ByteCode.News
RSS
SubmitLog in

Community

Chat Logs

Friday, October 9, 2026

  1. * skillbot quit (Remote host closed the connection)
  2. * skillnews quit (Remote host closed the connection)
  3. * skillbot joined #primate
  4. * skillnews joined #primate
  5. * lioloc joined #primate
  6. * lioloc quit (Ping timeout: 242 seconds)
  7. * nevet joined #primate
  8. * blue!~blue@user/blue changed the topic to: https://primate.run | this channel is publicly logged at https://bytecode.news/logs/irc/libera/primate and mirrored to Primate discord, channel #irc
  9. * lioloc joined #primate
  10. dreamrealyo blue
  11. dreamrealsorry wrong channel
  12. * nevet joined #primate
  13. * blue!~blue@user/blue changed the topic to: https://primate.run | this channel is publicly logged at https://bytecode.news/logs/irc/libera/primate and mirrored to Primate discord, channel #irc
  14. * lioloc quit (Read error: Connection reset by peer)
  15. dreamrealfeed subscribe https://deno.com/feed
  16. nevetSubscribed to "Deno"
  17. dreamrealfeed subscribe https://bun.com/
  18. nevetSubscribed to "bun.com"
  19. dreamrealfeed subscribe https://nodejs.org/en
  20. nevetSubscribed to "Node.js Blog"
  21. bluenice, thanks dreamreal
  22. bluefeed
  23. dreamrealfeed list
  24. nevetjustine.lol - https://justine.lol/rss.xml Distant Province - https://distantprovince.substack.com/feed tRPC Blog - https://trpc.io/blog/rss.xml MDN Blog - https://developer.mozilla.org/en-US/blog/rss.xml bun.com - https://bun.com/rss.xml FREEZE64.COM - https://freeze64.com/feed/ Feed - https://hollycummins.com/rss.xml Primate Blog - https://primate.run/blog.rss Rust Blog - https://blog.rust-lang.org/feed.xml Improbable Research - https://improbable.com/feed/ Apache NetBeans - https://netbeans.apache.org/blogs/atom That Privacy Guy! Blog - https://www.thatprivacyguy.com/blog/feed.xml The JetBrains Blog - https://blog.jetbrains.com/feed/ Paged Out! Magazine - https://pagedout.institute/rss.xml W3C - Blog - https://www.w3.org/blog/feed Typotheque - https://www.typotheque.com/rss ByteCode.News - https://bytecode.news/feed.xml Next.js Blog - https://nextjs.org/feed.xml Ziglang.org News - https://ziglang.org/news/index.xml Deno - https://deno.com/feed Random Notes and Cheat Sheets - https://www.kodewerk.com/feed.xml cekrem.github.io - https://cekrem.github.io/index.xml Vanish is... - https://vanish.is/feed.xml Svelte blog - https://svelte.dev/blog/rss.xml Ember.js Blog - https://blog.emberjs.com/rss.xml Open Liberty - https://openliberty.io/feed.xml Simon Willison's Weblog - https://simonwillison.net/atom/everything/ TypeScript - https://devblogs.microsoft.com/typescript/feed/ Eignex - https://eignex.com/feed.xml Docker - https://www.docker.com/feed/ Node.js Blog - https://nodejs.org/en/feed/blog.xml Chicory Blog - https://chicory.dev/blog/rss.xml Terminal Trove - https://terminaltrove.com/blog.xml Angular Blog - Medium - https://blog.angular.dev/feed Read the Tea Leaves - https://nolanlawson.com/feed/ The Astro Blog - https://astro.build/rss.xml Enigmastation.com - https://enigmastation.com/feed.xml The Vue Point - https://blog.vuejs.org/feed.rss Vlad Mihalcea - https://vladmihalcea.com/feed/ Spring - https://spring.io/blog.atom Christian Rackerseder | Blog - https://www.echooff.dev/rss.xml The Nuxt Blog - https://nuxt.com/blog/rss.xml Vercel News - https://vercel.com/atom WebKit - https://webkit.org/feed Poimandres - https://pmnd.rs/feed.xml Piotr's TechBlog - https://piotrminkowski.com/feed/ David M. Lloyd - https://word-bits.flurg.com/rss.xml React Blog - https://react.dev/feed.xml JoshLong.com - https://api.joshlong.com/feed.xml Nature Reviews Computing - https://www.nature.com/nrcomput.rss OpenAPI Initiative - https://www.openapis.org/feed Biome | Blog - https://biomejs.dev/blog/rss.xml thoughts from the red planet - http://feeds.feedburner.com/ThoughtsFromTheRedPlanet Apache Fory™ Blog - https://fory.apache.org/blog/rss.xml
  25. dreamrealhrmmm. That's... not right.
  26. blueand the active feeds for this channel?
  27. dreamrealfeed subscriptions
  28. blueanyway dreamreal, could be nice to have admins per provenance, if you get to it
  29. dreamrealfeed subscriptions for irc://libera/%23primate
  30. dreamrealOkay, found a latent bug, need to fix THAT NAOW
  31. dreamrealso both features are PRESENT - try the !feed list thing, I'd msg the bot to get it, let's see what it shows you.
  32. dreamrealI'm going to add a UI element to get the feed list (the API already has it, I just never wired "show the list" into the UIs.)
  33. dreamrealAs far as admin by provenance: still quite a struggle :/
  34. dreamrealadmins affect the *bot* config, the provenance admin stuff... I mean... it'd be another level under admin, doable but not simple
  35. bluethen maybe a new role
  36. * nevet joined #primate
  37. * blue!~blue@user/blue changed the topic to: https://primate.run | this channel is publicly logged at https://bytecode.news/logs/irc/libera/primate and mirrored to Primate discord, channel #irc
  38. dreamrealblue: yeah, but new roles ripple through the permission system, it's doable but not simple
  39. dreamrealbecause THESE roles would apply to provenances, so you'd have admin here but not in #java or whatever
  40. dreamreal"does user X have permission Y for provenance Z" - not illogical at all, but not part of the system either
  41. * nevet joined #primate
  42. * blue!~blue@user/blue changed the topic to: https://primate.run | this channel is publicly logged at https://bytecode.news/logs/irc/libera/primate and mirrored to Primate discord, channel #irc
  43. dreamrealfeed subscriptions
  44. nevetPrimate Blog - https://primate.run/blog.rss bun.com - https://bun.com/rss.xml Deno - https://deno.com/feed Node.js Blog - https://nodejs.org/en/feed/blog.xml
  45. dreamrealThere you go!
  46. dreamrealhey blue
  47. dreamrealif you think it's worth it - and ONLY in that case - can you submit that new BCN article to HN
  48. dreamrealhttps://news.ycombinator.com/item?id=50022246 please upvote if you think it's worth it
  49. nevetListening Too Fast | Hacker News
  50. bluedreamreal: upvoted
  51. bluekinabalu: thanks for helping
  52. blueI wasn't hope, would have done it otherwise..
  53. bluehome*
  54. bluedreamreal: the article raises an interesting point, and I think I can incorporate some of it into primate's AGENTS.md, especially red testing
  55. bluebut I do wonder how to skirt around the model cheating. I mean, nothing prevents it from writing a red test that goes `assert.fail()`
  56. blueof course, that example is stupid: there is no input that will pass the test. but any other trivial example would do
  57. bluemy conclusion, and perhaps your should be: the most important part of working with the agent should be, it produces red tests, and *you* review them
  58. blueyours*
  59. bluebut all of this doesn't save us from the fact tests need to express a behavioural contract, and agents often treat them as a prima facie gate
  60. blueiow, passing the test doesn't guarantee correctness, it doesn't even guarantee freedom from regression, it just guarantees that the whatever the tests test, is considered passing
  61. bluedreamreal: btw I just came across this https://www.wecodefire.com/p/kotlin-keeps-up-with-rust
  62. blueinteresting stuff
  63. dreamrealyeah, that's andrew again
  64. dreamrealfascinating results: the JVM beats rust and kotlin native, and as usual the balance is memory
  65. bluewho is andrew?
  66. dreamrealJVM = more memory, better performance
  67. dreamrealkinabalu
  68. blueoh
  69. dreamreala phalange made of pure flatus, but even so
  70. blueI'm missing zig from the dhh comparison, btw
  71. dreamrealnobody cares about zig, though
  72. * dreamreal runs
  73. bluebut I get the feeling this is because people are now taking zig's andrew less seriously after his lashing out at bun
  74. blueI gotta say: I was kinda turned off down completely by that behaviour. totally unprofessional
  75. blueturned off*
  76. dreamrealyeah, it was not a great look
  77. dreamrealif you're handing out upvotes: https://news.ycombinator.com/item?id=50023409
  78. nevetKotlin keeps up with Rust in DHH's Campfire rewrite | Hacker News
  79. bluelike dude, the LAST thing you get to complain about, is some dude deciding to move from your language to rust
  80. blueif you care enough, change your ai policy or accept their patches, and if not, shut up
  81. dreamrealyeah, it's a natural consequence of choices, live with it
  82. blueupvoted, that is some great work
  83. bluedreamreal: in between that, a rant about corporations and evil bun etc. etc. I'm like, come on
  84. blueand some kind of boasting about, we don't need bun's financial support anymore
  85. dreamreal"They haven't tried to buy me hard enough, is the REAL problem"
  86. blueI mean, have some self-awareness here... they gave you money, that's kind and humbling
  87. dreamrealI would love to have fuck-you money to the point where I didn't have to care about my clientele, but... I don't
  88. blueregardless of whether they benefit from it or not
  89. dreamrealyeah
  90. bluewhich apparently they didn't btw; they need to keep a lot of in-house patches to keep it running
  91. blueneeded*
  92. bluethis doesn't take from zig being great software: but again, they're failing at the human side; this guy might be a great engineer, but he's terrible as a leader
  93. dreamrealit happens
  94. blueanyway, I kinda hopped off zig in the last ~6 months and on to the Go train
  95. bluethere are many recent indications Go is a great language for very complex stuff: esbuild, the ts rewrite...
  96. blueit's probably still impossible to write a js engine in go, simply due to the fact there's no manual memory management, but anything that doesn't need that, it's great for
  97. dreamrealI'm still riding rust because of the native bindings and the ecosystem approach
  98. dreamrealbut my requirements are a little... much
  99. blueya
  100. * kinabalu peers
  101. bluedreamreal: I *really* need that agentbox feature in repopack. I'm running around 5 opencode sessions at once, and it's very hard on this poor computer
  102. dreamrealheh
  103. bluethey're also not sandboxed and what not, and then it keeps messing around in /tmp/opencode, until it runs out inodes for that, and it starts creating stuff in ~, it's all over the place
  104. bluethe thing is, this will have to be a paid feature, because it costs real resources
  105. blueeither that, or a `bring-your-own-box` kinda thing
  106. blueso rent a vps somewhere, put `rp` on it, run `rp agent` (which supports --headless), and then you're free to go
  107. dreamrealblue: worth noting: kinabalu is a LOT better at bizdev than, uh, the guy in my chair
  108. bluebizdev?
  109. blueyou mean devops, or?
  110. dreamrealbusiness development.
  111. blueoh
  112. bluewell, when I get to the point I'm confident enough to need investors, I know who to turn to
  113. dreamrealNo, that's too late
  114. dreamrealyou need to turn NOW so you're developing something that deserves confidence
  115. bluedreamreal, if *I'm* not yet satisfied with it, how will others be?
  116. dreamrealI'm not saying that's not where you are but the process needs to be started EARLY
  117. dreamrealIT is composed of two main things: the idea and the implementation
  118. bluemhm
  119. dreamrealbizdev is about the idea, the implementation is almost secondary
  120. bluethat's fair
  121. dreamrealFOR EXAMPLE: nevet. The implementation is almost ALL of it, because the project lead is a coder and not a business guy
  122. dreamrealand thus it's really good at the implementation but who'd want to buy it? Why or why not?
  123. blueya, I'm not a business guy either
  124. dreamrealand FWIW andrew was pointing that out from early days and he is not wrong about it
  125. bluemhm
  126. * skillbot quit (Remote host closed the connection)
  127. * skillnews quit (Remote host closed the connection)
  128. * skillbot joined #primate
  129. * skillnews joined #primate
  130. bluedreamreal: I will contact him after I'm done with a few things, including this chat feature
  131. dreamreal*nod*
  132. bluedreamreal: I gave you a reader role for the repopack project
  133. bluehttps://repopack.com/project/repopack/tasks/555
  134. bluethis is kinda brilliantly reported by AI, I believe. really good job there
  135. blueI had to stop the flog agent: it started implementing around it; I told it to stop and report the gap upstream
  136. bluethis is something that is really hard to teach a pleasing machine. it doesn't ask me almost ever if I want to stop and report the gap upstream, it just goes around it
  137. blueI had another thing just now, I told the agent to analyse #555, fork off follow-ups as stacked tasks, and define the contract, and it went ahead and parsed that as "do it yourself". I had to stop it
  138. bluegonna be better with chat and the 'controller' agent, but we're a few leagues away from there atm
  139. dreamrealsure
  140. blueseeing the road, seeing the distance, and not being there yet, is hard, lemma tell you
  141. * dreamreal is familar
  142. bluedreamreal: ironically, the ineffectiveness of my work right now *prevents* me from burning onto pro 100 in less than a day (~1.5d). once chat is done, I expected this to go down to roughly 0.5d, or be mostly bounded by the speed of the agents themselves
  143. bluebut I'm well aware that once chat is done, I will probably need pro 200, which does not make me happy
  144. bluepro 200 is x4 pro 100: so if I get down to 0.5d, that's still 2.5d in a 7d timeline. even with the limit busters, which usually come once a week that's still 5d / 7
  145. blue2d*
  146. blueso we'll have to see
  147. bluedreamreal: https://www.youtube.com/watch?v=XAaG0jalyjg
  148. nevetYouTube: Texas Lightning - No No Never | TexasLightningVEVO
  149. skillbot04,01► 14,01YouTube :: Texas Lightning - No No Never
  150. bluewee bit of entertainment
  151. blueshe has such a nice, calming voice
  152. bluestacks are so nice
  153. blue`rp job run 7 --source 2ce06f2cf64b24aaae5d6a39d2411fbc2e18f5db --build-arg FLOG_VERSION=0.1.0-job-test --attach`
  154. bluedreamreal: rp is *super* nice
  155. bluebut I'm a sucker for clis, for there's that
  156. blueflog 0.1.0 is building