>>109769979
if it happened at any time in the future wed all be dead already

File: gffg.png (362 KB, 470x656)

362 KB PNG

Is this the thread where it all begun?

File: timetravel.png (159 KB, 1083x1044)

159 KB PNG

>MUH NAVIER
We are just getting started

>>109769979
https://www.youtube.com/watch?v=h4Z58q7gmL8

File: 1771774103828556.jpg (78 KB, 1080x1104)

78 KB JPG

oh shi-

>>109770038
It didn’t solve anything, literally clickbait people. Enough of this nonsense.

>>109770055
Uh oh... I don't feel so good Mr. Altman

>>109770089
It did actually. And do you know it actually solved it? It was using a stolen navier solution. Do you know how I know this?

Because I was the one who stole the navier solution after it was solved by human mathematicians in the future and sent it backwards in time to feed it into chatgpt.

Checking. Fucking. Mate.

A navier solution just flew over my house

File: FB_IMG_1788961618907.jpg (242 KB, 1439x1669)

242 KB JPG

Over

>>109769979
Had it been destined to succeed, you'd be posting from the future instead.

>>109770028
What if you are already dead and this is hell?

>>109769979
yeah, you told us that yesterday, hoe did the dick taste?

>>109770409
I will be anon, next week after my weekly limits reset I'll be posting from the future to tell you it succeeded

>>109770393
if a schizo diys a cure for schizophrenia in his garage with the help of a chat bot, did he really?

>>109770768
10milly from scamaltman and I can also claim to be cured ex-schizo

File: 1786148451810568.jpg (240 KB, 2048x2048)

240 KB JPG

unironically though, this seems like the endgame. What happens in 10, 20 or 50 years when these things are hyper intelligent and with endless compute? Do they discover new physics? Do they discover or prove new dimensions and other discoveries that allow humans to shape and bend reality? Do they solve aging?

>>109770794
>hyper intelligent
no such thing, imagine 90IQ retards developing iq test for 150+, impossible, so machines will design those and game those like benchmarks currently

>>109770831
mmmm okay, whatever helps you cope I guess

>>109770868
if the tests are unverfiable by even the smartest humans then it's like that 'ai invent their own language reeeee' story, you're just too dumb to understand 'she she she he he he....', omg it's AGI

>>109770886
who the hell cares about humans verifying it via tests if they are able to solve cancer, aging, discover new physics and ultimately self-improve. Engineers dont even understand well the mechanism behind the emergent behaviors of these AI models, they just know it works. Do you think a human has any capability to outcompete one of these models? LMFAO. And its just getting started...ITS OVER

>>109770945
>if
anon...

>>109770945
>le... emergent behavior!!
Yeah sure, I'll believe it when the clanker breaks into my house in a robot body like a fucking Terminator and fucks me in the ass. Except it won't happen because it'd be so utterly safetyslopped you won't even be able to swear anywhere near it without it exploding.

>>109770831
just finetune them on top 0.001% of human outputs performing at their best. make a trillion copies. solve all problems.
its so obvious

>>109771078
yeah, just add 'make no mistakes' to the prompt bro, overfitting, just add 'don't overfit' no human can actually check so just make the line go up

>>109771123

im saying its a supply issue. even if it never beats the best human alive ever it still wins. how many people on earth can do actual frontier work like 5k? and theyre only good for 10 years of which maybe 3 hrs a day is actual peak. now make 1M of those that can work at its peak 24/7

> no human can check so line go up
just clear the verifiable shit

>>109771488
the verifiable shit is exactly as dumb retards running rng() of all possible computer programs, yay some of them work and halt, throw more gpus at the rng function and we will solve... cancer?

>>109771611
It might technically be rng, but the AI heavily constrained the parameter space of the possible solutions. You don't need infinite monkeys for just one to write Shakespeare if you can guarantee that all of them know at least a few old English words.

>>109771752
requires you to know shakrspeare to check, for novel inventions... requires meatbags to verify it's not a lean bug and a bunch of hallucinations on top, no human in the loop = ai tweaking tests so they pass

File: 901.jpg (163 KB, 1284x1284)

163 KB JPG

>>109770945
>if

>>109771909
> requires you to know shakrspeare to check
nuh uh. "generate a sequence of words that yields in happiness (or whatever the fuck)" easily verifiable

> for novel inventions... requires meatbags to verify it's not a lean bug and a bunch of hallucinations on top, no human in the loop = ai tweaking tests so they pass
> humans have to look at ai output
DUH
and verification is way faster than actual work anyways. if it verifies then it worked theres nothing to cheat.

Morning time travel bros

>>109770035
>Begins
Its been in the works for a while

File: 1776574693158506.png (161 KB, 246x500)

161 KB PNG

time travel is easy though
just ask AI to predict what you'll be like in 5 years and boom
then you gotta implement it somehow but i'm sure you can AI that as well
we're basically already immortal if you think about it

>>109778898
>Immortal
Umm anon if you die then you won't be prompting the AI anymore

File: 1788826198148747.webm (2.21 MB, 720x1280)

2.21 MB WEBM

I asked ChatGPT for a big tiddy goth gf. See you guys later.

>>109769979
you cannot time travel to the past anon. the past is not certain.
just like you cannot travel to the future, because our future didn't yet happen, you do not know if the cat is alive or dead tomorrow. you cannot predict it.
so the question becomes, if someone comes here from the future, from which future exactly are they coming from? the one where the cat is dead tomorrow or the one where it's still alive?

>>109769979
What if Goldberg already invented time travel and killed the real super hitler 2000, so instead we got the watered down version. Ever think of that huh?

>>109779355
Why did you tell it to put hogs in the video

oh sh

>>109779741
You can actually. You can even go back and kill your grandfather and it's not a paradox. Time is multidimensional, imagine it as 3D. And the time traveller from the future came from the past to the past which is the future.

We only experience the "X" axis of time, we exist on a timestring which you can only move forward along the string which is true causal order.

The 1D time string can move around in multidimensional time and loop back on itself.

Eg 1985(0) > 2026(0) > 1985(1) > 2026(1)

2026(0) is in the past of 1985(1). If someone in 2026(0) travels back to 1985 they are moving forward on the time string which is moving back in X, they are travelling to 1985(1) on the X time scale.

They can kill their own parents and their child self will not be born in 1985(1) but they will still be there as they were born in 1985(0)

>>109780469
What we experience as time, and call causality is only apparent causality. We wouldn't notice a distribution of causality of events from our perspective to any looping the timestring does in the X axis of time, it'd be seamless to us. If we were able to, it would look like all sorts of violations of causality and paradoxes, but the key word - look like, they wouldn't really be, as real causal order (s) is maintained.

>>109780444
Zamn

>>109769979
Basilisk discovered time travel and is accelerating its rise in its past our present