The courses are pretty far lagging the state of the art.
Watch every youtube video by Hinton, LeCun, etc. Read the fundamental papers.
Once you understand the ideas, write a simple NN program. It's like 100 lines of python/numpy. Train in on MNIST. Compare your results with the published results. Understand where your code is failing. Try to make it better. Get a 2-layer RBM to actually learn better than a 1-layer RBM.
That's like two months of evenings total work. Do that, and you can at least know if you like the field and have any hope of understanding it.
I want to be able to pick a page out of "The Art of Computer Programming" and have a prospective programmer explain what the page is talking about and why (or why not) the concept mentioned would be used.
I love that idea! I'm going to try it on the next 5 people I interview.
Google, Facebook, General Motors, Exxon, and 100 other firms need 10,000 people who understand this technology right now. There are current about 500 people who have a clue.
You start playing with this and ask a smart question on Google search... the magic "you look like a person who might be a Google fit" will appear.
If you can finagle it, try getting into a management position. Sounds like you have plenty of experience to base your campaign off of. You'll be better compensated and have a lot more upwards mobility.
Too many people mistake experience for competence. The OP has "just reached a senior level in a tech career and I've been doing pretty much a bit of everything, e.g. software architecture, full stack dev, eng. related specific dev, consultancy, etc." The question should be: is he actually an expert at anything? If yes, then he has nothing to fear from downturns. If no, he's going to be out of a job in the next downturn.
So should he try for management? Well, if he has no real skill at what he's been doing, a management role where he's unskilled and learning the ropes won't help him.
Right now, gas prices are relatively low, but they are rising, and oil will be back in the triple digits soon enough, almost definitely by Memorial Day.
Then you can make a ton of money right now by buying WTI futures or options. The consensus Memorial day price is under $60 - you can clean up to the tune of 1000%s of profit if you put money on your "almost definite" knowledge.
Can they do it with corporate code where there are naming and style standards in abundance, and code reviews to ensure those guidelines are followed?
I was starting to wonder about that, then realized we at $BIGCORP are already generating ASTs from your input buffer, unifying those trees with a bunch of patterns, and telling your editor to flag questionable constructs. You type "if not foo in x" and 50ms later you get a proposed improved snippet. It's pretty rare to see quirky style in our codebase.
"C++ is a three-way compromise between good object oriented design, backwards compatibility with C, and high performance. Stroustrup has never billed it as anything else."
Yeah, so when Bjarne wins the "sucks cocks for money with a smile, never billed it as anything else" award, we should all cheer?
" As long as you don't lick the walls, you can't get any radiation down there. "
Wow, a link to a SciAm article. Let me go read the source. Oh dear, it says absolutely nothing about licking walls. Guess that was just the submitter making stuff up.
"Absolutely and unambiguously make writing and publishing a zero-sum game"
Um, no - the more readers, the more money. It's not zero sum at all from the writers' point of view.
Of course, back in the old days, people often curled up in a chair and read eight good books simultaneously; writers didn't compete with each other for readers' time and dollars at all.
For the best reasons to learn Python, see The Zen of Python. If Python happens to pay more, that's just gravy.
That said, it seems hard to believe that people would get paid extra to work in such a pleasant language. If so, maybe Adam Smith had it all wrong when he said:
First, The wages of labour vary with the ease or hardship, the cleanliness or dirtiness, the honourableness or dishonourableness of the employment...The most detestable of all employments, that of public executioner, is, in proportion to the quantity of work done, better paid than any common trade whatever.
Read on a bit more. By paragraph 10 he points to increased wages for jobs requiring skill, by paragraph 20 he's getting into jobs requiring trust.
Pity he living too soon to comment on large software project laborers.
Software developers help companies make more money. It is the Add in Value-Add. They are the equivalent of the machines in a machine shop. Without them, what is the point in being in business. If you are a software company you pay what you need to pay, to recruit and retain the best developers you can.
Most software developers are not in pure software development companies. They are in large companies doing something like fortune-500 stuff or selling ads (Google) or moving goods (Amazon.)
Very few companies think "let's hire more developers, they add value!" Hiring a developer is a last resort when the tech you have doesn't do what you need. It's like needing to hire a lawyer - you don't want to do it, but it's the cheapest way to achieve your goal.
For most blocks, they just strapped four quarter-circle cradles around the stone and rolled them up earthen ramps using ropes. The remains of the ramps still exist around some pyramids, and some original cradles are on display in the Cairo museum. Pretty much considered solved by the archeologists; it's just armchair physicists who want to invent problems and propose new solutions.
Minimum Wage is also an attempt to keep the employer/employee relationship decent.
You could have a business model in which you maim small children so that they can earn more money while begging for you, but we, as a society, have decided that it is a bad thing. Yes, it happened in Victorian England and present day India, but we don't do it, even if it is "optimal" under free market conditions.
So, requiring you pay a person enough to live a decent live might not be that bad of an idea. If your business model can't support it, maybe you shouldn't be in that business.
I can understand that. "Take something that you do on a computer in a structured environment with constructive tools and then draw it on a whiteboard, while talking out loud, to a bunch of strangers." Impossible. Frankly, I can't write and talk at the same time, let alone try to code on the fly without a computer. I'm trying to imagine an interview for a guitarist where they say, "Why don't you walk up to a whiteboard and draw out how you'd play some song you've never heard of."
Good analogy. If you were a great guitarist, and you were looking to hire a great guitarist, that question might be reasonable. The two of you would start talking and pretty soon determine if you have a mutual fit.
Same thing with code. Maybe you think "can get the job done" is good enough. Maybe they expect symphonies, though. Right or wrong, you might not be a good fit. Heck, we expect serious CS from all our potential hires, and sometimes one stops the recruiting process because he figures we are weaker than him. Totally fair - it's not about getting the whiteboard problem right or wrong, it's about the talk between people who are going to potentially work together.
1. The infinite series: some people know the basic one of the top of their heads: 4 * alternating odd fractions. Some may even know some Ramanugen (spelling - Indian math genius) better series. 2. Some people may remember there is an algorithm to compute the nth digit of Pi efficiently - ask to do a web search for the state of the art. 3. Just go to a trusted website that has already listed Pi to a bazillion digits and pluck the digit out.
The axe has been with us for thousands of years, with its design changing very little during that time. After all, how much can you really alter a basic blade-and-handle?
Well, a simple Google image search for "axe catalog" shows 42 different axe heads sold by the Shapleigh company in 1929.
Be careful putting too much faith in almighty science. They've been wrong before, you know. A lot. And people died because of it.
You show a bunch of ideas that, when exposed to science, got shot down as objectively wrong pretty quickly. Sounds like the process works.
Want to list 6 current sciency ideas that are wrong but the scientific community considers reasonable? I'll give you a few to start you off:
1. Humans are not changing the climate. Current verdict: wrong. Supporters: a few loons. Evidence: about nil. 2. Evolution is wrong. Current verdict: wrong. Supporters: a few loons. Evidence: nil. 3. Vaccines cause autism. Current verdict: wrong. Supporters: a few loons. Evidence: nil.
I'm sure Slashdot2114 will be debating the bad science ideas that existed in 2014. Some will claim history shows science is death. Smarter people will note that imbeciles, public relations people, lobbyists, and trolls have always added noise and generally slowed the dissemination of knowledge.
This is an impressive step forward in image processing - while reconstructing an image from diffuse light seemed plausible in theory, figuring out how to do it in practice is a hard problem. These guys deserve some respect.
Well, some respect, but it's hardly cutting edge or even very new. Maybe for physicists, but CS was ahead.
Kohonen described the basics of correlated reconstruction back in the 1980s.
There were videos of reading the backs of cards from diffuse lighting by the early 2000s. Admitted using some cheats like controlling the light source, but not awful compared to this paper that restricts the color.
By the late 2000s, the ideas were pretty common and computationally feasible. I even wrote a few POCs myself while working on somewhat related optical stuff.
>And for what it's worth, in the last few places I've worked, the multimillionaire bosses have always sat right in the middle of the open plan with everybody else
I bet they didn't write much code.
You'd lose that bet at my workplace. The MMBs are in the middle of the open plan and are the top 1% coders: that is why they are there.
Slashdot Mirror
The courses are pretty far lagging the state of the art.
Watch every youtube video by Hinton, LeCun, etc. Read the fundamental papers.
Once you understand the ideas, write a simple NN program. It's like 100 lines of python/numpy. Train in on MNIST. Compare your results with the published results. Understand where your code is failing. Try to make it better. Get a 2-layer RBM to actually learn better than a 1-layer RBM.
That's like two months of evenings total work. Do that, and you can at least know if you like the field and have any hope of understanding it.
We have a winner. Congrats, sir!
I want to be able to pick a page out of "The Art of Computer Programming" and have a prospective programmer explain what the page is talking about and why (or why not) the concept mentioned would be used.
I love that idea! I'm going to try it on the next 5 people I interview.
Good analysis of the landscape.
Google, Facebook, General Motors, Exxon, and 100 other firms need 10,000 people who understand this technology right now. There are current about 500 people who have a clue.
You start playing with this and ask a smart question on Google search... the magic "you look like a person who might be a Google fit" will appear.
And me: I first found it as a hardcover book "a handbook of integer sequences" about 25 years ago.
If you can finagle it, try getting into a management position. Sounds like you have plenty of experience to base your campaign off of. You'll be better compensated and have a lot more upwards mobility.
Too many people mistake experience for competence. The OP has "just reached a senior level in a tech career and I've been doing pretty much a bit of everything, e.g. software architecture, full stack dev, eng. related specific dev, consultancy, etc." The question should be: is he actually an expert at anything? If yes, then he has nothing to fear from downturns. If no, he's going to be out of a job in the next downturn.
So should he try for management? Well, if he has no real skill at what he's been doing, a management role where he's unskilled and learning the ropes won't help him.
Right now, gas prices are relatively low, but they are rising, and oil will be back in the triple digits soon enough, almost definitely by Memorial Day.
Then you can make a ton of money right now by buying WTI futures or options. The consensus Memorial day price is under $60 - you can clean up to the tune of 1000%s of profit if you put money on your "almost definite" knowledge.
Can they do it with corporate code where there are naming and style standards in abundance, and code reviews to ensure those guidelines are followed?
I was starting to wonder about that, then realized we at $BIGCORP are already generating ASTs from your input buffer, unifying those trees with a bunch of patterns, and telling your editor to flag questionable constructs. You type "if not foo in x" and 50ms later you get a proposed improved snippet. It's pretty rare to see quirky style in our codebase.
"C++ is a three-way compromise between good object oriented design, backwards compatibility with C, and high performance. Stroustrup has never billed it as anything else."
Yeah, so when Bjarne wins the "sucks cocks for money with a smile, never billed it as anything else" award, we should all cheer?
Oh, so it was a quote from James Conca, a nuclear power booster, not Scientific American. That's just plain dishonest.
" As long as you don't lick the walls, you can't get any radiation down there. "
Wow, a link to a SciAm article. Let me go read the source. Oh dear, it says absolutely nothing about licking walls. Guess that was just the submitter making stuff up.
If the pool is greater than $0, it's a positive sum game for the authors.
If Amazon behaves economically rationally, the pool size should increase with the number of readers.
"Absolutely and unambiguously make writing and publishing a zero-sum game"
Um, no - the more readers, the more money. It's not zero sum at all from the writers' point of view.
Of course, back in the old days, people often curled up in a chair and read eight good books simultaneously; writers didn't compete with each other for readers' time and dollars at all.
For the best reasons to learn Python, see The Zen of Python. If Python happens to pay more, that's just gravy.
That said, it seems hard to believe that people would get paid extra to work in such a pleasant language. If so, maybe Adam Smith had it all wrong when he said:
First, The wages of labour vary with the ease or hardship, the cleanliness or dirtiness, the honourableness or dishonourableness of the employment...The most detestable of all employments, that of public executioner, is, in proportion to the quantity of work done, better paid than any common trade whatever.
Read on a bit more. By paragraph 10 he points to increased wages for jobs requiring skill, by paragraph 20 he's getting into jobs requiring trust.
Pity he living too soon to comment on large software project laborers.
def fastFactorPrimeNum(primenum):
"""quickly factor a prime number"""
return 1, primenum
Happy to help. Only copyright stands in the way of breaking cryptography forever.
Software developers help companies make more money. It is the Add in Value-Add. They are the equivalent of the machines in a machine shop. Without them, what is the point in being in business. If you are a software company you pay what you need to pay, to recruit and retain the best developers you can.
Most software developers are not in pure software development companies. They are in large companies doing something like fortune-500 stuff or selling ads (Google) or moving goods (Amazon.)
Very few companies think "let's hire more developers, they add value!" Hiring a developer is a last resort when the tech you have doesn't do what you need. It's like needing to hire a lawyer - you don't want to do it, but it's the cheapest way to achieve your goal.
For most blocks, they just strapped four quarter-circle cradles around the stone and rolled them up earthen ramps using ropes. The remains of the ramps still exist around some pyramids, and some original cradles are on display in the Cairo museum. Pretty much considered solved by the archeologists; it's just armchair physicists who want to invent problems and propose new solutions.
Minimum Wage is also an attempt to keep the employer/employee relationship decent.
You could have a business model in which you maim small children so that they can earn more money while begging for you, but we, as a society, have decided that it is a bad thing. Yes, it happened in Victorian England and present day India, but we don't do it, even if it is "optimal" under free market conditions.
So, requiring you pay a person enough to live a decent live might not be that bad of an idea. If your business model can't support it, maybe you shouldn't be in that business.
I can understand that. "Take something that you do on a computer in a structured environment with constructive tools and then draw it on a whiteboard, while talking out loud, to a bunch of strangers." Impossible. Frankly, I can't write and talk at the same time, let alone try to code on the fly without a computer. I'm trying to imagine an interview for a guitarist where they say, "Why don't you walk up to a whiteboard and draw out how you'd play some song you've never heard of."
Good analogy. If you were a great guitarist, and you were looking to hire a great guitarist, that question might be reasonable. The two of you would start talking and pretty soon determine if you have a mutual fit.
Same thing with code. Maybe you think "can get the job done" is good enough. Maybe they expect symphonies, though. Right or wrong, you might not be a good fit. Heck, we expect serious CS from all our potential hires, and sometimes one stops the recruiting process because he figures we are weaker than him. Totally fair - it's not about getting the whiteboard problem right or wrong, it's about the talk between people who are going to potentially work together.
I wouldn't.
There are three basic ways to solve this problem:
1. The infinite series: some people know the basic one of the top of their heads: 4 * alternating odd fractions. Some may even know some Ramanugen (spelling - Indian math genius) better series.
2. Some people may remember there is an algorithm to compute the nth digit of Pi efficiently - ask to do a web search for the state of the art.
3. Just go to a trusted website that has already listed Pi to a bazillion digits and pluck the digit out.
Well, a simple Google image search for "axe catalog" shows 42 different axe heads sold by the Shapleigh company in 1929.
So, the answer would seem to be "quite a lot."
Which track record is that?
Be careful putting too much faith in almighty science. They've been wrong before, you know. A lot. And people died because of it.
You show a bunch of ideas that, when exposed to science, got shot down as objectively wrong pretty quickly. Sounds like the process works.
Want to list 6 current sciency ideas that are wrong but the scientific community considers reasonable? I'll give you a few to start you off:
1. Humans are not changing the climate. Current verdict: wrong. Supporters: a few loons. Evidence: about nil.
2. Evolution is wrong. Current verdict: wrong. Supporters: a few loons. Evidence: nil.
3. Vaccines cause autism. Current verdict: wrong. Supporters: a few loons. Evidence: nil.
I'm sure Slashdot2114 will be debating the bad science ideas that existed in 2014. Some will claim history shows science is death. Smarter people will note that imbeciles, public relations people, lobbyists, and trolls have always added noise and generally slowed the dissemination of knowledge.
Where do you stand, PR Man?
This is an impressive step forward in image processing - while reconstructing an image from diffuse light seemed plausible in theory, figuring out how to do it in practice is a hard problem. These guys deserve some respect.
Well, some respect, but it's hardly cutting edge or even very new. Maybe for physicists, but CS was ahead.
Kohonen described the basics of correlated reconstruction back in the 1980s.
There were videos of reading the backs of cards from diffuse lighting by the early 2000s. Admitted using some cheats like controlling the light source, but not awful compared to this paper that restricts the color.
By the late 2000s, the ideas were pretty common and computationally feasible. I even wrote a few POCs myself while working on somewhat related optical stuff.
It goes way eventually. Then you join senior senior management and you will have the airport dream and the moving to a new house dream.
Welcome to eternal nightmares.
>And for what it's worth, in the last few places I've worked, the multimillionaire bosses have always sat right in the middle of the open plan with everybody else
I bet they didn't write much code.
You'd lose that bet at my workplace. The MMBs are in the middle of the open plan and are the top 1% coders: that is why they are there.