bitcrshr

I’ll start with saying that I think it’s good you’re asking about this. A lot of things are possible these days with all these AI tools, but they don’t solve the problem of unknown unknowns.

I think the most honest thing I could say is that it’s almost certainly not “engineered well” but there’s nuance to that. The most important bit is that it almost certainly has at least one vulnerability or issue that could lead to a whole host of problems for both you and your users. Given, that’s true of all code, but without being vetted by someone who knows what to look for there’s a much higher chance some low hanging fruit was missed.

If it’s safe and secure and actually works, then the rest doesn’t matter so much. Not saying it’s gotta be Fort Knox, but even the smartest models can easily miss something that an experienced human wouldn’t.

I would strongly recommend hiring an experienced developer to at least audit it. Will be a lot cheaper than an out of control cloud bill or worse, lawsuits.

I commend you for opening yourself up to potential torment here for the sake of your product and its users, but a) nobody is gonna do it for free and b) even if they did, well, you get what you pay for!

Odd, I got right in. I think it was maybe 5 seconds, didn’t have to do anything. Curious if that’s a bug or if there’s some automated trust or something. Cool concept though, echoes of GPG web of trust days.

2 years ago I bought a Dell R630 for about this much with 128GB of RAM and 2 beefy xeons (for their gen, anyhow). Oh, how the times have changed.

I needed this. Thank you.

I tried Keycloak for a while, it’s really good too. Given it has an admkn dashboard, it’s a bit more “batteries included” than Ory.

Kratos is awesome, especially alongside Hydra, OathKeeper, and Keto. Super powerful combo, if not a little intimidating at first. There’s a LOT of configuration involved, but that’s to be expected if you want to host your own Auth0 replacement.

Their dynamic forms stuff is really cool too, always liked how they chose to go about that. Only complaint I really ever had is that while their docs were overall serviceable, I remember some areas were pretty lacking and I had to dig really far to find answers to some fairly common issues.

Well for one my editor doesn’t support plugins (yet)!. But I have found that learning my editor well enough removed my need for the pesky edits / boilerplate writing offerings of AI. Other than that, I don’t trust it to write anything big, and don’t need it to write anything small.

I’ve tried it a few times and it’s decent for writing unit tests but otherwise often made a mess of things. I understand there’s an art to it but I’m just not interested in putting too much effort into it. I’m always going to go through whatever it generates with a fine toothed comb, so I don’t see it saving me much time anyhow.

I have watched some very senior engineers really dive in with it though, and seemingly with a lot of success.

No worries of HTTP 418 here.

i think you’re right, though i think i was thinking about about this at a higher level. i more meant that the languages in the JVM ecosystem I have experience with (java, kotlin, and scala) have all given me similar unpleasant experiences building them. same tools too, though you probably wouldn’t use sbt for non-scala projects even though you could.

the main pain points for me are dependencies, packaging, and configuration. best i can tell, those pains are shared between anything that targets JVM, especially those that want to have good interop.

This is great. I have been daily driving Helix for a few years now and every release has been really exciting. Still anxiously waiting on a plugin system, but that’s more of a bonus than a need.

love the syntax, and excited to mess with it, but man i’m sad to see it’s on the JVM. if i had to guess, a lot of langs like this are on JVM because that’s a lot simpler than writing a whole backend with anywhere near the same performance or reliability, and i totally get that.

that being said, bearing in mind that i’m not a Java/JVM developer and only rarely have to use it, for the few nontrivial projects i have shipped with it the build system was by far the most challenging and frustrating. it’s so complex and has such a large surface area.

no hate at all, and the trade offs are completely reasonable, but i am hoping during my career we’ll start seeing either a massive simplification of JVM builds or a lot of innovation that would make native compilers easier to build.

(as a side note, it is nice to have langs like this for when JVM is the only option)