Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
apatheticonion
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
8 ms
·
91.
▲
by
apatheticonion
6mo ago
I've been running a production workload for a cost sensitive customer on a small cheap Vultr VPS for 10 years. Updates have zero downtime and there's lots of compute headroom so it never locks up. No, there's no redundancy an
92.
▲
by
apatheticonion
6mo ago
West coast isn't too bad, but being remote, I might move overseas where the time zone is more friendly. I have considered moving to the US on an E3 visa, but with no job lined up, no contacts, it's a bit of a hurdle
93.
▲
by
apatheticonion
6mo ago
I'm happy to work as a contractor with no Australian labor protections, taking the tax obligations on myself. Though I'm failing at the first hurdle of getting in touch with a US company, let alone one willing to pay via contract.
94.
▲
Ask HN: As an Australian, is it possible to get a remote US role?
4 points
by
apatheticonion
6mo ago
|
8 comments
95.
▲
by
apatheticonion
6mo ago
I've just been doing some reading. The driver situation in Linux is a bit dire. On the one hand there is no stable driver ABI because that would restrict the ability for Linux to optimize. On the other hand vendors (like Orange Pi, Sam
96.
▲
by
apatheticonion
6mo ago
I have always found it perplexing. Why is that required? Is it the lack of drivers in upstream? Is it something to do with how ARM devices seemingly can't install Linux the same way x86 machines can (something something device tree)?
97.
▲
by
apatheticonion
6mo ago
Huh! Well I'll be
98.
▲
by
apatheticonion
6mo ago
There are valid use cases however the issue is rooted in lacking browser APIs. For instance, - if you want to do statistics tracking (how many hits your site gets and user journeys) - You have a widget/iframe system that needs to teard
99.
▲
by
apatheticonion
6mo ago
Sorry, I meant that decompilation is practically illegal. You can look at a circuit board and figure things out and there are clean room techniques for decompilation - but neither of these are practical
100.
▲
by
apatheticonion
6mo ago
A big hurdle to this is hardware vendors locking bootloaders and making it impossible (or impractical) to write or use existing drivers. Manufacturers maintain long running forks of Android (often very old Linux kernels) with their drivers
101.
▲
by
apatheticonion
6mo ago
Hopefully this results in investment in desktop environments and Wine!
102.
▲
YouTube: HTTP Server in Rust (For JavaScript Developers)
(youtube.com)
1 points
by
apatheticonion
6mo ago
|
0 comments
103.
▲
by
apatheticonion
6mo ago
Personally this library isn't to my taste - but I successfully use classes (without inheritance) along with reactivity primitives to create beautiful, tiny and high performance React applications
104.
▲
by
apatheticonion
6mo ago
I have noticed that inheritance is largely ignored by experienced developers but it's a hard argument to make that "all good JS programmers do this". Classes are invaluable and are an extremely efficient and ergonomic way to
105.
▲
by
apatheticonion
6mo ago
Same: https://github.com/alshdavid-public/mvvm/blob/main/examples/...
106.
▲
by
apatheticonion
6mo ago
Who does serious work on a phone? I have found app-only experiences that cover looking for a job, looking for a house, doing accounting/personal finances. How can you even stay organized or do any meaningful data entry on a phone?
107.
▲
by
apatheticonion
6mo ago
Ignoring the AI article contents. God I wish Apple offered first party support for Linux on Mac computers.
108.
▲
by
apatheticonion
7mo ago
Same. I started writing a high level Rust that was based on typescript. Then realized Rust wasn't that hard.
109.
▲
by
apatheticonion
7mo ago
Hi thanks! My focus was developer ergonomics so it's rewarding to hear some of that has translated across. The benchmark claim is based on a naive, simple "hello world" GET req/response, comparing the Go stdlib vs the eq
110.
▲
Show HN: Rust HTTP library modeled after Go but 2-3x faster
(github.com)
5 points
by
apatheticonion
7mo ago
|
2 comments
111.
▲
by
apatheticonion
7mo ago
There is the clean room problem though. If a human reads GPL code and outputs a recreation of that code (derivative work) using what they learned - that is illegal. If an AI reads GPL code and outputs a recreation of that code using what it
112.
▲
by
apatheticonion
7mo ago
While web development is fine enough - spending a bit of time doing native development makes it feel like web technologies are very limiting in what experiences are possible. It's also very obvious how limiting they are in the quality
113.
▲
by
apatheticonion
7mo ago
1) Try to use multi-threading on the web 2) It's not just me, all of the apps listed have the same issues. I know that in addition to Atlassian, Meta and Canva have sunk millions of dollars into writing custom bundlers.
114.
▲
by
apatheticonion
7mo ago
1) Wasm wouldn't prevent you from running those client scripts (e.g. remove fixed elements). That would still just be a script that modifies the DOM from an extension after the page is rendered/dom mutations occur. 2) wasm makes s
115.
▲
by
apatheticonion
7mo ago
This is the part about the "wasm won't replace JavaScript" argument I see being slept on and why I am so disheartened about how practically no progress has been made on it. Most trivial apps don't need to be optimized, a
116.
▲
by
apatheticonion
7mo ago
G'day, haha. Yeah Atlassian. 1/3rd of my team were given the boot sadly. One guy had 12 years at the company - crazy times
117.
▲
by
apatheticonion
7mo ago
I wonder if that would extend to training LLMs on decompiled firmware. A new clean room method unlocked?
118.
▲
by
apatheticonion
7mo ago
AUD* - so $450 USD But yes, that's very expensive and surprising to me.
119.
▲
by
apatheticonion
7mo ago
I wonder if there is a "loaded" lawsuit here that could be a win-win for license enforcement case law in LLMs. Take a litigious company like Nintendo. If one was to train an LLM on their works and the LLM produces an emulator, tha
120.
▲
by
apatheticonion
7mo ago
Me too, and I use LLMs often for personal and professional work. Knowing that colleagues are burning through $700/day worth of tokens, and a small fraction of those tokens were likely derived from my work while I get made redundant is
More ›