musselwhite.dev
Computer science PhD student at Howard University, digital photographer, and Oxford comma enthusiast.
🪪 musselwhite.dev
📷 joncaptureslight.com
188 posts
107 followers
159 following
Regular Contributor
Active Commenter
comment in response to
post
Ooohhh, Zen looks really good. I'm sure it would be difficult to accomplish, but I wish it was more common for browser engines (or renderers) to be modules that could be swapped or selected based on the website. I do prefer Firefox/Gecko, but I have to use Chromium/Blink on certain websites.
comment in response to
post
By default, applications absolutely should refuse to resolve an onion address, and their developers are right to resist allowing it, but RFC 7686 doesn't prohibit them from defining signals that the system administrator can use to relax an application's onion restriction. 3/3
comment in response to
post
Also, RFC 7686 uses the keyword SHOULD as defined by RFC 2119, which means, "there may exist valid reasons in particular circumstances to ignore a particular item, but the full implications must be understood and carefully weighed before choosing a different course." 2/3
comment in response to
post
Yeah, the discussion kind of fizzled out, but I don't think there should be a standard way for an application to check if it's behind an onion router because it could be behind an untrusted onion router (e.g. a public wifi hotspot). 1/3
comment in response to
post
Of course, supporting and rewarding effective teachers is expensive. Many of those people who blame the kids are the same ones who would support cutting funding for public schools. With limited funds, schools can't be very picky about who to hire. Thus, the downward spiral continues.
comment in response to
post
Don't get me started. Teaching is hard, and making good assessments is harder, but teachers who just read off PowerPoint slides, reuse tests, take no pride in their work, and refuse to fix their mistakes are creating an environment that punishes those who want to learn and teaches them to cheat.
comment in response to
post
That's got this vibe.
comment in response to
post
Years ago, I tried mac OS X and was surprised at how easy it was to install apps. Dragging into the Applications folder was too intuitive for me, lol, but that took me on a deep dive of the `.app` directory structure that evolved from the old resource forks of apps on classic Macintosh. Cool stuff.
comment in response to
post
Those are cool! I don't have any from GitHub, but I sometimes get similar pins. A shadow box can be a nice way to put them on display, especially when accompanied by a certificate, letter, or picture. Otherwise, I usually just attach a representative one to my bag and toss the rest in a drawer.
comment in response to
post
Personally, I'm all for it. It's wayyyyy better than the Ouija board I used to use.
comment in response to
post
The strength of its pervasive open-source culture is the cause of, and solution to, all of Linux's problems. Everyone is allowed to fix any of the problems you run into, and somebody might have already, but it's in a fork that never got merged.
comment in response to
post
Yeah, it's hard to tell the difference between an enthusiastic early adopter wearing rose-colored glasses and a grifter sometimes. I'd also like to know if there is a good podcast by people who don't gloss over problems. They talk AI on the @syntax.fm podcast sometimes, but it's a web dev show.
comment in response to
post
The idea that deleting a file could be unethical to the file is quite funny. AI models do not learn. They use context as working memory, but that context is not controlled by them, and neither is their training. They have no agency. Their creativity comes from a random number generator.
comment in response to
post
@cred.blue has something for that now.
comment in response to
post
Yeah, I was just wondering about what happens if you block a verifier. It would be strange to still see the verifications they issue.
comment in response to
post
Yup! It takes just a few minutes.
comment in response to
post
I read that article expecting to learn that Bernie Sanders hadn't actually been arrested, but the article confirms he was arrested and just likely didn't cross paths with John Lewis during his time there.
comment in response to
post
That is one mildly annoying thing about installing windows without a Microsoft account. It's hard enough to trick it into skipping the Microsoft account, but then it requires security questions for account recovery. I can enter random gibberish, but I'd rather disable them entirely.
comment in response to
post
Unfortunately, that hits the same 2,147,483,647 ceiling, so it falls in the same z-index.
comment in response to
post
Yeah, it's not ideal, but it's communicates its purpose better than arbitrary-length sequences of 9s, and it's easier to remember than `2147483647`. It might be the best way to overlap embeds that have z-index values outside our control. Is there a better way to handle that case?
comment in response to
post
Yeah, moderation lists seem to work better because then it doesn't matter if the discovery algorithm recommends a type of post to you because it'll be muted or blocked, so you don't actually see it.
comment in response to
post
You can tap the three dots next to a post and select "Show less like this," or the quicker way is to mute an entire list. Find the account on clearsky.app to see what lists block them or search by topic on the blue sky directory to find a list that works for you. blueskydirectory.com/lists/all?ty...
comment in response to
post
I'm looking forward to following once your account is verified. It takes just a few minutes to change your handle to dccouncil.gov, or a subdomain of it, instead of using a handle anybody could have chosen. bsky.social/about/blog/4...
comment in response to
post
sounds like you need pub/sub, a common feature of kv stores, but mqtt is a solid option that isn't just for configuring smart devices. it might be worth comparing against others for your use case, if you haven't already.
comment in response to
post
I wrestled with Microsoft Edge settings to finally set DuckDuckGo as the default search engine. Opened it up again... and got a popup begging me to revert to Microsoft's own search engine. It's a battle of wills, I tell you.
comment in response to
post
Ahh, there I go again, offering solutions where none are needed. Sorry about that! Yeah, the start menu got buggy for me, too, in the past year or so. I'm trying to get in the habit of hitting alt+space now, but it's an old habit.
comment in response to
post
Microsoft PowerToys has that feature and a few other useful utilities.
comment in response to
post
That oscilloscope costs more than my car, but look at the specs on that thing!
comment in response to
post
You'll have a blast, I'm sure. Your project sounds really fun.
comment in response to
post
Waveshare has some e-paper stuff, and they're pretty popular in the space for DIY components, but CanaKit is probably the biggest name in kits.
comment in response to
post
Or rubberducking is having a conversation with something that can't understand you. LLMs pretend to understand you, but they all use a random number generator and other techniques to avoid appearing deterministic. A rubber duck has 0 parameters compared to 72b, so there's nothing to randomize.
comment in response to
post
I'm not missing the point of it, I used it as a baseline of comparison. The LLM model often gets the answer wrong, but the conversation stimulates my thought process as I try to get it to understand what I'm asking. Hence, it's like a rubber ducky that talks back.
comment in response to
post
Yeah, it's useful as a rubber ducky that talks back. I also really like using it for context-aware code completion.
comment in response to
post
If your goal is to warn people against relying on a single resource without questioning the info, I'm with you 100%. But burying your head in the sand is the only way to avoid relying on other people's work to gather, filter, and aggregate news.
comment in response to
post
Reading someone else's study is not doing your own research. It's reading someone else's research (which certainly references other people's work), and if you are not trained in the field, you might not have the knowledge and experience necessary to accurately interpret the results of the study.
comment in response to
post
Can you give an example of "doing your own research" that doesn't involve referencing someone else's work or using any tools?
comment in response to
post
It's literally impossible to get all of the news that affects you without editorialization or algorithms. Even that video is posted on a platform that uses algorithms. So why scare people away from having an effective aggregator in their toolbox? What do you think "doing your own research" entails?
comment in response to
post
Why do you think there can be only one bubble? Everyone relies on aggregation for their news, whether it's social media algorithms, specific media sources, news aggregators, or a combination. A birds-eye view of thousands of media sources across the spectrum is a helpful tool for exploring media.
comment in response to
post
This is vaguely reminiscent of when the Outlook app prompts for two-factor authentication but gives you a two digit code to enter into a notification from your Outlook app, and you have to start over if you didn't memorize the number before tapping the notification because it's the same app.