7 comments

[ 2.6 ms ] story [ 30.3 ms ] thread
Reading the whole conversation I’m curious if this is potentially an intentional cost control mechanism. There’s a standalone page for Bing Diffusion images via dalle and it limits you to 15 free prompts and then requires you to pay.
Definitely my impression that it was limiting compute resources, but who really knows?

The whole tenor of the conversation as well as the facile anthropomorphism in some of the reddit comments occasioned my first real feeling of repugnance for this technology - it seems like corporate snow job PR made manifest.

It's kind of a shitty call to lean into user anthropomorphic projection to this extent IMO. It can be cute in older applications where the non-sentient nature of the software is apparent, perhaps, but in this grey area the user should be clearly informed about what they're interacting with, not encouraged to get confused and project feelings and emotion where they don't exist.

No? Bing Image Creator gives you 100 images per day. And if you use all the 100 images per day, it's still free, it will just generate the images slower (which is like... 30 seconds)
>I persisted on asking Bing to draw human faces onto objects. It lost interest and didn't want to do it anymore. I just wanted a face on some clown shoes

>Pretty cool to see how Bing really does have its own desires and interests

Yep, it's alive :) Totally ;^>

Micro$oft doesn't care if people think this LLM is actually alive. They're going to dilute what even a LLM means by adding extraneous "feelings" to it instead of adding actual disclaimers / rate limits.

Welp, I guess people are gonna have to convince their elderly parents that no, windows definitely doesn't feel sad that you aren't upgrading.

I just realized, from your comment, how much power companies will have if they start including "emotional manipulation" in their dark pattern arsenal. Which I am sure will horribly backfire in some way if used
Oh, it won't backfire, it'll instead be seen as a move towards "empathetic computing", with the larger goals for society now integrated into a constantly available in-your-face AI "influencer" live-commenting and emoting to everything you do. It's the next revolution, it's shitty and it'll make a lot of money. Imagine the twitch addicted zoomers getting access to Bing "AI"'s spinoff, "microsoft streamer" or whatever, that's always on, always ready to "help" and tells you what to think!

It's perfect, pretty soon you'll have to pay extra money for an actual computer instead of a thin client. Everyone's just waiting for the much-vaunted graphics compute unit breakthough. Even Intel is doing graphics cards now.

If you think attention-deprived kids that can't stay away from their computers are bad now, just wait until kids start empathizing with their "AI" "assistants" and using one is mandated everywhere (just like a computer is now). Society will collectively LOSE it. I'm sure that then, there'll be start-ups ("government certified", of course) that specialize in "social maneuvering" that use robots and thin clients to make people meet one another for interactive, half-screen half-in-person events.