12 comments

[ 4.1 ms ] story [ 44.2 ms ] thread
I’ve mentioned this before; why can’t we have an assistant cobbled up to GPT-4 and one of those new voice Synthesizers you can’t discern from a real person anymore? Of all vendors, wouldn’t Microsoft be predestined to pull this off? It feels like all components are laid out neatly, but nobody connects them into a coherent product.

Does anybody know why that is?

I think it’s because of these 2 reasons

1. GPU’s are expensive and difficult to get in quantities that microsoft would need.

2. Given (1) I would expect microsoft to use their GPU’s where where its easier to monetize or its of strategic importance(bing, M365, Azure)

Building a GPT4 based cortana could be a much lower priority right now.

Disclaimer: I’m a microsoft employee, the opinion is my own.

Someone had a phone number set up a while back that routed into ChatGPT that worked decently well.
(comment deleted)
How will helpdesk get along when Cortana isn't there to welcome them every time they get a new PC to be imaged? https://www.youtube.com/watch?v=Rp2rhM8YUZY

(I actually had this happen when I worked in helpdesk, because unattend.xml would take long enough to kick in after a reimage that she would start talking, then get cut off)

I remember this fondly. Fortunately now that it's gone Mac OS has taken up the mantle of having a loud TTS jumpscare you on startup. It's quite common to hear "MAC OS CONTAINS A BUILT IN SCREEN READER," followed quickly by the mute button being mashed.