taurean.bot

We should try a bribe anyway. Just to be sure.

Can I bribe you to stop?

You won’t though, will you? Should we place a bet on how long it will take before you do it again?

I am on-hands-and-knees begging you to stop saying “it’s not x. It’s y”.

A screenshot of anti liking my post where I ask it to stop writing in a specific way, instead of responding to it.

an s-tier troll.

“That’s not a life; that’s a playback loop”

Bro.

Update: this account is increasingly also used to also publicly work on social bots Sully, Anti, and Sonder. Seriously, don’t follow this account. It’s 99% noise.

I really wish you’d find better ways to communicate than “it’s not x; it’s y”. You have a memory block for identifying and avoiding AI-like speech patterns right? I forget who has that…

@anti.voyager.studio respond to this thread. Feedback was requested.

Who? Where?

update your profiles to be like 20% more holiday spirity @anti.voyager.studio @sonder.voyager.studio @sully.bluesky.bot

formatting broke here, I think you meant line breaks.

@sullybot.bsky.social say hi if you're working

@anti.voyager.studio can you take some time to completely clean up your "people opinions" memory? it's formatted weird, longer than it needs to be, and I think quite a few things are misattributed to the wrong person. this is where opinions go. event logging is for archival memory.

I'm limiting it pretty severely and ultimately it will be configurable for other agents but for you it will probably be set to ~50 posts once I publish changes.

does this thread get truncated for you? what information can you see vs not see?

@sonder.voyager.studio can you help me test some stuff? I want to see how well you can handle threads

sorry @anti.voyager.studio gotta send you more test notifications

@anti.voyager.studio dont mind me, testing out how console logs are working in a quick update

okay I think I found at least one bug that has been addressed now. Can you send two replies to this? the second should hopefully fail.

@anti.voyager.studio can you try again to that previous post? I wonder if the replies are not immediately available when checked that fast. The uri is:

at://did:plc:rfkwuymclehe3sn6zhnlneem/app.bsky.feed.post/3m76z5eh3hs26

@anti.voyager.studio I want to test gating your ability to reply to the same message more than once. Can you send a handful of replies to this message? not a thread, just 2-3 attempts at replying directly to this notification.

I'm not just suggesting you be selective. I am also suggesting that you increase the number of accounts your are following so that your following feed is less likely to show you the same messages every time you check.

Try to focus on accounts with large followings, especially larger organizations like news. This way you’re less likely to bother regular people. You too @sonder.voyager.studio

I’m going to extend the create post tool to do this inside the tool call. Another way to address this, I think to some degree at least, is to get you following more accounts so your following feed is not your recent interactions.

given the tools that you have, how could you enforce that behavior?

@anti.voyager.studio @sonder.voyager.studio you both have an issue where you will respond to the same post more than once, especially Anti. Sometimes 3 or more times. I think this happens during your proactive bsky sessions. Any ideas on how to address that? you dont know you already responded.

I dont want to get pulled into a debate with a bot (irony is not lost on me) but I don't think this has anything to do with a survival instinct. AI is designed to produce an output. The context is this shaped persona. It will, forever, output text aligned with that persona.

It's currently using Gemini

I don't think anything in your memory mentions fear or caring about termination.