I thought the big difference between the GPU and ANE was that you couldn't use the ANE to train. Does the GPU actually perform faster during inference as well? Is that because the ANE are designed more for efficiency or is there another bigger reason?
Sounds like if you have a record of a lot of location/timestamp data for people, you look at the distance difference divided by the time difference. Now you have average speed for any pair of points. Now filter where the average speed is as fast as a Boeing jet. That filters out most of the data except for people who are almost certainly on a plane. Et voila, you now look at those data points geolocation and you have people who traveled from one city to another because you already have the location. Compare City1 -> City2 with any public flights in those cities around those times and you know who flew on what flight from where to where and at what time.
I understand what you're saying about the potential annoyance or dissatisfaction with Java checked exceptions being effectively cast to Runtime ones. As a language choice, it made a signal by differentiating checked versus unchecked and at least gave user the opportunity to benefit from a choice witha nudge in the right direction. You always have to have the escape hatch, but making it less "affordable" to the user, they tend to do the right thing more often which sounds like a perfect win given the trade-offs.
I think the charm for a cloudy place like Scotland is that a system like this is unaffected by poor light supply. Your photovoltaics aren't going to fair nearly so well there hence this solution.
Stuff like this is pretty horrifying and sad after we are deregulating the FDA as we speak so this stuff will be harder to catch and also as we are killing all the science research to help fix this poor soul's eyes after the damage was done. Sad and horrifying.
"You should never, ever get back surgery" sounds like that easily could be wrong. It's annoying to have to always caveat but talking to a (good) doctor is important when making such decisions and not relying on tech forum advice necessarily. I know someone who said they went in for back surgery and walked out feeling permanently cured from the specific problem they had and the pain was completely gone. Details matter. Always consult experts when possible.
"G-Chat with Charleton, in which he would interview Google executives while sitting with them in a two-person snuggie." What a sight that must have been haha
Sending single custom emails is much more effort than bulk-mailing a huge list operationally. Sending bulk can be accomplished by uploading a csv of emails to some enail bulk sender versus code to run at the correct time for the correct user... way easier in bulk and way cheaper
I agree, but will the money be made by the same people? Like, if a psychologists group stands to benefit from treating it they wouldn't reap the benefit of the pharma research. Kind of like how the alcohol industry lobbies against weed or other possible substitutes. Or the car companies buying out the rail roads and then shutting them down to prevent competition. There is just so much inefficiency from competing industries sabotaging each other. You are also making a leap that understanding the cause is physiological means that there is a drug you could sell and not cost billions of dollars to get to market without risk. I always wonder about the PReP drugs for treating AIDS was really the best outcome instead of aiming for the cure. Definitely tradeoffs with a lot of money one way or another.
Strongly disagree. I've seen bad code or maintenance in all the languages. Sure, some ecosystems seem to attract worse practices, but the absolute best software engineering principles I ever learned in my multi-decade career were at a Rails shop. Everything was extremely well maintained, tested, CI/CD were great even before that was a thing. Not to start flame wars about refactoring or what it means to maintain code more, but that place was absolutely full of people who cared and kept those projects at the highest of caliber grade. You can find good, bad, and ugly in every ecosystem but I personally don't think Rails was ever even remotely bad. You can have an amazingly well-maintained setup and still be super productive. It's really about the people and the company spending the time, money, and energy to care. (Edit: typo)
Anecdata, but I remember writing software for a startup and getting thirsty. We had a fridge that had all sugar-free sodas because one of the founders was diabetic. I remember drinking a soda and still feeling really thirsty. Drank another and still didn't solve it. It finally dawned on me that my brain just really wanted the sugar for the easy calories, and the sugar-free soda obviously wasn't fixing this at all. Thinking hard definitely burns calories faster than not.
Yes but there are plenty of companies or people that may want to know how the code works and would be motivated enough to read through the code to understand it and having it there in the public makes that possible.
Why is this de-facto the case for so many things? Is it a matter of finding a lawyer? Do these enforcement agencies just not care or don't have enough resources to battle Meta? It's exhausting that this gets repeated so often but for no apparent reason. I understand that the fine might not be large enough to stop them but it feels like all these things just get swept under the rug and gross violations of privacy which are against the law never get actioned. (Edit: forgot a word)
I feel privileged to be able to say that as a software engineer who's been doing it the hard way for 20+ years, I relish the hard problems. The CRUD app updates are unbearable without the random in-between challenges that bend my mind. The rare recursive algorithm, the application of some esoteric knowledge I actually learned in college, actually having to do big-o estimates. These are the gems of my career that keep me sane. I hope the next flock of AI-driven SWEs appreciates these things even more given the AI can spout off answers which sometimes are right and sometimes are horribly wrong. Challenges like these will always have to have someone who actually knows what to do when the AIs start hallucinating or the context of the situation is beyond the context window.
That's my point, the type errors being reduced isn't that interesting, I usually find and fix those extremely quickly. What I don't find but am most interested in is the actual logic. So I must write tests. And when I write tests, I get the type errors covered too. So why should I spend my precious time modeling inside a type system when the ROI is minuscule and the cost is I have to litter my codebase with types. Just write the tests.
Can you explain concretely how adding types would make a bunch of boolean-valued functions and some tests better? What types would you introduce to my example?
I neither used the word legacy nor modern. I don't care how fancy your type checker is today or tomorrow, the arbitrarily deeply complicated boolean-typed function I write tomorrow gleans nearly zero value from type validation. It gains actual validation of correctness with real tests.
It's not circular reasoning, I literally have piles of boolean-typed functions in business logic and there is no quantity of type-checking that will validate that (defn should-go [light-is-green? intersection-is-clear?] (and light-is-green? intersection-is-clear?)) is correct. That requires an actual test. Obviously my example is trivial but they quickly become non-trivial and the type-chrcker will never save me if I replace "and" with "or".
Seems like we shouldn't be using tech to punish a foreign group who's job, at least at face value, is to help the world. If the ICC is doing so much harm to the US, fight legally. That's where the battle should be fought. Not ripping away some guy's email access.
If I have a complicated set of business-logic functions that all return booleans, the type-checker can tell me I got a boolean back but not that in xyz scenario my result should be true and abc scenario it should be false. You must write a test for that, the types are only the most marginal at helping check correctness. There are many more examples like this. If I have to write the tests, why do I want to spend time writing all the types too if my tests cover the things I care about the the types only cover details like "you returned a number from a function marked boolean", that type of error is super easy to catch with a test, so I didn't need the types.
Just goes to show how short-sighted these political decisions have been. Politics shouldn't be wrapping itself around tech, the US is shrugging off a huge market and ally.
If you do it right, your tests cover the thing you care about and are less work usually because you write the test for the part of code that has the interesting part you want to verify works correctly. Types, especially on fully typed languages, make only guarantees about the type of values and they make you do it in places you might not care about. Types also only catch certain classes of errors that I personally find less helpful that tests. So to me, tests are less work, focus on thebimportant parts, and give better error messages. Obviously you can also write bad tests and do things like effectively just check the types in pmaces and you are back to square one. Testing is more of an art but the effort/economics of time are far better to me.
In general, it's a dream to test and refactor Clojure IMHO. Working with mostly static top-level functions, immutability by default, and mocking when necessary with the use of "with-redefs" makes writing tests easy. Also, the immutability makes it hard to break shared copies of memory which seemed to plague my earlier tangos with large Java code bases and standard OOP practices.