3 ms·

Are you serious?

dijksterhuis 2y ago
not the parent, but yeah. they’ve basically described the system of discord there. chat stuff is just one possible use case for the system. like, when i read the root comment i was a bit like “huhhh?” but reading Spivak’s comment it now makes total sense. also to add — * discord gives you easy topic queue debugging. just login yourself and read the queue entries. * something like 2000 character text content per message * RBAC for topic queue access * can probably turn it into a pub/sub topic queue system if you start adding webhooks and bots.
- stavros 2y ago
  Let's try to send a few hundred messages per second over it, a task any queue should be able to perform easily, and see how well it performs!
  - LelouBil 2y ago
    I mean, huge bots are doing just that. Your own non-approved bot might get rate-limited hard though.
- paulddraper 2y ago
  This is just such a terrible idea. You really don't need to defend it. You'll get rate limited in two seconds, for no upside.
  - dijksterhuis 2y ago
    if i walk in and out of a room 100 times in a day, that is an average of 0.001157 messages per second required to turn lights on/off. i highly doubt i’ll be rate limited for that. were not talking an enterprise use case here. not every problem is faceboook scale. also, discord is free to use and has very thorough documentation, on top of all the things previously mentioned. so there’s definitely upsides to it.
golergka 2y ago
That's a serious part of Midjourney business strategy. Hard to imagine how much money did they save on the frontend.
- throwaway71271 2y ago
  or how much money they lost on missed business
  - reissbaker 2y ago
    You don't have opportunity cost of missed business, if the alternative is that you don't exist. If it allows you to launch faster, you aren't losing business by not having every feature; you're gaining business that you would've lost by not launching sooner. You can always build the extra features later. Midjourney avoided having to make a frontend at first, meaning they launched faster. Once they got product/market fit, they built a web frontend, which is now better than the Discord interface. Seems like a pretty smart strategy, that worked well, for a completely bootstrapped company that never raised VC.

Are you serious?

not the parent, but yeah. they’ve basically described the system of discord there. chat stuff is just one possible use case for the system. like, when i read the root comment i was a bit like “huhhh?” but reading Spivak’s comment it now makes total sense. also to add — * discord gives you easy topic queue debugging. just login yourself and read the queue entries. * something like 2000 character text content per message * RBAC for topic queue access * can probably turn it into a pub/sub topic queue system if you start adding webhooks and bots.

stavros 2y ago
Let's try to send a few hundred messages per second over it, a task any queue should be able to perform easily, and see how well it performs!
- LelouBil 2y ago
  I mean, huge bots are doing just that. Your own non-approved bot might get rate-limited hard though.
paulddraper 2y ago
This is just such a terrible idea. You really don't need to defend it. You'll get rate limited in two seconds, for no upside.
- dijksterhuis 2y ago
  if i walk in and out of a room 100 times in a day, that is an average of 0.001157 messages per second required to turn lights on/off. i highly doubt i’ll be rate limited for that. were not talking an enterprise use case here. not every problem is faceboook scale. also, discord is free to use and has very thorough documentation, on top of all the things previously mentioned. so there’s definitely upsides to it.

golergka 2y ago

That's a serious part of Midjourney business strategy. Hard to imagine how much money did they save on the frontend.

throwaway71271 2y ago
or how much money they lost on missed business
- reissbaker 2y ago
  You don't have opportunity cost of missed business, if the alternative is that you don't exist. If it allows you to launch faster, you aren't losing business by not having every feature; you're gaining business that you would've lost by not launching sooner. You can always build the extra features later. Midjourney avoided having to make a frontend at first, meaning they launched faster. Once they got product/market fit, they built a web frontend, which is now better than the Discord interface. Seems like a pretty smart strategy, that worked well, for a completely bootstrapped company that never raised VC.