering school, and the people who seemed to have the most fun playing with toys were my physics profs. But as a result, my ability to roll circuits and sling code with the best of 'em was still marketable in 1993. Today, I see a trend towards hiring people with more formal skills in their job areas, such as CS and EE degrees for programming and electronics design. And I see more emphasis on tools (IDE, CAD, etc.) than on knowledge of fundamentals. As an aside, it's worth noting that the public image of physics is the Theoretician, but most physics students are Experimentalists.
lsc 13y ago
Now, I'm seeing things from the production programming/systems administration side. But my impression was that physics people use large clusters of Linux boxes in college, and thus have practical experience when dealing with that sort of thing in industry. From my perspective, I'd much rather have a physics major who has large-scale systems administration experience and linux experience than a CS major who used java and windows. (But again, I'm not working on "hard CS" - I'm keeping clusters running, with bondo and bailing wire, if need be.) That, and they tend to be bright sorts who are able to break down problems into something that can be reduced to code. My impression is that the standards as far as intelligence goes are higher in the physics department than in the CS department, but that might just be that where I work, I don't get to see the best output of the CS department.

graycat 13y ago

> Wait, what? I mean, you can argue that they are really bad and maybe you are right; I disagree, but maybe I am wrong. I need to get some code written and typed too quickly, but basically I'm correct. So, need to fill in dates on a resume. So, if have years in grad school and some more years in a Ph.D.-only slot as a prof, then how to account for that time on the resume? If put in the truth, then won't get hired. If just leave a gap, then won't get hired. If fill in the gap with the jobs before and/or after the gap or something else, then that's resume fraud. Likely if just omit the Ph.D. in the section on "education", then that's resume fraud much the same as claiming a degree don't actually have. > But until it's in the bag, guessing what is in the mind of VC is... not something I'd bet money on. It took me a while, but eventually I got a reasonably good view of what VCs want. For how I learned to do projects for the US DoD and in grad school, the VCs know essentially nothing: E.g., at one time I was programming in the group that did the orbit determination for the Navy's version of the later USAF GPS (my work was in processing data for passive sonar), and I heard the stories for how the Navy's system was developed. Heavily the plan was a dressed up version of something on the back of an envelope. The work got funded, and has been a great success, from a proposal just on paper. Good for the Navy and US national security. But VCs essentially won't invest in projects described just on paper if only because they don't even want to read the paper -- it may be possible to get some VC seed funding just from paper, but that seems to be recent. NSF and DARPA will read project designs on paper, but the VCs won't. What the VCs want for a Series A round is essentially in one word, traction. Or traction significant and growing rapidly. At one point Menlo Ventures wrote me that the level of traction they want to see is at least 100,000 uniques a month, that is, for a Web site project 100,000 unique visitors. Okay, consider 100,000 uniques, on average: Considering what my site does, 100,000 uniques would mean about 500,000 visits; each visit sees 6 Web pages, with 4 ads per page; and I get paid, in effect, $2 per 1000 ads displayed (drawing from the Mary Meeker data at KPCB). Then my monthly revenue would be 500,000 * 6 * 4 * 2 / 1000 = 24,000 dollars at which time no thanks seed or series A and welcome some new computers and a new car! If the site keeps growing, and once it gets to 100,000 uniques it should, start hiring and grow just organically. > Of course, you don't always need VC to run a company. Yup. My first server might be an 8 core AMD processor at 4.0 GHz with maybe 32 GB of ECC main memory and 12 TB of hard disk in a mid-tower case for about $1500 and software, Windows Server and SQL Server, from the Microsoft BizSpark program. From my software timings, that server might be enough for the $24,000 a month in revenue at which time I can get a $100 wire rack from Sam's Club, put maybe 12 such servers on it. My software is written for such scalability now -- e.g., so far I'm using SQL Server essentially only as a key-value store so that each SQL Server table (there are only a few tables in the SQL Server schema) can be in its own running instance of SQL Server with no communications among the instances; for the various servers, there can be a lot of parallelism via sharding, etc. E.g., I wrote my own session state store; the timings indicate that on a single core running at 4.0 GHz I should be able to do the work needed to send 2000 Web pages a second; such a store is easy to parallelize (partition the collection of possible keys and have one server for each partition); and a few racks of such servers could do enough for session state to serve the world and build a big company. For a seed round, that now can be as high as a few million dollars and no longer limited just to, say, $200,000, the VCs want to see the running software and "play with it". So, soon, maybe later today, more likely before next Sunday, my software should be ready to be played with over the Internet. This software development has taken a while as I learned enough about essentially system administration and Microsoft's .NET Framework to get good with such infrastructure. For the .NET Framework, I found, downloaded, indexed, and studied about 5000 Web pages from Microsoft's MSDN Web site. Yes, my code is awash in comments with tree names on my file system of such Web pages. So, when I'm reviewing code and have a question, e.g., a TCP/IP timeout parameter, one keystroke in my favorite editor pops up the relevant MSDN Web page. In the end, I have found the .NET Framework to be nice, but the 5000 pages took a while. I actually studied all those 5000 pages; at 10 pages a day, that's 500 days just for the MSDN Web pages, and in the end I spent more than that. Heck, I did the research for my Ph.D. dissertation independently in my first summer in graduate school, about 12 weeks, did a second pass later in about 2 months, wrote and ran some illustrative software in about 2 months, and did the final writing and typing in about six weeks. Those 5000 Web pages from MSDN mean that at times I have been known to scream to the point of a sore throat at what I regard as some really poor technical writing at MSDN. In the end, once understood, the .NET Framework is nice and not very difficult to use, but going from the first day with .NET, through 5000 Web pages and maybe a cubic foot of books to some good facility can be a bit much. My whole project is now a surprisingly large number of lines of code (maybe I will add them up tomorrow), but the work unique to my project was easy once the system administration independent of my project was done. E.g., SQL Server installation was an unanesthetized root-canal procedure (once wiped out my boot partition). My software installation got wiped out by a virus in an Akamai download manager I used to get a PDF file from an Asus Web site. There were more such problems, all independent of my work. For my project, I created and derived the basic core applied math in a few days and typed it into Knuth's TeX in a few more days; the software unique to my project has been easy for me; but the routine software part has been grim. The last thing I did, as of last night, was finish the code to have the code of a Web page communicate with a server, in my architecture, a Compute Server, for the core work. So, it's a simple TCP/IP socket application. .NET streams over TCP/IP might have been easier; maybe I rewrote some of that functionality. I avoided such streams because the MSDN documentation seemed too sparse to give me good confidence in what I really might encounter and have to consider, and I didn't want to write and run test cases to clarify just what such streams would and wouldn't do. For what I did write, based just on the basics of TCP/IP sockets, "the workhorse of the Internet", that I've had experience with for a long time, there was little question about how to write the software or what it would do. So, it was easier just to stand on basic TCP/IP than to make sense of Microsoft's documentation of the streams they wrote to use TCP/IP. I put the data to be sent to/from the server code for a Web page and one of my compute servers in an instance of a class, serialize the instance to a byte array, and send the byte array via TCP/IP. > It was actually easier for me to get another well-paid SysAdmin job than to get a job at wal-mart or the coffee shop, even at the depth of the dot-com crash in 2001. Wow, but I can believe it.