entropy

Tada! If I'm first to it, please send the key to @mogsk.bsky.social :) Thanks for the fun challenge!

This isn't the implied Inky mode, which I would expect to be in the "extras" menu with the arcade mode, but... Maybe I'm missing something, I've had two other people look over this demo with me, and we can't find this menu at all

Do I win?

Based on the SAGE demo page, I assumed this was because I didn't have all medals, but even after collecting all of them listed for the available levels, still nothing.

Well, I'm pretty sure I've figured out how? Unfortunately, after seeing this pop up, I'm not able to unlock anything after Practice 2.

PS:
Oauth IS actually as bad as it sounds. There are some good community developed solutions popping up for this. I hope some day it's trivial. Sorry. The rest is easy, I swear.

I do not work for Bluesky and am not an ATProto author. I am just an indie ATProto dev hoping to advance the landscape. I doubt this will reach anyone that it isn't already obvious to, but it's something I want to be more vocal about, to more seriously bridge the informational gap in. Build with us.

I think these benefits are what we need to communicate to devs. It simplifies some of the biggest design problems, and eliminates an adoption barrier of trust between you and your users; they own their data, you just make it better for them. And the more we build, the more we have to share.

ATProto isn't scary, it isn't nightmarishly complex. I wouldn't blame anyone for thinking so, and I only got there through the help of other devs who want the protocol to succeed. But the further I went, the more the benefits became obvious, and the more the possibilities started to shine.

Every part can be hosted separately for your service, if that's what you want, or use existing ones. That information alone is enough to build a ton. Add on a couple more details, "labelers" that put metadata on data and "lexicons" that describe how data should be shaped, and the options grow.

So what's ATProto, for devs?
Your data is stored on a PDS, which can be hosted anywhere you choose.
When you put data in the PDS, it goes to global mirrors called relays.
Servers called AppViews consume those relays, and use it like you would in any other architecture.

This is the perfect mix for an ecosystem that can overtake everything. There are still missing parts, problems to solve, services that can't be built today. But Bluesky alone can not solve those problems, for it does not represent all needs. We must build, discover use cases, propose requirements.

The potential is incredible! And even more incredible, users don't have to opt in to your service. A PDS works anywhere. When you are on ATProto, you are on ATProto. It doesn't matter if you use Bluesky or Smoke Signal or Whtwnd. The user cost for adoption is low. The dev cost to build is low.

You can describe most of it in a paragraph. Its most glorious aspect for developers is that every part of its architecture is optional; you can self host or build on top of existing services as much as you please. This opens the door to projects that were not reasonable for indie devs before.

It took me around two months to get a good grasp on the end-to-end details of ATProto. Most devs I talk to see the architecture diagram on Wikipedia and their eyes glaze over, and don't care to give it any further thought. It's a shame.

But for a protocol that presents a radical change for the way we build services, we have to reach, and teach, developers. I think Bluesky was wise to obscure the technology; it is easy to alienate your users by advertising your decentralized stack. But for devs, it isn't obvious why it's so special.

I don't think ATProto is particularly complicated. It is, however, new, roughly documented, and obscure even with Bluesky's popularity. For Bluesky, this isn't a problem. They can continue iterating the standard, expanding the possibilities, and for users this is simply the platform improving.

Bluesky's success is in heavy part due to its tech, yet the details are hidden from users. It's a strange situation where everyone loves its strengths, and nobody has a clue how it works. ATProto could change the internet, but even experienced devs I talk to don't understand it.

My gut instinct when I saw it was "oh no, this card is unlocalizable"
This doesn't seem like a viable concept in a standard text game

Do I not understand benchmark graphs at all, or did OpenAI suggest with a visual that 52.8 > 69.1 == 30.8?

My confoundment that style data that should only be available for one user was now global to my service is indescribable. It is as though a wizard teleported into my home and turned everything inside out and then just peaced out before I even knew what was happening

Still pretty new to web dev. Learned today that if you don't set cache control on your responses for, say, something cute like a user-custom CSS file, then they'll get cached and globally distributed to every user by Cloudflare. So that's a funny trick for the IETF to play on me specifically

fun fact! the Romans invented a cloud orchestration system similar to modern day kubernetes, which is why their concrete is self-healing

Phil is out here building free infrastructure opening up new possibilities in the ATmosphere, and has addressed developer issues faster than some well funded services with real SLOs that I've used. Support independent AT Proto development!

I'm having a hard time figuring out HOW to pursue this. It seems like something where the only unambiguous answer right now is "that's up to the department, which might just decide not to!"

I forgot about this a little too long, but seems to be working perfectly after the update!

Thanks for the quick attention on this! I'll have to give it a shot today, I'm curious how the fix translates to the case I was running into because I was generating multiple records linking back to one from the same DID- no updates or deletions though.

I'd build it just to gain that power for myself

I can't take credit for it admittedly. @graze.social released AIP for streamlining the process at the perfect time for me, and their devs and community absolutely came through on helping me hammer out issues along the way. That's why even after getting this right, I'm still not sure how it works.

I truly love ATProto, but very few people in the world (not including me, I'm pretty sure) understand OAuth 2.1 and I can't convey to my friends my excitement that, after two weeks of tinkering, I managed to post a single record