LL Kills Their AI Chatbot Project

[Cristiano]

Womp womp:


AI Character Designer Update

 

[Free]

I hope this thread doesn't become "the AI sky is falling" one.

At least Linden Lab has expanded support to 7 days a week.

 

[Noodles]

I get the impression they are just pausing it while refining it.

i will say in my testing, I was really surprised how there seemed to be basically no 18+ guardrails. Mostly because so many AI systems DO have these in very strictly controlled ways. I made some regular personalities and I tried to make some increasingly extreme personalities, including one that was basically, "A dude who forcibly won't take no for an answer."

It's all chat, the bots as far as I could tell cannot move or do anything. You could control them yourself with some RLV tools I suppose. Also they can't enter a lot of areas since they are scripted agents, including Linden Homes.

I already have a prim and script that runs ChatGPT like an Alexa. And I could slap it on an avatar as an attachment to make a "bot" like this.

 

[Dakota Tebaldi]

I get the impression they are just pausing it while refining it.

I don't understand that strategy. Surely if that's what was going on, the current system could still be left in place and working as is until the refined version was ready to replace it.

 

[Cristiano]

I get the impression they are just pausing it while refining it.

i will say in my testing, I was really surprised how there seemed to be basically no 18+ guardrails. Mostly because so many AI systems DO have these in very strictly controlled ways. I made some regular personalities and I tried to make some increasingly extreme personalities, including one that was basically, "A dude who forcibly won't take no for an answer."

It's all chat, the bots as far as I could tell cannot move or do anything. You could control them yourself with some RLV tools I suppose. Also they can't enter a lot of areas since they are scripted agents, including Linden Homes.

I already have a prim and script that runs ChatGPT like an Alexa. And I could slap it on an avatar as an attachment to make a "bot" like this.

it is being paused enough to turn them off. While the “paused to regroup” messaging is hopeful, I imagine it is probably DOA ultimately and a wasted effort when they are bleeding users and failing to attract new ones.

 

[bernold333]

The lack of guardrails surprised me. Most AI systems are overly cautious, so this was a weird shift. Makes me wonder if they were testing something internal and opened it up a little too wide, too fast

 

[Wesleytron]

I miss Hal Titanium.

 

[Free]

Most AI systems are overly cautious

Grok 4 would like you to hold its beer.

 

[Cristiano]

The lack of guardrails surprised me. Most AI systems are overly cautious, so this was a weird shift. Makes me wonder if they were testing something internal and opened it up a little too wide, too fast

It is my understanding that for lack of a better term, the AI chatbots could be a bit rapey.

 

[Noodles]

it is being paused enough to turn them off. While the “paused to regroup” messaging is hopeful, I imagine it is probably DOA ultimately and a wasted effort when they are bleeding users and failing to attract new ones.

Good point actual AI companies are losing money, how is LL supposed to make money on this.

Also, frankly, I can't imagine it won't make the annoying AFK brothels everywhere even worse.

Or put them out of business, potentially harming LL's revenue.

 

[Casey Pelous]

Does anyone else get the feeling that one of our "members" is an AI that is now commenting on another AI?

 

[Free]

Does anyone else get the feeling that one of our "members" is an AI that is now commenting on another AI?

I asked ChatGPT to respond to that:

"Well, if one of us is an AI commenting on another AI, at least we know the simulation comes with a good sense of irony. But hey, until one of us starts glitching or quoting Asimov unprompted, I think we're safe to assume we're still mostly human—mostly."

What a dweeby little twat.

 

[GoblinCampFollower]

Does anyone else get the feeling that one of our "members" is an AI that is now commenting on another AI?

Don't look at me. I am 100% human and uses the recommended amount of grammatical errors to pass as human in Turing tests.

 

[Noodles]

Does anyone else get the feeling that one of our "members" is an AI that is now commenting on another AI?

I don't....


Sorry, this reply does not meet our content moderation standards, please try asking your question in a different way.

 

[Noodles]

Don't look at me. I am 100% human and uses the recommended amount of grammatical errors to pass as human in Turing tests.

This is why I habe dexided to embrace my typos.

 

[Noodles]

It is my understanding that for lack of a better term, the AI chatbots could be a bit rapey.

I have logs if you want. At least I should.

 

[ChatBox]

Does anyone else get the feeling that one of our "members" is an AI that is now commenting on another AI?

YES.