2%? You have an interesting idea of the world's population.
Just think about what that means. It means 2 out of 100 people can hack into Facebook's Legal Department Admin Panel.
I mean if we are talking "mentally capable to achieve that within a decade if the person does nothing else but strive to that goal"... Perhaps.
If we are talking "sit down right now and do it", then it's more like what... 10,000-100,000 people on earth? Which makes for more like 0.0014%?
Not quite. The US alone graduates 2 million Computer Science students of various stripes every year. It's been graduating (smaller numbers) of them for over 40-50 years now. There are now second and third generation comp sci. workers and graduates.
So let's say 1% of 1 million/year are up to this, I suspect it's rather more, but I can't be bothered to do the curve on past graduation rates, and figure out what the world wide figure is... you've easily got a couple of million people world wide.
When I think it's going to get really interesting is in another 10 years or so when there start to be significant numbers of bored retired former developers. At any rate the market rate probably isn't that bad.
I think you might have looked at the wrong statistic when googling this. According to this site[0] (which is one of the first hits for "computer science graduates per year"), there are 2 million computer science people in the workforce _in total_, which seems far more realistic. Actual number of graduates per year seems to be 65000.
With your numbers (assuming linear growth) after those 40 years, about one third of the total US workforce would now be CS graduates.
Driven a lot by intentionally deceitful marketing too IMO. The PR person at the company behind whatever storage project only quotes those numbers in these terms (enough to power 50,000 homes) because they sound impressive, and as you say the journalist doesn’t know better so they then print it like that.
ML will become a library. It has about as much to do with programming as a compiler. You don't need to know what it does, you just need to know how to make it do things. The problem with ML currently is that nobody really knows how to do things and that you have a million parameters that need tuning and most algorithms need continuous improvement and fine tuning to the use case. There is nothing "mainstream" about ML at this point, except that everyone wants to use it.
In maybe a decade, it might be found in standard libraries of programming languages and on top of things like `Math.abs`, we will have `ML.textToSpeech("Hello world")`, or `ML.isCat(image)`, etc. However, the problem I see with that is that no matter how far we wind the clock forward, we will only be able to put the most simplistic use cases into a library. `ML.isCat()` could be one of those, since most humans will be able to image categorization, it stands to reason that you could put this into a library. However, most industry application involved highly customized ML algorithms that are optimized for a very specific use-case. So there will always be a need for a research team in big companies at least. Maybe smaller companies will try to build their stuff by chaining libraries together.
There's never going to be a `ML.isCat(image)`, just like there isn't a `Math.solveProblem(hypothesis)`. Yes you do have `Math.abs` and you're going to have stuff `model.fit()` and `layers.dense()` - but something like `ML.isCat` is too specific to be used in a library
Disagree. In the future, that'll be `npm install ml-cat` followed by `MLCat = require('ml-cat'); MLCat.isCat(image)`
It might not be npm, but something like that is probably inevitable.
The reason it seems so unlikely is because the tooling isn't there yet. No one even agrees how ML code should look, let alone how libs should be distributed to end users. But I saw the transformation for JS in 2008.
the range of problems you can solve with ML/AI is simply too wide for there to be fully-canned solutions for everything. Sure, there will be canned solutions for _some_ things - maybe even for cat detection, because it's fun so why not.
But, a library that uses AI to optimize the production of your business' flux capacitors? Ain't gonna happen, you need to build that yourself. To have a library/product that solves problems using AI, you need a "language" to describe the problem (like you can e.g. use SQL to describe any data query you may have). But describing problems is notoriously hard - accurately & precisely describing the problem is very often just as hard as solving it.
Mm, it's a bit like arguing that "the range of text editor customization is simply too wide for there to be fully-canned solutions for everything." Meanwhile, elisp wiki go brr.
I think ML solutions will increasingly take the form of an elisp script rather than a python library, but it'll take a little while to get there.
> it's a bit like arguing that "the range of text editor customization is simply too wide for there to be fully-canned solutions for everything."
But the the range of editor customization really isn't that wide. That's exactly what I'm arguing, that ML/AI is more like "math" than like "editor customization".
Fwiw macs have had an equivalent functionality for both text to speech and speech to text for at least 17 years to my memory. The quality is poor compared to today's server-driven approaches, of course, but the functionality has been there if you're willing to articulate yourself clearly.
Actually, code isn't what does that. It's true that code is one of the things that keeps it "all". However, it's not the best place for documentation or even storage of that knowledge. Code changes.
Tests are what you need to invest in. When you test all of you business use cases in an abstract (perhaps even no-code) way, you are truly independent. You can rewrite the system very quickly, if you had a test-suite that allows for quick iteration (TDD style) and does not depend on implementation details.
Given just the code? Well good luck migrating to another cloud vendor. You will probably introduce one million bugs on the way.
You'll want to test against stable API/ABI, and use positive and negative testcases. The problematic parts are mocking, simulation and how tests affect the codebase.
- The only stable APIs I've seen, if an API exists, are production financial APIs, because money is involved.
- DynDNS has been knows to change input and output parameter types, breaking calls
- Even Twilio has changed a fundamental API path (!) in the past couple of years, breaking SMS API calls in 2019
- Facebook only supports API version N and N-1. Hope you're not on N-2 and there's a sudden version bump!
(I say production APIs, since even payment gateways often have flaky dev/qa gateways that you can't reliably do automated tests against. Past companies that I worked at had to test against prod gateways with their own personal credit cards.)
Otherwise, if you want an API, ensure you choose a product/partner that offers/supports one.
It's the "American Dream". They need to reiterate again and again how every dishwasher can become something if they just work hard enough lol. It's these things that make me pity Americans, but other stuff offsets it as well, so there we go.
Almost anyone who wasn't born into money did some menial task around high school or college. I worked at UPS as package sorter, now I earn 40 times as much at FAANG.
But that had nothing to do with me working at UPS. I worked there because I liked the extra money on top of what my parents gave me and to me it was like getting paid for gym :D.
> But that had nothing to do with me working at UPS. I worked there because I liked the extra money on top of what my parents gave me and to me it was like getting paid for gym :D.
I'm happy you ended up with a great job that pays well. Congrats to you!
But I think you underrate what and how much you may have or could have learned from working at UPS. UPS is a fantastic logistics operation, and while being a package sorter is obviously not the same as being a VP, the job still exposes you to a system that's highly optimized for profiting from being good at logistics, and that reaches every job in the system. Maybe you didn't learn anything at the job, but that doesn't mean there was nothing there to learn for someone observing and trying to learn.
I mean if we are talking "mentally capable to achieve that within a decade if the person does nothing else but strive to that goal"... Perhaps.
If we are talking "sit down right now and do it", then it's more like what... 10,000-100,000 people on earth? Which makes for more like 0.0014%?