h3mmy

@Purple@woof.tech It's me, I am that senior engineer! I often see the most mishaps occur with junior devs that haven't learned enough to understand the codebase pushing code that they don't understand. And then when asked for justification for why they approached something a certain way, I get mixed responses and reverts. Juniors also suffer here because they don't learn the code if they're just believing what AI says about the code. genAI is often woefully incorrect about what certain code is doing, and with how big enterprise codebases often are, there's no way it can paint the picture as well as a dev that's been working in that code (without any AI) for years. The juniors either fail to thrive as a result, or take a lot longer to come around to the realization that genAI is actually slowing them down. @cwebber@social.coop @KimPerales@toad.social

@ShadSterling @mjd Tests need to be written effectively for them to be meaningful. Too many devs simply want to make the tests green, even if it means not actually testing the case it's meant to. I'll catch these in code reviews because it's going to have an impact for anyone working with that code in the future hoping that the tests will catch breakages.

My tagline has read "Professional Computer Booper" while I've been in a Software Engineering role. Now that I'm an Architect, what should it change to? (If at all) I'll probably stew and simmer on it for a while, but if any of all have something that might be inspiring, I'd love to hear it. #AskFedi

Why don't you just try writing good tests and let AI generate the code that makes them pass? #GenAI #Meme Backstory: One of our devs was complaining about how they tried to use the AI tools work has been pushing on us and the dumb ways it failed to be useful. A manager jokingly asked if the dev had considered writing good tests and letting AI generate the code. Both myself and the other senior dev groaned at the thought. I made this meme to showcase why. Disclaimer: Assembled by moi, still theoretically a human.

@SteveRudolfi Direct link to the patent https://patents.google.com/patent/US12536233B1/en The Forbes article is too narrow imo. This seems like it's going to be marketed as "making sites easier to navigate" or for "previews" so one "doesn't have to click through." I will bet that this is going to be used for their "agent mode" or "AI mode" that keeps becoming more prominent in the tabs. Like "this blogger has a bunch of queer content, and we've determined that you don't like that so here's an option to generate one tailored to your worldview that shows you only the parts you want while hiding the rest" Or "this storefront doesn't have easy to find buttons, would you like your agent to buy this for you?" I'm too sleepy to write up a whole thing, but I don't think it's going to be simple. Rather it has much more potential for widespread damage to how folks engage with what's left of the internet.

@tarajdactyl There are plenty! Reolink, Amcrest, Dahua, and Unifi generally all rank high as far as camera quality goes for offline cameras. You'll need an NVR to capture and store footage. I know a lot of folks that self-host frigate to do this, and some that have a unifi setup use that as their NVR. The backing storage can be a dedicated machine or a general NAS with some space budgeted out for camera use. With offline networked cameras, it's generally recommended to have them on their own vlan, and also have firewall rules that prevent them from being able to access the internet at all. Personally I prefer wired cameras that are PoE compatible. It's too easy to deauth wireless cams for my comfort, and PoE means I only have to run one cable instead of two.

@octothorpe@mastodon.online This doesn't do anything to stop spam. It's like nobody checked how easy it is to spoof a number. It adds overhead to simply being able to communicate via the telecom networks, and it removes privacy. Meaning that companies and scammers can target people _more_ effectively. And it's not just them either. Burner phones are extremely useful for journalists, folks working on escaping abusive situations, etc. This is not how you craft good policy. It is simply another avenue of removing options for anonymous and quasi-anonymous communication.

@CiscoJunkie@techhub.social Since you already have sync working via the live sync plugin, for sharing you could use the share-note plugin and point it to your own hosted backend: https://github.com/note-sx/server https://github.com/alangrainger/share-note/

@koalou I've had a junior literally copy paste the logic from the code into the expectation without any actual test data. It gets picked up a coverage but is functionally testing nothing. I'm thinking I need to add a test-mutation tool to analyze quality of tests and push to get that enforced. It sucks not being able to trust devs to want to put out high quality work. Definitely had run-ins with dynamic date time tests and edge scenarios. Having only the tests break sounds frustrating. I've gotten deployment overrides when testing environment issues arise. It's straightforward because I keep meticulous paperwork, but I'm happy to not have had to do that for a while.

@0xabad1dea That felt like someone left alone with an LLM unsupervised. AI-scannint tools rated it extremely high on "written by AI" but barring that... When I saw they wanted to add auth using JWT tokens for active layer 2 routing... I was like "They missed April Fools by a couple of weeks. And it's not even funny." Then saying it's "backwards compatible" and in the same paragraph highlighting the packet header being 8 octets bigger... Anything that sees a "Version=8" packet will drop it as a 'failed to parse' I don't even want to tear into the assortment of "requirements" like the new socket type,new DNS record, new packet types (e.g. icmpv8) etc because I'd be putting in nore effort than the author did which as a general rule I try to avoid.

@NyxMir That all sounds amazing! I'm a bit envious :neobot_heart_green: I checked with the querant and they can't be approved for museums, etc because it's "leisure," however if it's a workshop at a museum then it's fine.

@NyxMir 👀 I do believe travel for something like that is covered but I'll check with them

@geekygirldawn I like to make it very clear to newer devs I interact with that they are expected to ask for help. And it's a balance of giving them the answer and guiding them through the process of figuring out the answer. Sometimes a problem breaks down to them not knowing something and being too embarrassed to ask directly. Like "Oh, so that's how you check the logs in that environment" or "I don't know how to use a java debugger" and I try to be reassuring that it doesn't matter that they don't know those things. It matters much more that you communicate and ask for some help figuring it out instead of languishing.

@pointlessone@status.pointless.one @skaverat@skaverat.net Oh the NULL mishaps are so prolific. It still causes so much grief for pain. There is a radio lab episode that goes over some of the NULL cases. There's also a transcript available. https://radiolab.org/podcast/null

@dw_innovation@mastodon.social It's a little amusing that this wasn't archived on all the archives listed already. It's been added to archive.today as of right now though #Archives #DigitalForensics

@skaverat@skaverat.net Most common implementations of time_t use a signed integer, so it's trivial to track last dates. So I would be surprised if anyone had issues, and would be curious to hear about this as well. 1969-12-31 23:59:59 UTC would be -1 which I can see being easily confused with an error but anything other than -1 is pretty clear.