Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
lionkor
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
16 ms
·
331.
▲
by
lionkor
6mo ago
I have an alternative! Regulation. A government can simply regulate what is and isn't legal, and in most of the world, that's been what governments do. I'm sure a country like the US, which is filled with lawyers, can come up
332.
▲
by
lionkor
6mo ago
Yes, you do owe the inspiration money if the result is close enough. Welcome to intellectual property laws!
333.
▲
by
lionkor
6mo ago
That's a very strange view. So if I publish a paper with some novel method of compression, for example, it's fully okay for the first person who sees it to open it on screen 1, open an editor on screen 2, transcribe it, register a
334.
▲
by
lionkor
6mo ago
Well, the way intellectual property works, anything I write on the internet is, by default, all rights reserved. Different website's policies will impact this, of course, and different laws (and quirks like "fair use") as wel
335.
▲
by
lionkor
6mo ago
> I quickly started hitting limits developing several apps at the same time > I'm a career engineer I'm trying really hard here to be nice, but what the hell are you doing? Are you vibe coding multiple apps in parallel and c
336.
▲
by
lionkor
6mo ago
Every cent you spend on this, remember: The people who made this possible are not even getting a millionth of a cent for every billion USD made with it (they are getting nothing). Same with code; that code you spent years pouring over, fixi
337.
▲
by
lionkor
6mo ago
But you BET that person is never, ever making that mistake again!
338.
▲
by
lionkor
6mo ago
Those human assistants can be held accountable.
339.
▲
by
lionkor
6mo ago
It's just an LLM.
340.
▲
by
lionkor
6mo ago
The kind of code, with the kind of quality, that LLMs can output has become cheap. Learning has not, and neither has genuinely well designed, human designed, code. This might be surprising to the majority of users on HN, but once a really
341.
▲
What Claude Code Chooses
(amplifying.ai)
1 points
by
lionkor
6mo ago
|
0 comments
342.
▲
by
lionkor
6mo ago
That would put itself back in the bottle by running killall to fix a stuck task, or deleting all core logic and replacing it with a to-do to fix a test.
343.
▲
by
lionkor
6mo ago
I agree fully. Not sure what they are on about, with "no labs??" in the replies. You still do all the same things, and they are graded, but this doesn't affect your final grade. Instead, you need to pass a threshold to enter
344.
▲
by
lionkor
6mo ago
Weaponized FOMO to drive everyone to use AI isn't really a great idea
345.
▲
by
lionkor
6mo ago
I haven't seen it.
346.
▲
by
lionkor
6mo ago
It also writes lots of bugs which it'll catch some of, in an independent review chat. This is bogus. If you think LLMs write less buggy software, you haven't worked with seriously capable engineers. And now, of course, everyone ca
347.
▲
by
lionkor
6mo ago
People get arrested for running large cocaine operations, that's the difference.
348.
▲
by
lionkor
6mo ago
The first example is tic tac toe. Why would anyone bother? None of those eash things are relevant for people who use AI. They don't care about learning, improving, exploring how things work, creating, being creative to that degree. The
349.
▲
by
lionkor
6mo ago
Yes, but every time the capabilities, security, accuracy, or any other quality of LLMs is challenged, the default answer is that we'll essentially have AGI in a quarter or two. It's very tiring to try to argue with people about cu
350.
▲
by
lionkor
6mo ago
It's already YEARS down the line from when this was promised, we can't keep saying "but in a couple more quarters it'll all be different!".
351.
▲
by
lionkor
6mo ago
How can you argue they don't lie, as if they have any idea of correct vs wrong? There is no brain there. When statistics overwhelmingly say "yes" is the correct answer to something, it will say "yes" -- completely i
352.
▲
Silicon Valley's Billion Dollar Design Scams [video]
(youtube.com)
4 points
by
lionkor
6mo ago
|
0 comments
353.
▲
by
lionkor
6mo ago
How does that not make any sense? It turns into a statement. If it no longer makes sense, the original title was bad clickbait, no?
354.
▲
by
lionkor
6mo ago
I can't feel bad when the subscription circumvention uses the same method they use to get their stuff to rank high in search results. News pages want their content indexed, so they can pull a bait and switch.
355.
▲
by
lionkor
6mo ago
If they are built on Godot, Unity, Unreal, or are built with SDL+OpenGL or SDL+DirectX or C# MonoGame any kind of combination that is reasonably common in the last 20 years, it'll run well
356.
▲
by
lionkor
6mo ago
There are a suprising number of games with a native Linux version, like ArmA 3 (which is an ancient, broken, outdated port, but it exists and used to install by default), or Minecraft, and plenty of other games have native linux versions th
357.
▲
by
lionkor
6mo ago
Its a matter of language. When non-technical people here "Linux doesn't support anticheat", that's a lot different than when they hear "anticheat chooses to block Linux".
358.
▲
by
lionkor
6mo ago
These aren't faults of Linux, they are faults of hardware vendors. There is nothing(!!) Linux or the community can do if the hardware vendors decide to not support Linux at all. I know that's not what you meant, but to less techni
359.
▲
by
lionkor
6mo ago
Lutris solves that
360.
▲
by
lionkor
7mo ago
Time is money, get talking! Yes, last time (recently) I tried, the original games ran very well, with no (Linux specific) issues!
More ›