mgraczyk

52 ms·

751.
by mgraczyk 4y ago
If you have never worked at a large tech company like Microsoft, you'll probably have a bad feeling because there's a lot you don't know about the business process of shipping features like this. It's reasonable to be cy
752.
by mgraczyk 4y ago
As a user of software, I don't care about how reasonable the optimization is, I care about how responsive the software is. Code that feels fast but is poorly optimized is fine. The problem is often that developers don't know how t
753.
by mgraczyk 4y ago
Wouldn't the exception actually come from throwIfFailed? Or does it come from resultNow? How does the tool show it as corresponding to the call to findUser?
754.
by mgraczyk 4y ago
If Java adds some nice standardized helpers like this, they will look equivalent. The current proposal is not this clean but that doesn't mean it won't be possible. The key difference is that async/await implies cooperative m
755.
by mgraczyk 4y ago
With virtual threads, you need to write fork/join code to do two subtasks. With async await, you call two async functions and await them. So the virtual threading model ends up requiring something that looks like a worse version of asy
756.
by mgraczyk 4y ago
How does the tooling know to hide the call to "fork" in the scoped task example?
757.
by mgraczyk 4y ago
I mean what the article calls a "synchronous blocking method", which javascript (mostly) does not have.
758.
by mgraczyk 4y ago
I think the async/await debugging experience is easier to understand. For example in the structured concurrency example, it seems like it would require a lot of tooling support to get a readable stack trace for something like this (in
759.
by mgraczyk 4y ago
I agree with your point that for CPU bound tasks, the threading model is going to result in better performing code with less work. As for the point about locks, I think this one is also a question of IO-bound vs CPU bound work. For work tha
760.
by mgraczyk 4y ago
The first part is semantics, yes I understand that python is running one OS thread at a time with a GIL (for now). Just pretend I used the word "concurrent" instead of parallel in all the places necessary to remove the semantic di
761.
by mgraczyk 4y ago
This makes sense, especially the bit about tooling. I'm unfamiliar with the state of Java tooling besides very simple tasks. On the other hand using things like debuggers and reading stack traces in python/js "just work"
762.
by mgraczyk 4y ago
Two separate threads run in parallel, but one thread cannot do two subtasks in parallel without submitting parallel jobs to an executor or a StructuredTaskScope subtask manager. It's basically forcing the developer to do all the hard w
763.
by mgraczyk 4y ago
How do you implement doThingOne? You should read the "structured concurrency" link in the article. You have to explicitly wrap the call to doThingOne in a future under a structured concurrency scope. The code example you wrote is
764.
by mgraczyk 4y ago
There are a few reasons. async/await allows you to do multiple things in parallel. I don't see how you can do that in the virtual threading model, although I haven't used it and only read this article. You would have to spin
765.
by mgraczyk 4y ago
The section "What about async/await?", which compares these virtual threads to async/await is very weak. After reading this article, I came away with the impression that this is a dramatically worse way to solve this pro
766.
by mgraczyk 4y ago
Well I don't work on ads or on any product that has or will ever have ads, but... It's really the other way around. I work at Google because I have no problem with ads. It's another example of opportunities available if you a
767.
by mgraczyk 4y ago
I have the same TV and have literally never thought about the ads on it except when reading articles on Hacker News. It seems like a huge benefit these days to be the sort of person who just isn't bothered by ads, like you can enjoy pr
768.
by mgraczyk 4y ago
Where? That's not in the article. This is not talking about actual brains or synapses, if that's what you're thinking. It's just using the same words but these systems are only "inspired" by actual neurons, the
769.
by mgraczyk 4y ago
Where are you seeing "memory enhancement" or "memory augmentation"? This article is about analog computation hardware, which can potentially solve computer vision tasks with far less power than typical hardware with digi
770.
by mgraczyk 4y ago
Since the issue was in both the iOS and Android versions of the app, and it was caused by an integer overflow, does that mean that the bug was in a bundled C++ library implementing webRTC? Is there any information about the source-level cau
771.
by mgraczyk 4y ago
This is more secure overall. The problem with short sessions is that users who do not think about them sign in more often. This makes typing your credentials more normal and makes phishing attacks more common. Reducing the rate at which use
772.
by mgraczyk 4y ago
Every photo sharing and storing site compresses photos for view. The originals are still there (read further down in the issue).
773.
by mgraczyk 4y ago
Hi, I work on the processing algorithms at Google like the one here (not sure about this particular issue) and I use Google photos for all my old photos. People at Google read Hacker News and fix things, they just don't comment on spec
774.
by mgraczyk 4y ago
In terms of practical engineering it's not solved, I mean that that SOTA in photogrammetry is good enough to create high quality textures and meshes directly from pictures.
775.
by mgraczyk 4y ago
Generating good video game assets from pictures is solved, but this does more than that. It generates modified versions from words.
776.
by mgraczyk 4y ago
There is reason to believe that won't be true much longer, for example https://wikibon.com/qlc-flash-hamrs-hdd/ It's likely within the next 10 years HDDs will no longer be viable. In terms of TCO with replace
777.
by mgraczyk 4y ago
For languages like C and C++, the idiomatic way to get the first optimization mentioned is to use formatting strings. Instead of splitting these two calls into four: printf("Good day professor Jones"); printf("Good
778.
by mgraczyk 4y ago
It's not what puts bread in my mouth though. I don't work there now and don't work on anything related to ads or messaging. CA happened but that has nothing to do with this. The policies that allowed CA to collect data were v
779.
by mgraczyk 4y ago
It's hard to believe because I worked there and worked on this stuff (data and ML side) and know that they aren't.
780.
by mgraczyk 4y ago
One of the main things were learning from the current trajectory of large models is that this kind of supervision isn't necessary. It's better to focus on bigger models, more data, and better datasets. Models will improve faster t

More ›