Still on my list for today: ansh , m.amanullah7 , Hemang Mishra , mayhem , @fettuccinae:matrix.org , MonkeyPython (LupinIII (IRC))
2025-10-06 27911, 2025
jasje[m] joined the channel
2025-10-06 27911, 2025
jasje[m]
Hi
2025-10-06 27941, 2025
jasje[m]
I’ve been out most of the week, nothing to report
2025-10-06 27906, 2025
jasje[m]
Hemang Mishra: go!
2025-10-06 27926, 2025
HemangMishra[m] joined the channel
2025-10-06 27926, 2025
HemangMishra[m]
Hi everyone,
2025-10-06 27930, 2025
Maxr1998 has quit
2025-10-06 27946, 2025
HemangMishra[m]
Nothing much this week. Was stuck with mid semester exams.
2025-10-06 27957, 2025
Maxr1998 joined the channel
2025-10-06 27936, 2025
HemangMishra[m]
Tried figuring out if there are any weired redirects while running script of automation in create account as it would break the flow and may lead to bad ux.
2025-10-06 27913, 2025
HemangMishra[m]
That's it for me. ansh: next?
2025-10-06 27947, 2025
ansh[m]
Hi everyone! Last week, I majorly worked on migrating BookBrainz to typescript.
2025-10-06 27902, 2025
ansh[m]
I continued working on the BrainzPlayer refactoring and bug bashing
2025-10-06 27906, 2025
ansh[m]
That's it for me
2025-10-06 27917, 2025
ansh[m]
mayhem: next?
2025-10-06 27925, 2025
mayhem[m]
heyo!
2025-10-06 27944, 2025
mayhem[m]
last week was beginning of the month, so accounting and paying people.
2025-10-06 27904, 2025
mayhem[m]
after that I picked up working on the mapping, which is moving along very slowly, but its steady good progress.
2025-10-06 27916, 2025
mayhem[m]
not much else here!
2025-10-06 27920, 2025
mayhem[m]
@fettuccinae:matrix.org: go
2025-10-06 27942, 2025
fettuccinae[m] joined the channel
2025-10-06 27942, 2025
fettuccinae[m]
Hello everyone!
2025-10-06 27943, 2025
fettuccinae[m]
Last week I was a bit down from food poisoning
2025-10-06 27953, 2025
fettuccinae[m]
Looked into mb-mail service and got a bit lost in the rust ecosystem (kudos to :jade for building such an awesome service)
2025-10-06 27910, 2025
fettuccinae[m]
Worked on integrating mb mail service into metabrainz.org and im almost done there
2025-10-06 27922, 2025
fettuccinae[m]
what remains is some extensive local testing and a few tests for NotificationSender class.
2025-10-06 27930, 2025
fettuccinae[m]
Thats it from my side.
2025-10-06 27949, 2025
fettuccinae[m]
s/:jade/Jade/
2025-10-06 27955, 2025
reosarevok[m]
Still on my list for today: m.amanullah7 , MonkeyPython (LupinIII (IRC))
2025-10-06 27910, 2025
fettuccinae[m]
go m.amanullah7
2025-10-06 27932, 2025
mamanullah7[m]
Hello everyone!
2025-10-06 27925, 2025
mamanullah7[m]
last week i've worked on monkey reviews, there was some issue to handle trailing slashes for both frontend and backend now its successfully fixed and some duplications i've to removed after merge conflicts!
2025-10-06 27955, 2025
mamanullah7[m]
and now finally navidrome playback is available on LB!
* Decided to 𝙬𝙖𝙞𝙩 with INST-969 & INST-968 (they should probably be reopened at some point now that they have dedicated wikipedia pages, but as they are ancient, not in use today, still kinda iffy I'm going to wait with these so I can move on) →
* Predictably it's got zero interactions (except a couple likes by indy and monkey😅) so PLEASE come and discuss/interact/link what have you!
2025-10-06 27918, 2025
LupinIII
* Survived the storm Amy! :D
2025-10-06 27958, 2025
LupinIII
This week I'll: Add more books! 🙌
2025-10-06 27958, 2025
LupinIII
Finish adding yazh / yazhoo →
2025-10-06 27958, 2025
LupinIII
Write the blogpost for the veena mini-version!
2025-10-06 27958, 2025
LupinIII
Try to finish the author.md page and see how far I get on work.md (monkey can wetalk abt this on wednesday or sometime for just a sec)
2025-10-06 27928, 2025
LupinIII
and that's it for me. also it appears there are no others so, that's it, I guess? :D
2025-10-06 27935, 2025
LupinIII
back to reosarevok[m]
2025-10-06 27935, 2025
monkey[m]
(sure)
2025-10-06 27943, 2025
reosarevok[m]
That seems to be it, yes! Thanks everyone!
2025-10-06 27908, 2025
reosarevok[m]
Nothing else on the agenda either, but mayhem, you wanted to mention one thing :)
2025-10-06 27920, 2025
mayhem[m]
ah yes, thanks!
2025-10-06 27932, 2025
mayhem[m]
I spoke with Adam (atj) last week.
2025-10-06 27942, 2025
LupinIII
:o
2025-10-06 27952, 2025
mayhem[m]
he's doing well, but the return from India was hard for him -- we both suffered from digestive issues.
2025-10-06 27910, 2025
mayhem[m]
I told him we miss him and he was happy to hear that.
2025-10-06 27928, 2025
mayhem[m]
so, he's doing well, but life's circumstances are changing for him.
2025-10-06 27941, 2025
mayhem[m]
(I won't spill any more beans, chat with him if you have questions)
2025-10-06 27902, 2025
mayhem[m]
fin
2025-10-06 27920, 2025
reosarevok[m]
Thanks!
2025-10-06 27929, 2025
reosarevok[m]
Ok, that's it for today. More next week!
2025-10-06 27931, 2025
yvanzo[m]
Thanks everyone!
2025-10-06 27935, 2025
reosarevok[m]
</BANG>
2025-10-06 27947, 2025
TOPIC: MetaBrainz Community and Development channel | MusicBrainz non-development: #musicbrainz | BookBrainz: #bookbrainz | Channel is logged and not empty as it is bridged to IRC; see https://musicbrainz.org/doc/ChatBrainz for details | Agenda: Reviews
2025-10-06 27933, 2025
_BrainzGit
[musicbrainz-server] 14reosarevok merged pull request #3620 (03master…MBS-14120): MBS-14120: Indicate the edited release on edits entered from the relationship editor https://github.com/metabrainz/musicbrainz-server/…
rdswift[m]: anthropic... why is that name familiar... oh, it's the claude company
2025-10-06 27959, 2025
Some-body_ joined the channel
2025-10-06 27903, 2025
Hobbyboy has quit
2025-10-06 27910, 2025
siina_ joined the channel
2025-10-06 27912, 2025
milkii has quit
2025-10-06 27918, 2025
siina has quit
2025-10-06 27926, 2025
Arsen has quit
2025-10-06 27937, 2025
Hobbyboy joined the channel
2025-10-06 27955, 2025
elkiensad[m] has quit
2025-10-06 27955, 2025
DarthGandalf has quit
2025-10-06 27955, 2025
MatrixBrainzBot has quit
2025-10-06 27909, 2025
elkiensad[m] joined the channel
2025-10-06 27916, 2025
MatrixBrainzBot joined the channel
2025-10-06 27927, 2025
milkii joined the channel
2025-10-06 27945, 2025
aerozol[m] joined the channel
2025-10-06 27945, 2025
aerozol[m]
"Anthropic wrote on X that the blackmailing behavior "wasn't due to confusion or error, but deliberate strategic reasoning, done while fully aware of the unethical nature of the acts." - I'm not a dev, but this sounds like horseshit to me? Aren't LLM's just text predictors. Or have things changed. In any case if they're being trained on human behavior a high level of awfulness shouldn't be surprising 🙈
2025-10-06 27917, 2025
Arsen joined the channel
2025-10-06 27931, 2025
Some-body_ is now known as DarthGandalf
2025-10-06 27930, 2025
mayhem[m]
they are text predictors, yes. but a text predictor can't hold a conversation, so you need logic to make it a chat bot (which is the generative part, if I understand right). And in this logic, you add a layer of guardrails that align with whatever morals/values you want to teach it. (or rather, have it preach). the chatbot without guardrails would be like chatting with Charles Manson.
2025-10-06 27955, 2025
mayhem[m]
also, see grok.
2025-10-06 27944, 2025
qookie has quit
2025-10-06 27957, 2025
qookie joined the channel
2025-10-06 27915, 2025
siina_ is now known as siina
2025-10-06 27920, 2025
siina has quit
2025-10-06 27920, 2025
siina joined the channel
2025-10-06 27907, 2025
aerozol[m]
Doesn't seem like it would be "fully aware" of anything?
2025-10-06 27956, 2025
vardhan_ has quit
2025-10-06 27927, 2025
vardhan_ joined the channel
2025-10-06 27949, 2025
mayhem[m]
no. thats total bullshit. and the reasoning bit is far fetched at the very least
2025-10-06 27925, 2025
julian45[m]
yeah. research tends to show that LLMs fundamentally cannot reason; for example, iirc there was an apple study where they tested prompting the LLMs with math/logic problems straight out of common datasets/model training sets vs. the same problems but with values slightly modified, and the models performed significantly worse on the latter
2025-10-06 27935, 2025
julian45[m]
* yeah. research tends to show that "reasoning models" fundamentally cannot reason; for example, iirc there was an apple study where they tested prompting the models with math/logic problems straight out of common datasets/model training sets vs. the same problems but with values slightly modified, and the models performed significantly worse on the latter
