Your browser (if you're using one of the "usual ones") doesn't really do much with the response's status code if it doesn't match a few specific ones for redirecting/caching/protocol shenanigans.
Anything in the 4XX range is going to be treated as just a regular ol' response, just like 404. (You could serve an entire site with all responses set to status=404, and be fine... other than probably never getting any cache hits) If you don't include a body in the response, the browser might sub in it's own error page, but it will just communicate that the user agent made a bad request.
204 has weird behavior in Safari and Firefox for example. Entering a URL returning 204 in the URL bar will not change the URL bar to it, leaving its contents to whatever was there before. Similarly if you click on it it would not actually navigate to the page.
I've seen sites that use unexpected HTTP response codes, I think to try to defeat bots. The front page would return a 503 Service Unavailable, but the body was just normal content that would load a bot detection script and then redirect you to the actual content.
I successfully wrote a bot that would bypass it all, but it was weird, and became a slight challenge since I couldn't rely on response codes to determine if I succeeded. When I solved the challenge, it would return a 400 Bad Request while serving me the content I was looking for.
Once upon a time, Internet Explorer used to substitute its own error pages if the body of the error response was too short for its liking. Those depended on whcih error code it got. (I expect nobody has used an old enough IE to see those pages for at least a decade.)
> These agents are showing the early signs of swarm intelligence.
Ehhh... it's not that impressive is it? I think it's worth remembering that you can get extremely complex behaviour out of conways game of life [0] which is as much of a swarm as this is, just with an unfathomably huge difference in the number of states any one part can be in. Any random smattering of cells in GoL is going to create a few gliders despite that difference in complexity.
100% anecdata, but I think YouTube nudges your ad profile towards some averaged out cosign product of everyone’s ad profile at regular intervals.
I’ll discover something new, then get pushed a ton of things related to it, which is really good! After a very long break of ~4 years, I started playing oldschool RuneScape again, and that interest weaved its way into my recommended feed perfectly for a month. Felt like I was picking up where I left off, new folks making OSRS video essays, folks I remembered from a long time ago that I had unsubscribed from, exactly what I want out of an algorithmic feed when I’m freshly into something.
Then BAM, gaming content. Some sort of threshold gets hit and now I’m being pushed hyper popular gaming content regardless of RuneScape-y-ness. There’s still a nudge towards it, but I got placed in some “gaming” cohort and it totally crowds out my recommended feed. I don’t really do much gaming outside of this stupid old MMO!
All that’s to say: it might have been a year since you last had one of these inflection points where YouTube will let your ad profile exist as an outlier for a bit.
I would not trust this as-is. I do not like the `curl | sh` install strategy generally, but especially with something like this it feels sketchy.
> We couldn't read your secrets even if we wanted to.
Yes you can, you got to run a shell script with root privs when the cli was installed. You might only store ciphertext in your DB but skimming the shell script, it's dumping a mystery binary off your digitalocean spaces bucket and giving it all-user execute privs. There is no way to verify that binary isn't skimming my key.
to be super candid, this isn’t open source because i don’t have the bandwidth to maintain/support another open source project. that may change as time goes on, though.
i get it’s a trade off, though, and i respect anyone not wanting to use it because of that.
I mean, to be super candid back at you: if you don't have the bandwidth to maintain/support another open source project, I also doubt you have the bandwidth to maintain a custom-built key/token/password store entirely on your own, for free.
Your pitch for storing "API keys, tokens, and credentials" puts you personally in a rather liable position if someone uses this exactly as described, and you've made a mistake in code no one else has seen that either gives YOU those credentials, or leaks them somewhere another party can see them. (Analytics, logs)
For yourself, this is kick-ass and solves a real problem. But I might refrain from pitching it for use by others because there's basically only downside for you in that.
to clarify, i meant i don’t have the bandwidth to run this as a business and an open source project. not either, but both.
that said, i thought more on it last night, and i’ve decided to open source it. just going to be explicit in the README that i wont be offering support for anyone wanting to self host atm, just wasn’t built to be easy to self-host (external service dependencies and etc).
> ...for a one-time purchase of $4.99 - No subscription.
That's really neat! But you may want to scale your pricing with your ambitions. That's very cheap, and it will be much harder to increase single-purchase pricing if you start at 4.99$. Totally unrequested advice, but maybe a 4.99$ "essentials" version that acts like a flipbook, and a 49.99$ "pro" or "full" version that includes future features you might add down the road, like more tweenable properties or IK bones.
Really nice job on the app so far! I also really like your AI use statement. I'll keep Animate in mind when I need something like this!
Thank you! My intention was to keep it very low cost so a lot of people have access to it. I definitely do not anticipate going above $15 at any point, and since it's still early stages the current price felt right!
This is an immensely cool thing that fits the space I imagine for AI within art. And I am not someone that likes framing AI output as art.
If you had instead drawn this, there would be charm and fun details EVERYWHERE. Little buildings you know would have inside jokes, there would be references snuck into everything. Who YOU are would come through, but it would also be much smaller.
This is HUGE, and the zoomed out view is actually an insanely useful map. It's so cool to see reality shifted into a style like this, and there's enough interesting things in "real new york" to make scrolling around this a fun thing to do. I have no impression of you here other than a vague idea you like the older sim city games BUT I have a really interesting impression of NYC.
IMO, that's two totally different pieces of art, with two totally different goals. Neither takes away from the other, since they're both making something impactful rather than one thing trying to simulate the impact of the other. Really nice job with this.
I feel like I've seen enough of these "X is good/bad for Y" articles on these popsci explainer sites surrounded by the same ads for an entire lifetime. Not that I don't think this is actual research, it's just the canonical link we end up with here on HN is probably... not the ideal one.
I wanted to grab the journal link, but going back to the article now, I'm trapped in a "keep reading" loop that blocks the article, so I can be told "Mouse Study Suggests Nose-Picking Has a Surprising Link With Alzheimer's". If I could use an adblocker on the work PC I would.
I've fiddled with Open Props [0] a bit lately, seems like a nice middle ground! Colours/fonts/spacing/etc that look nice together are there, but it's still up to you to use them. (And you're still writing CSS, so might be a deal breaker if that's the part of tailwind you like... but CSS is rather nice nowadays.)
I might be missing something, but was this project started in 2016? I'm not sure what line in the sand you're drawing. That was some minima for developers "knowing UI actually matters" I presume?
Anything in the 4XX range is going to be treated as just a regular ol' response, just like 404. (You could serve an entire site with all responses set to status=404, and be fine... other than probably never getting any cache hits) If you don't include a body in the response, the browser might sub in it's own error page, but it will just communicate that the user agent made a bad request.
reply