@deepfates
Yes I’m constantly rubbing my finger in small circles while thinking. It’s like my finger is a stone skipping over metallic water or something
The new Meta AI will create images as you type, give you the option to animate the images it creates, and then lets you create a video of the images that were generated while you typed.
Pretty cool stuff.
@adcock_brett
@johnvoncolln
I’ve thought to myself for years, if you can drop a pile of unfolded laundry in front of a robot and have neatly folded stacks of clothes within an hour.. then the future has arrived.
For those who didn’t know, we sell merch designed specifically for techno-optimists.
Here’s a preview of our upcoming collection - “UTOPIUM”
Would you ever cop some?
@skirano
GPT-4 is definitely the better debugger, but debugging might be the *only* thing I’d argue it’s better at.
Claude Opus is able to understand complex jokes when GPT-4 is not.
@yacineMTB
Dude Ive been thinking about this daily since GPT-V. A vision model should be able to embed the dom and then see elements and know their coordinates.
An AI being able to use a computer in this fashion makes Sam Altmans vision of an AGI being an “average co worker” very feasabl
@AISafetyMemes
@sama
Just struck me that we might actually start thinking in this way. BA and AA.
Before AGI and After AGI
Maybe even BA (before AGI), AABS (After AGI, Before ASI), and AA (After ASI)
@SmokeAwayyy
Name: Caelux
Voice: TARS from Interstellar
Personality: Sarcastic and witty, brief with answers (signal > noise), insightful if you probe it
Appearance: 2D ascii art robot
Take just a still frame from one of the SORA generations, and its somehow better than Dalle 3.
Does training on video (as opposed to images) allow for a more coherent output?
Say I wanted to build an at home 4090 rig to run some open source small models -- image, language, video, voice, etc.
Does anyone know of a good tutorial for how to go about this?
"yacine how are you going to hit your mrr KPIs??"
HOLY SHIT!!!!! DID YOU SEE THAT!!?!?!?!?! NEW ON DINGBOARD DOT SHOP!?!?!? HOLY SHIT I AM SOYFACING RIGHT NOW?!?! I CAN GET A SHRIT WITH A GUY WEARING A SHIRT THAT SAYS I ❤️ AGI???
go to dinbgoard<dot>shop NOW
@Icalbert101
A movie or TV show about a powerful AGI that’s able to bring humanity to a utopia would crush. The backlash from EAs would just promote it further
@mattshumer_
All I’ve heard is that context window starts falling apart after 8k tokens. Is anyone trying needle in a haystack tests for these longer lengths?
quick schizo theory:
top and bottom quarks were originally called "truth" and "beauty" quarks
"satya" in sanskrit = truth
"sundar" in sanskrit = beauty
the 2 mfs competing in the race to AGI are named after opposing fundamental particles
the simulation is fucking with us
I’m convinced that this new “Udio” tool is an AI music generator on par with Suno that’s paid a bunch of influencers to generate hype about how much “better” it is.
@multivrsstudios
@sama
This is probably very possible now, but the generated audio would probably be more “ambient” noise rather than noises synced with the events in the video.
A solid voice assistant is a trillion dollar idea.
Right now, you can have a pretty good voice chat experience with chatGPT’s voice mode, but it still feels like you’re lobbing a message over a wall and waiting for it to come back.
That’s not how humans converse.
Humans give
@julesterpak
I really think you’re onto something with this tik tok edit stuff. I honestly got inspired to learn how to make them after watching your YT vid on the topic.
Since my brand is all about being optimistic for the future, I’m gonna try to carve out a new niche called “future core”
@firesta27536185
So reality is a fractal. That the idea I usually land on. Trying to fathom how the first intelligence emerged in bass reality will literally drive you mad.
But then again, maybe our brains are only limited to thinking things have beginnings and ends.
A decade ago when I first watched the movie “Her”, I thought to myself “this technology will just be an smartphone app, why does he use a separate device?”
I’m starting to realize that the producers of Her might’ve been spot-on.
It's telling that every AI companion company is building their own hardware (Rabbit, Humane, Meta, Tab, Rewind). Current state of mobile is just too restrictive for big ideas.