seems like most of the time i bring up the firefox task manager lately, there's one or more "subframes" from cloudflare just hanging out taking up ~10% of my cpu. i assume this is some kind of bot detection thing? is this just what using a web browser is going to be like from now on

god damn linus is so fucking stupid https://lore.kernel.org/lkml/CAHk-=whTs09_ozN+e09GU68N1wPP56MUzDjrg9rnKxQUa0A_4Q@mail.gmail.com/

(whoops meant to mention that these are all from Emily Short's _Bronze_, an all time classic imo)

(i can very easily imagine a page of Grapefruit that reads "DOOR PIECE (1963) / Walk through a doorway. If the door is closed, open it first.")

the thin line between inform 7 source code and fluxus event scores

i refuse to do anything for the time being—the time being is already far too powerful

We now note that that we do not waive our attribution requirements for Spritely Goblins, which are present in Apache v2 (and for that matter, CC BY, CC BY-SA, and most FOSS licenses) https://codeberg.org/spritely/goblins/pulls/998/files

So if you train on our codebase, you have to retain attribution, which we consider the most important requirement to motivate development of technology in the commons.

If, somehow, you are using a statistical soup which has no way to preserve attribution, well, we don't permit or approve of dropping such a requirement, and actually, the licenses we use never have.

it really seems like every paper that I read about LLMs from the science-y side of academia goes something like "as you can see, our experimental data clearly show that LLMs are comically useless and/or harmful. in conclusion, we hope LLM researchers and engineers are inspired to think about making a plan to consider slight changes to what they're doing"

@mewsleah oh wow i didn't know about the mcl68+ project, very cool! my 68k is plcc, though, so i'd have to make an adapter board or something if i wanted everything to carry through. hmmmm

(i haven't been able to find a document/tutorial that is like "so you have a 68k processor, here's how to build a computer around it" BUT i've looked at various little homebrews like https://github.com/74hc595/68k-nano so i know it's possible. i just don't want to go to the trouble of putting together a whole prototype only to discover that the processor is cooked)

context: i found a board in the trash with an MC68HC001 processor on it. i wanna build a little computer with it, for laughs. the board itself (a 90s video compositor?) is kind of trashed, so i'm not all that eager to spend a lot of time figuring out how fix it.

question: is there a way to see if the processor is functional *without* having "the rest of the computer" (e.g. no RAM/ROM/serial)? if i just hook up VCC, GND and a clock source, will I see bus activity?

i am simultaneously (a) glad that i tried out cython for this project, because now i'm more familiar with how it works and behaves; and (b) convinced that you should basically never use cython. [alanis morissette voice] you live, you learn

i'm sure there's an option to turn on warnings about this, but a thing that sucks about cython is that leaving off a single type annotation on a function sets off a chain reaction in which anything that has the mis-annotated function in its call stack slows to a crawl (because now that function is using python values that have to be dynamically allocated and converted back and forth from C values). and then i end up having to comb through the generated C to find the problem

jesus christ, Linux has had a 100% LLM takeover.

this project is completely cooked.

even after the AI bubble pop, it's not clear how to recover from this level of bullshit from this level of bullshitter.

https://lwn.net/SubscriberLink/1099594/be20382b54a9b4cd/

> Linus Torvalds believes that blaming LLMs for the kernel community's problems is wrong, when the problem is LLM users who do not understand how to use them or how the kernel works. A properly trained LLM is a useful tool, but it takes a lot of time to reach that point.
> Torvalds and others repeatedly said low-quality contributions from new developers should receive low priority, and that the community should stop trying to make contribution easy. Many developers said, repeatedly, that trying to mentor those developers is almost always a waste of time; they do not listen.
> There is some concern about companies maintaining private kernels with LLM help, but there does not seem to be much to be done about it.
> Kernel developers should not be required to use LLMs to work in the community, but those who do need to have enough resources available to generate acceptable results.
The requirement to add Assisted-by tags to patches created with LLMs will be removed, but maintainers still want to know about how a patch was created.

so how's it going already?

> The development community is struggling to keep up with the flood of security patches, and will continue to for some time. Some of the most urgent fixes may bypass maintainers and go straight to Torvalds. Maintainers were cautioned against telling their employers about vulnerabilities before public disclosure happens.

Really, I'm not out to suck as much ass as Microsoft. That will just be a completely unintentional side effect.

i think if i'm going to implement sprite rendering i might as well implement audio. sigh

(the accesses in green are coming from the DMA peripheral, and the bus conflicts are happening when both DMA and the PPU are trying to put a value on the address bus)

(i wonder how hot you could make a game boy SoC if you did this on purpose)

(it's genuinely bewildering to keep track of which bus accesses from which components are possible under which conditions)

hmm, you can create some very pretty conflicts (in red) on the video ram bus if you start a DMA from VRAM while the PPU is in mode 3 (tile/sprite rendering)

[me, after i've archived all of the images in your camera roll] oh no no no no. I don't have your data. I just have blocks of discrete cosine transforms *derived* from your data. it's all very secure

i also don't understand the point of apple's "secure exclaves." okay, so you can't access the audio or video data directly. but you *can* access records of the individual people who are using the device, tied to their apple account, and what they're talking about, presumably with timestamps? how is that any better at all

remember when a big tech company could release a new product and you could think about anything other than surveillance, planned obsolescence, vendor lock-in and e-waste