Hi, I'm James. Eternal dilettante and purveyor of nonsense, much of it about #Python or #physics. I work for a computer vision company whose customers actually care about results, so the current """AI""" craze is slowly melting my brain. I boost more than I post. TANSTAAFL
has anyone, in the history of websites and apps, ever found a chatbot useful? when they were powered by humans? what more with random gibberish with AI agents? i don't understand why that is groundbreaking
@skinnylatte we did our weekly show-and-tell meeting today and one person's thing to talk about was showing how Copilot mangled some tests they were working on and how they could BY HAND fix the tests
It's really odd and concerning to me to see that Debian rolled out a huge change to their packages for @keepassxc with no clear rationale, no clear governance process, and with snide and insulting comments left on the upstream project by packagers.
That's not a great open source community or governance approach.
> This will be painful for a year as users annoyingly do not read the NEWS files they should be reading but there's little that can be done about that.
> ...
> Users who need this crap can install the crappy version
I have spent a lot time explaining to people that Debian is a very opinionated distribution when it comes to packaging software, trying to stay short of accusing them of thinking upstreams are necessarily inferior, and this guy just comes out and says it…
@glyph@xgranade a lot of my resentment of this attitude comes from the Python analogue of this.
python3 is a fundamentally broken experience for a user who wants to write programs on a Debian system unless they are writing "system programs" explicitly designed to fit Debian's worldview. But all the user frustration ends up on Python's doorstep, and Debian's response is still too close to a scoff
Underappreciated reality that a package named pkg REALLY means "what Debian thinks pkg is"
I hate it how transit apps will have a mode that lets you say you're about to get on a bus and then tracks your trip but if you're already on the bus and you just want it to estimate when you'll get there this is always awkward or impossible in the UI
@mcc the Transit app that I use here has UI for this but I've never actually tried it on the bus, only the subway. I remember there being a quirk to how it's displayed (in particular two UI elements that seemed inconsistent) but the information is at least there without doing anything special, and it's been accurate when I've used it (which is rarely)
@skinnylatte people really do lose their minds over the idea of exclusivity, it's incredible how many products the American market supports that are basically "I hate other people unless they're the RIGHT SORT of people"
Mastodon filters are really what sells me on the platform.
Earlier today, I created a "paywall" filter to label news websites with a paywall label, so I can skip over them more easily.
Creating the 'paywall' filter brought back memories of the early days of Twitter, when users created hashtags before Twitter introduced UI support for them.
I have dozens of filter categories which are nice to take the edge off either temporarily or forever.
@webology the "temporarily" is such a game-changer for me. Sometimes I'm just not on the same wavelength as someone else whose posts/boosts I generally enjoy seeing. It's nice to be have "take a break" as an affordance.
When AI hype has settled some, I'd like to see neural primitives be considered to be part of standard CS education along with other ADTs.
Hype makes ML look like too good to be true magical algorithms, and then fails because it was a grift all along. But, there is legit value.
When you watch educational videos on auto encoders, U nets, etc etc, they talk about specific things they are good at to fit in a larger solution.
@demofox do you think the primary value of adding these models to the 'standard' education would be in direct awareness of them (anticipating moments like "aha, we could use an autoencoder here"), or a more generalized familiarity with the basic bag of tricks that make up the current world of convnets (i.e. come up with some suitable architectural design and throw gradient descent at it)?
#python fixture config is magic and I don't like it.
def test_something(fixture):
...
So in pytest. What this does is get the name of the param fixture to see if it matches the name of a previously defined fixture function. If you don't know that it looks bizarre. That IS NOT a parameter passed into a function but a sentinel that is used to look up a fixture by it's parameter name.
@pkw if you would prefer to write tests in the more explicit unittest style where you do all the setup yourself, pytest supports running tests so-written. This may be more acceptable if you find the pytest style offensive.
Other replies have already covered why pytest chose this style.
@pkw it does read as very critical of a piece of software that is new to you, but I am personally not interested in arguing about it. Cannot speak for others.
hiring a hitman is a complicated game theory problem because the hitman always has the option of just taking the money and calling the cops. you have to convince them that you are capable of hiring another hitman to come after the first hitman, but the very act of trying to hire the first hitman indicates that you don't have a second, more reliable hitman ready to go. so i guess the moral of the story is don't be a landlord. https://www.latimes.com/california/story/2024-05-07/landlord-sentenced-20-years-in-prison-in-murder-for-hire-plots
props to gbaned for the continued effort to get a reviewer to look at one (1) line of code, this is SO FAR beyond the typical treatment of issues or PRs in that project, which ranges from apathy to open contempt for bug reports/contributions
@xinit@benroyce it is extremely unpleasant to have the thing that was to be left behind brought here deliberately, whether or not it's attached to a message about how that other site is bad.
But that's why Mastodon offers users ways to stop seeing this kind of thing.
@whitequark an infinite number of mathematicians walk into a bar. The first one orders a beer, the second orders half a beer. The bartender sighs, says "you're all idiots", and pours two beers.
@onelson oh yea the perils of hyperspace gone wrong in general are definitely established. It's cool that 8 decided to bring that to space combat, but it does feel like it leaves a hanging "uhhh well why not always do this"
Probably they figured that fans would (do) eagerly debate the reasons if they just never bothered to address it
