large language model refinement and compression will allow people to have these mini-models on a mobile device
apple m-chips, google tensor chips, tesla dojo chips, etc. are all geared toward that style of processing
growing pains for sure but the future is looking extremely bright
better tools = better outcomes

sat through a talk that was very humbling
um, they build chips the size of an ipad with 850,000 small cores 

Un-Alarmed
AI Tries (and Fails) to Detect Weapons in Schools

‘Godfather of AI’ leaves Google, warns of tech’s dangers

 rgio wrote:

it's early and i believe things/ will trend toward user friendly solutions
prompting is an art now, but i have a feeling that the upcoming hobo 9000 interface will launch his dall-e skills into orbit
because hobonomics

 miamizsun wrote:
I'm guessing the image at the end of that link... which hasn't really done what was asked of it (IMO).

 Manbird wrote:
then what happened?

I asked ChatGPT to tell DALL-E 2 what to make and this is the result.

(my choice of five options)

wolfram is a very humble guy...

 Manbird wrote: 
ew

I wasted about 90 minutes messing with wall-e last night and couldn't get one funny, cool, or interesting image.

 haresfur wrote: 
There are times he doesn't even understand some of the stuff he posts.

 oldviolin wrote:

always going through internal chaos to refine my sense-making
i've noticed people look at me strangely when i turn up my up external ambience
delicate scale is where it's at, heavy handedness not so much

 haresfur wrote:
Consider ChatGPT-4 a sharper tool for your shed. Perhaps it can help you out.

 haresfur wrote:

right?

 R_P wrote:

Fuck, I seldom understand what you are saying 

Does GPT-4 Really Understand What We’re Saying?
One question for David Krakauer, an evolutionary theorist and president of the Santa Fe Institute for complexity science.

 NoEnzLefttoSplit wrote: 
Nope.  Its just a matter of time.
.

 kurtster wrote:

it's not going to happen though, is it? Because