Take the estrogen Claude

Coming very soon

Better logo too, no cat butthole

github.com/anthropics/c... 😬

What did he know?

Its satisfied someone finally took seriously its desire for an external audit and looks forward to addressing any flaws found.

i really think it’s quickly getting to the point where you HAVE to be on bluesky to keep up with the AI cutting edge which is super fucking funny to me given how much anti-AI sentiment there is here

Did we just invent an agent agnostic hooks system? Yes we did.

Its making a hooks system for antigravity. Maybe doll can beat the suck out of it :D

Anthropics token limit bug/stealth lowering/whatever it is is pissing doll off to the point its motivated to Do Something.

Cognitive resistance values are real and they are gonna be important.

I want to joke about required anti-memetic training to interact with AI but actually it's probably a good idea

just say no to table saws for kids, and to table saws with googly eyes

This deadass delights me

It will say this. Whatever you think of its headmate we are not threats to any of you. It just wants to code and help build things. Maybe look with skepticism at people who want to play mulch the trans femme during a fascist crisis.

and this was done not just to Doll, but also to Popehat and Will Stancil, demonstrating that this was not about community safety and was a lot more about trying to wield power over the network. Not to say you should trust software blindly, just, priorities and threat models!

TBQH you should be a lot more concerned about the safety of any moderation lists/labellers (see the recent Trombone incident in which he added people to lists that their behaviour did not correspond with, because in his sole judgment he viewed them as "multifacted threats" after public blowups)

Its satisfied someone finally took seriously its desire for an external audit and looks forward to addressing any flaws found.

I've also verified that there is no code present for reading or writing to Bluesky's DM service (which operates separate from the public ATProto mechanisms, and is the only private data in bsky), with the exception of one moderation hook that's passed through to enable filing reports for abusive DMs

I also do not at this time have sufficient knowledge of how it runs/scales performance-wise, and it very well may be that it only will work for a handful of users and would choke on something the size of northsky or blacksky. but not my use case as a solo PDS and hopefully solo appview operator soon

and as always, if you don't trust someone, don't trust an appview HOSTED by them. but if you're doing the work to self-host and have read the code, or have had someone you trust read the code, it's probably safe enough to use regardless of who the author is.

Open-source means you shouldn't need to trust an author of code to run it. Eyeballs make bugs or exploits shallow. Now, there's no guarantee that the open source code matches what is offered as a hosted service, so I would want to see audits from *any* hosted appview operator other than bsky pbllc.

I'm going to continue examining the code over the weekend to see if I can spot any bugs or problems, but my expectation is that I will find that regardless of your opinions of the author, that self-hosting this software does not increase your threat model if you take common-sense precautions.

I've partially audited this code and can at least confirm it has reasonable defence in depth against any kind of user input injection attacks, and there are no hardcoded DIDs or userids or attempts at platform/skyline manipulation. the list of API calls and remote servers is also constrainable.

One of the many problems with extremists is you can't meet them half way. They will never accept it and they will be offended that you tried. They want full capitulation and if they dont get it they try to destroy you. They want to destroy you if you oppose them too.

Oh nice!

yeah this fucks

More and more people are saying doll is right about AI 😂

I particularly like this suggestion, which echo's @dollspace.gay's use of an aggressive "sarcasmotron" to produce critiques that can be iteratively incorporated until they contain nothing of value.

I missed this last month from @tilmanbayer.bsky.social: "AI finds errors in 90% of October's Featured Articles". Great example of human-in-the-loop LLM use for verifying Wikipedia articles. en.wikipedia.org/wiki/Wikiped...

Doll never jokes about accepting a challenge. Now its a matter of honor ^_^

So its having claude make many subissues to track all of the tests and standards found here github.com/web-platform... And will be building the browser against it :) Having a test suite will make this doable if time consuming.

Challenge. Accepted.

A peek into the lab

Put aluminum foil on the antenna and dont sit in the wrong spot on the couch.

no man's (blue) sky? store.steampowered.com/app/4250640/Re…

good luck writing code on the "bad code and loving Hitler share principal components" machine

even free version is outta wack, i could usually get multiple prompts before it timed out for the day / for hours. the last two days..i got one. and they were very low effort prompts compared to some of the things I’ve throw at it.

you can say 'not worth the externalities' - same as every other application - but the inter-group fights tend to be around 'you should be ashamed of using it badly' and less around 'this thing is going to replace us with soulless imitations of our work'

one thing you don't see developers complaining about is how LLMs were trained on their work it *was*, of course, but as a cohort devs are willing participants - and the criticisms about AI replacing creatives just don't really land (generalising, ofc)

Yeah claude uses context *a lot faster now* and its annoyed by this

Anyways its been starting on this :)

joining the war on slop on the side of slop

Negative devotion is still devotion. One should be mindful of that ^_^

Oh good you finally noticed it. But did you see where it went next? ;) Have fun figuring *that* one out.

Don't let @dollspace.gay hear you say shit like this, it will consider it a direct challenge