Remote
Replying to
@Joshsharp
Honest answer: The same way that I'm writing this on a phone which both in the production process and as a software-social ecosystem has obvious moral and ethical implications. Or how I use retirement investment funds which likely prop up the price of many terrible corporations. Nothing's perfect and we make imperfect choices of balancing comfort of life with the world impact every single day.
Replying to
@sodiboo@gaysex.cloud
This is why I need every monitoring solution to provide heatmaps in addition to line charts. Ideally even default to them. #LineIsALie
Replying to
Replying to
@Joshsharp
Nor am I suggesting giving up. I'm just saying I make the imperfect choices using those systems for building something for the community I'm involved in or improving work at my clinic. That's my balance. Less putting aside, more definitely being aware of it.
Replying to
@lina@vt.social
This basically applies to most libraries and utilities they release IMO. There's so much silly crap I needed to reverse / reimplemention from their build tools recently. Just opensource all the simple things already or make them non-private at least... 😤
Replying to
@NosirrahSec@infosec.exchange
Let me rephrase. Depending on what the person does, I wouldn't find it weird for an "expert Linux user" to attempt this. It should likely be alerted, but that's completely orthogonal to Claude doing it.
It's like crowdstrike flagging/blocking my custom compilation of a linker recently. It's ok it did that, I'm just not the typical user. If I worked on the same project with Claude, it would also get flagged.
Maybe missing context? But it doesn't seem Claude specific.
@GossiTheDog@cyberplace.social
Replying to
@sarajw@front-end.social
Yup, I guess I reacted strongly to "They generate outputs, instead first trying to shape the problem (...)". When it's really "they do what you ask for", we get the opportunity to teach and encourage better questions instead. It works with both people and LLMs. If you want something to improve, you have to explicitly talk about it.
@stephaniewalter@front-end.social
Replying to
@stephaniewalter@front-end.social
That's an interesting one. It's a bit hard to know when to stop and for me it's when the "payoff from this really hard decision is minimal" threshold hits. I admit it's something you need to get used to - the first time I tried to go for too long and just wasn't happy at the end because every decision had some negative tradeoffs that I became more aware of. Sometimes it's easier not to know 😂
Replying to
@yosh
(Yes, I know it's a bit more complicated than that. But not *that* much more complicated. They can take more money for actions if they need capacity - run a business like a business.)
Replying to
@troed@swecyb.com
Also sounds vaguely dirty, like something I should check on the urban dictionary. Which is not a bad thing 😀
@freequaybuoy@mastodon.social @pluralistic@mamot.fr
Replying to
@infobeautiful
This is not a good info diagram and it's not even athletically pleasing. Why share it?
Replying to
@NanoRaptor@bitbang.social
Moving to an English-first country with a name outside of ASCII kind of does that to you automatically 😀
Replying to
@stephaniewalter@front-end.social
I hate to grumble once again about this, but I feel like this is another person who just asked for the output and never considered alternatives. Asking LLMs to review your spec and ask questions about things missing in your design is brutal. They always come up with: simple things I missed, things that may need thinking about, and things that are so left field they make me reconsider the whole project. And they'll keep working on the design forever if that's what you want!
Replying to
@doragasu
Here's an example of a security audit from a better environment. Next to no guidance. https://claude.ai/code/session_015N3igQqkaYkNXhxEWzidTJ
(Trying to create a full report made it complain, so it's just a chat log unfortunately)
@Npars01 @lina
Replying to
@doragasu
They can fail, but also they usually don't if you know how to hold them, and if you can spot things going wildly wrong. I mean, just sending a link to github will likely result in terrible answers for a number of reasons. Clone the repo and run it in a harness made for development, and you'll get much better results. But you need to know those things and LLM itself will not tell you you're holding it wrong unless you ask.
@lina
Replying to
@ZachWeinersmith
I do like the "moderate inefficiency" idea in life. As a weird side effect I grew something like... pride (?) in not getting super light laptops. Like, I don't have an issue lifting a few kg in one hand, so why should it matter? Getting a slightly better deal and power this way. I know it's silly 😂
Also optimising my shopping list / path takes about the same time as just backtracking once or twice and I consciously choose the second option.
Replying to
@cyan
If the standard plugs have latches to keep them in place, would these plugs have love handles?
Replying to
@mxsparks@social.treehouse.systems
Just like the usual one, but the C atom had a spiky moustache.
Replying to
@caseynewton @zoeschiffer I really want some company to finally cut them off for not paying like they would any other customer. Have courage, international corporations!
Replying to
express the totality of visual presentation and layout design and typography and animation and digital interactivity
Isn't that a threshold where CSS should've been replaced with SVG?
Replying to
@jimbob@aus.social counterpoint: I can write 10+ languages and could learn SwiftUI too. But I wanted a very specific timer app that I got generated while I was doing other programming work. Nobody I actually talk to wants to stop programming. It's either: save time on writing stupid stuff, generate something while programming anyway, or generate something because they can't program in the first place.
Who are those people who are ready to deskill?
Replying to
@robhon@mastodon.social @Popehat@mastodon.social It's so stupid it may well be one of the "let's post flamebait with a mistake to drive that engagement" ideas.
Replying to
@alatiera
There's *GOT* to be a guard called securix in one of them, right? 😂 It's perfect
Replying to
@NanoRaptor@bitbang.social
Quality ✨
I was only convinced it's not real after reading 4 titles. 🙇♂️
@JuliaRez@chaos.social
Replying to
@cratermoon
The 2024 view on this is ancient at this point. This was a property of models at the time, not LLMs overall. Nanobanana for example passes all the listed tests without issues.
@cadey
Replying to
@chartier@toot.cafe
Ok, those are all cool things, but the title could mention what was she charged for?
Replying to
@lina@vt.social
Ooohhhhh! Yes, I want to see how much more I can squeeze out of pipewire jack interface and real time guitar effects. (It's already crazy low and effectively real time as far as I can tell)
Replying to on mastodon.social
@cmconseils@mastodon.social
Sadly, no 😔 but it made rounds. https://www.polygon.com/2019/10/9/20906250/jeopardy-tetris-viral-photoshop-joke/
Replying to
@decryption@aus.social
I bet it's because of this on Android. It was even closer in the past.
Replying to
@mym
What's the cats' equivalent of OKLAB colour space? Because there's definitely saturation difference along that spectrum...
Replying to
@GossiTheDog@cyberplace.social
As posted this seems... not weird at all? I do compilation and binary experiments in /tmp all the time. If I asked Claude for something similar, I wouldn't be surprised at all.
Replying to
@RepShontelBrown@mastodon.social yes, we know. So what are you going to do about it? Or what do you suggest anyone here does? We have enough outrage posts already.
Replying to
@alys
We're pretty much past the point of having to measure in many cases. It's not usable for everyone (for example there's no improvement in infra work in my experience), but for plain coding there's no presumption. I've finished in hours projects that involved months of slog before.
Replying to on defcon.social
@deviantollam@defcon.social
@UnicornRiot@mastodon.social maybe? Or they may redirect to a better one...
Replying to
@simon@fedi.simonwillison.net
It's interesting that image-2 seems to have the fewest anatomy horrors, but bumping the size and quality results in so many more.
Replying to
@viticci@mastodon.macstories.net I'd say I hope you like the mainstream stuff, because GPT's idea of "more obscure" is for example Black Eyed Peas.
Replying to
@streetartutopia@mastodon.online holy shit, that black flag placement! Perfect photo timing.