vrighter

same here. Then I started writing a path tracer and now i think in probabilities

i recently heard a colleague telling another "vscode can't really help you when writing yaml files". I said that the yaml lsp supports loading a json schema, all you need is just one comment at the top to point the lsp to it, then you get reliable error reporting with red squigglies on every single keystroke.

He just shrugged and said "oh i just use ai".

What takes me one keystroke takes him a full minute. And he didn't even consider the use of actual tools

agent security is a very old lesson UNlearned.

data and instructions don't mix. And agents can't separate them

i still have a gtx 1080 in my machine. But nvidia just dropped support for them, and some games will forever remain glitching (ex 007 first light flickers like crazy on proton)

so a kind of setfenv type thing, but only from c++. That could work

So the cases where you loadfile a file just for it to have a separate sandboxed _ENV that the function that created it, is flat out unsupported, forever?

finally.... my time 2 should be shipped by the end of the month. I've been waiting for it for ages! can't wait!

Interesting how the AI companies are resorting to releasing devices themselves. Pretty clear the AI side of things can't sustain itself.

maybe you should stop overhauling stuff and fix the stuff that's there.

You haven't even finished overhauling control panel yet...

When I was making an EEPROM programmer with an arduino, I decided to pick up the datasheet, ditch the arduino builtin stuff (I made a separate "board" which stripped the serial code out, and the time functions, because I didn't need them and they fired an interrupt needlessly wasting resources for me), and write your own interrupts for serial. It's obviously more work, but you get to implement whatever you want. You can set up an interrupt for the flow control line. The arduino libraries are shockingly inefficient and limited sometimes.

Or you could do software flow control and use XON and XOFF for a simpler but, imo, clunkier solution

didn't some platforms already do this?

there is absolutely nothing that indicates that it is. Have you considered that it might be complete? Software doesn't need to keep accumulating cruft for it to stay alive. Sometimes it just does the job and is finished

nah, it's just people wanting to pad their resume. And your resume is usually first read by someone who's only impressed by stuff they can see

that's a completely different project

python? that extremely brittle ecosystem?

in that case, github is not the place for them. Nor show hn, or anywhere.

If they're to be trated as disposable, then they should be treated as disposable

it is insanely slow. I regularly need to wait for minutes just for the final crate to compile when compiling some tui program from scratch

brutal for them, hilarious for me

as a maltese dude: fuck her. Our government has long since stopped representing us citizens

i loled at ansible being called stable. When playbooks broke left right and center on updates

i have never used odin.

I did write a compiler that targeted brainfuck

it is malware when everyone is explicitly asking to not have it.

isn't this like the ps3's otheros thingie? Where the advertised functionality of the device was crippled after the customers bought them?

this guy seems so full of himself. Everything I read of his triggers my bullshit alarm. Stuff like claiming feasible solutions to problems that have been mathematically proven don't have any

I think of it like this. Imagine a network with two inputs and one output. What's happening during training is to orient a set of 2d planes in 3d space. Then for each x and y coordinate you can iterate through those planes and figure out where the normal at (x,y) hits them on the z axis, take the highest of your result and that's your network output.

Sometimes the solution needs one of the planes to make a big change to its orientation. and in the process of doing so the fitness will go down. Now if some other plane happened, by luck, to be in a better position to get to that state without lowering the fitness, then it will generally dominate over the other one, and the first one would start to "evolve" to be suppressed by the other one. The first one then becomes (usually, but not always) pretty much dead. Since the network has a fixed number of weights, it can't decide to just "add another plane". It has to make do with a fixed number of them. If too much of them become useless your training can't really recover, you're stuck at that minimum. The overparametrization allows for a bigger chance that some other plane's orientation will happen to be in a position to shadow counterproductive ones.

I think a language having a central repository of libraries that anyone can publish stuff in is a stupid idea in the first place. It's not a question of if, it's a question of when it goes to shit.

it unlocks even more vendor lock-in

Why anyone would even consider using an online password manager is beyond me. Keepass works perfectly offline, and it's easy to sync. Just copy one file. I set up syncthing on my pc and phone and I basically just forget about it. For other places I have a particular subdomain (guid) where I serve the current copy of the file, gated by a password. The database itself is protected by both a password and a yubikey.

as much as humanly possible. If you're not already a good developer, you can't make "good" use of ai

you forgot that to not have a knowledge cutoff and fall behing, you need to always be training new models. It matters jack shit if inference is cheap, if you are forced to do training anyway to stay "competitive"