Agreed, it's a matter of degree. You can auto-scale up to a certain amount, at least in Heroku, so in practice you can limit it, say up to 8 servers or something, but nothing beats your own box if you can afford the opportunity cost.
Can't think of any, but managed hosting tends to do better. For instance if you are doing your testing in Ruby, I'd suggest hosting on Heroku while you do all your preliminary work, where it's clear(er) what the cost will be, then move to AWS later when the bank account is, well, beefier. :)
Another beautifully designed piece of Emacs software is magit, where the defaults are extremely sane. The package is self-documented and you can learn it all by typing ? anytime. Being able to git log the file you are in is also a huge plus.
This is already being done in places such as the university of Arizona (Chomsky and his former students). The subject is narrower of course (computational linguistics and some neuroscience), but there are taking an approach that is more Galilean in nature, by designing experiments that reduce externalities rather that simply looking at massive amounts of data. I think that's what's going be the most useful, at least in areas that continue to be challenging for the current trends in AI, namely language.
Genuinely curious if this has been done, but considering that a $1000 is not a lot of money which suggests you are not running a big operation, why not use it and tie the rest to a card with a limit such as from privacy.com? Granted Amazon might deny such cards, in which case, if you really want it, you can set up a new debit card whose account has say a maximum you can spend?
I agree with you on most points, and they are working hard to fix the generics issue in a way that does not make you lose all the nice things you mentioned. The only part I didn't get is:
> The use of zero values is a crime against nature
Depends. If it's a bug or a new feature built on top of code that is already tested, then I will write the tests because they fit nicely into already existing testing infrastructure. The test therefore "proves" the bug, and the fix eliminates it.
For outright new code (think new objects, new API, and so on), I tend not to write the tests first because they become a cognitive load that affects my early design choices. In other words, I am now writing code to make the tests pass, and have to exert effort not to do that.
This always comes up, so I am curious if one fixes for "where the Tesla is driven" when doing safety analysis? Considering how difficult it is to get a Tesla (and for older models, how expensive), surely if one fixes for location, such as suburbs and well designed metropolitan areas as well as the type of people that drive Teslas, then perhaps other vehicles will do just as well?
I am by no means saying Tesla is less safe, in fact I do believe these claims, I am just not sure if they are truly thorough.
Well, I am not sure if you've come across some of Bertrand Russell's work, but he qualified Mathematics as having what he called "supreme beauty". The older I get, and I am not a mathematician by any stretch of the imagination, the more I get what he means. This stuff is surely a gift for all of us.
I think there is a middle ground. Apps could announce the upcoming railroad crossing just like they do with a left/right turn. Somewhat of a silly example, but if there is a river at the end of a road and your only option is a left or a right, the app will not help stop you for driving the car into the river, but the instruction to go right or left does (albeit accidentally) help you avoid that. I don't see why it won't do the same with rail crossings.
Although you should focus on road signs more than you do the app, there could be a dangerous situation in which the app says "make a left turn at the light", but before reaching the light there is a railroad crossing. Of course you need to stop at that, but the app can enhance the situation but treating it like another instruction. I see no problem here but an overall safety enhancement.
Your comment reminds of this wonderfully written post by Roger Ebert[1]. It's not the same topic, but there is something to be glanced at in his writing where he eludes to choice, editorial-ism, and doing a story regardless of what the audience wants.
Sure, I see what you mean, but he is trying to help people who are navigating "machine learning", whatever that is. Your argument is the same as "well, why did they call it gravity, shouldn't they have called it attraction?" It is what it is, and you teach based on the term already in the literature.
That's certainly true, you are making the leap of faith that your customers are running standard compliant browsers, which, all considered, is true. All the techniques you mention require you to break either the customers or some element along the network chain to the server, and once you do, well, CORS is the least of your problems.
It also creates huge discrepancy between what a developer using an ORM like ActiveRecord writes and how the engines end up optimizing, meaning that should you want to write a performant one, you have to learn both, or, short of that, notify your DBAs of what you did, and in both cases huge resources are wasted. I never understood combining both from a business perspective anyway.
What do you mean the AC units are not allowed in Western Europe? I lived in Germany and never heard of this. In my experience, it's mostly cultural. The French for instance expect a "canicule" once or twice a year and that's generally survivable, although I think they are realizing that their frequency and power are increasing, so we might see uptake in air conditioning.
Edit: Or perhaps you meant regulations? If so, then yeah those exist, such as permits for the building and so on.
What's the definition of "fair" here? There are many scenarios under which that makes no sense. Is the price inflation adjusted? What if the currency in question tanked in the mean time? What about actual labor done on the copy (annotations, notes, ...). Not faulting Microsoft, but we really need to rethink this model.
Any examples? Perhaps you mean either Generics or the module system (the two elephants in the rather large and accommodating room)? I read almost all proposals for Generics, and, perhaps to my inexperienced eyes, the problem seems hard. The ramifications on the language design are not trivial, and the benefits, again, from my quite limited world view, are not so easy to quantify, otherwise how can one justify the extreme productivity spawned by the language? In the cloud business no less.
I have almost no stake in this, I don't even use the language professionally, but I find the supposed criticism a bit over the top. I'd wager that an example implementation of Generics that fits nicely with the language and works for a big and healthy code base would absolutely not go unnoticed by Russ and the other core members, but we have to accept that it is not easy.
I think the definition has evolved. Since usually the DB is reset between tests (and such reset is snappy), it's transparent enough to appear as unit testing. That we do this in a buggy manner or do not understand how to properly reset the DB does not negate that fact in my opinion. You could mock it or inject it, therefore doing unit testing in the traditional sense, but again you could introduce even nastier bugs and a whole lot of indirection overhead.
For me, having to put it on a credit card is effectively not being able to cover it, so the numbers do check out, from this perspective. Again, we are talking an emergency, not a mortgage.