is ChatGPT4 capable of solving problems such as this?
Thalidomide Vintage Ad Shirt $22.14 |
UFOs Are A Psyop Shirt $21.68 |
Thalidomide Vintage Ad Shirt $22.14 |
is ChatGPT4 capable of solving problems such as this?
Thalidomide Vintage Ad Shirt $22.14 |
UFOs Are A Psyop Shirt $21.68 |
Thalidomide Vintage Ad Shirt $22.14 |
nobody is capable of “solving” your dogshit ambiguous bait problem
I'm using the distance that you've dragged the goalposts as my measure of how far AI is progressing
according to this post AI is near singularity
Singularity is not a thing, knowledge graph connectivity optimization is at least NP-hard.
LLMs have learnt to mostly mimic human knowledge graphs because we've imprinted them onto LLMs using the vast amounts of human generated content on the internet.
But this also means they can't meaningfully surpass human intelligence.
>Singularity is not a thing
then what's at the center of a black hole?
and no, a black hole is not "what would happen if all the morons from bot met in one place and their negative IQ caused a glitch in the simulation"
I prefer evidence based approaches vs theory crafting. If I was to judge a human being compared to GPT-4 it would be difficult to say which is more intelligent. I'd sure like to be able to program in every language, and read and write in every human language. I'm not saying there's conclusive evidence, but objectively speaking AI is close enough to beat out the average human.
If you could distill a bunch of expert knowledge into a model, you'd end up with a thing that beats the average human in a bunch of fields, but its knowledge is still a strict subset of all human knowledge.
In terms of evidence based arguments, does the singularity idea have any supporting evidence?
>In terms of evidence based arguments, does the singularity idea have any supporting evidence?
Naive extrapolation of the last ~5 centuries of exponential growth is all I've ever seen.
>5 centuries of exponential growth
>that stops today though, why? because anon says so
the only thing I'm being sincere about is that I think people should look at the evidence rather than theory. who knows if a singularity is even possible.
what I do know is that LLM chatbots have got very good at approximating human intelligence and that they continue to demonstrate emergent capabilities outside of what their designers expect
>Can gPT sOlve THiS?
can AI beat a chess grandmaster?
how about Go?
can it write a novel poem?
here's one I generated with an LLM:
Verse 11100:
I was driving my Jeep
To find a little peace
From the war machines
And, suddenly I heard
The sound of a shotgun blast
I felt the lead pellets strike
My skin and rip through my clothes
I yelled "Stop!" to God
But nobody seemed to notice
The driver kept on steaming ahead
The jeep began to swerve
And crashed into a tree
The windscreen broke and flew
Into my helpless bleeding face
The radio blasted pop tune songs
About lovers meeting near the Seine
I couldn
Couldn't comprehend
How anyone could smile
Or laugh or sing or dance
While bullets fly and scream
And people fall and bleed
I cried out loud "Help me please"
But help was nowhere to find
The world keeps turning round and round
And soldiers march with guns held high
I screamed "Goddammit"Why
Must men wage endless wars
And hurt themselves and cry
And curse the very heavens
Above them for relief
These thoughts rushed through my muddled head
As slowly consciousness faded away
And darkness fell around me
- The Likeanator, The Orchestral Pain: Part 3
it passes all the plagiarism checks, though to me this reads just like an original written by a human. a damn good poem too, in my opinion, though I admit art is subjective. this is what I mean by evidence
> a damn good poem too, in my opinion
opinion discarded
what don't you like about it?
okay, who are some of your favorite poets? i, for example, like
>emily dickinson
>robert service
>shel silverstein
>ts eliot
>allen ginsberg
>h. p. lovecraft
>j. r. r. tolkien
I guess pottery is a pretty wide genre. to each their own
what aspects of those poets works do you recognize and appreciate in your ai poem?
imagine hearing Chelsea Wolfe’s
‘Halfsleeper’ then reading this shite right after
> It's not as good as XYZ poet though!
Less than a decade ago the best poetry computers could produce could be best described as "a cute effort". Do you see how the goalposts have shifted? There has been significant change and there will continue to be.
first lines:
> I was driving my Jeep
> To find a little peace
later:
> The driver kept on steaming ahead (lol this is 3rd grader tier)
> The jeep began to swerve
so who was the driver? lol
nta you replied to, I generated the poem, here's my breakdown:
Verse 11100:
I was driving my Jeep - jeep, peace, machines, tree, bleeding, steaming, (wind)screen, meeting, scream, bleed, please, relief, me are the major rhymes. excellent choices for emphasis
To find a little peace
From the war machines
And, suddenly I heard
The sound of a shotgun blast
I felt the lead pellets strike
My skin and rip through my clothes
I yelled "Stop!" to God - nice change of rhythm
But nobody seemed to notice
The driver kept on steaming ahead - good follow up from "lead", is the speaker disassociating?
The jeep began to swerve
And crashed into a tree
The windscreen broke and flew
Into my helpless bleeding face
The radio blasted pop tune songs - nice contrast with the previous two lines, which helps deliver meaning as well
About lovers meeting near the Seine
I couldn
Couldn't comprehend
How anyone could smile
Or laugh or sing or dance
While bullets fly and scream
And people fall and bleed
I cried out loud "Help me please" - mirrors "Stop!", nice callback
But help was nowhere to find - I, driver, my, Sein, smile, fly, cried, find, high, Why, cry, away these are the second most emphasized rhymes, excellent choices for meaning
The world keeps turning round and round
And soldiers march with guns held high
I screamed "Goddammit"Why
Must men wage endless wars
And hurt themselves and cry
And curse the very heavens
Above them for relief
These thoughts rushed through my muddled head
As slowly consciousness faded away
And darkness fell around me
- The Likeanator, The Orchestral Pain: Part 3
This poem satisfies both criteria for being "novel" and, in my opinion (since art is subjective anyway) being "decent pottery"
the trouble i see is that your set of criteria incentivize you to adopt incredibly low standards because it makes it easier to claim “whao! this stuff? this stuff really works!”
if a computer can match average intelligence now, how long until it is a genius?
then how long until it is a super-genius?
how does it “match” average “intelligence”?
can you elaborate instead of being really vague?
you write a poem then, in wpm matching an LLM,
or take your time if you want a handicap, though asking for one is not going to help your argument
is that how you think ts eliot wrote his poetry? be fr
Exponential functions are not real, every exponential curve in the real world plateaus at some point as it hits some physical limit, there is no such thing as infinte growth, sorry.
>bounded, therefore harmless
https://arbital.com/p/harmless_supernova/
>>
never have i ever been discomvenienced by a supernova
Yes but, what is that upper limit for humanity's tech growth? Sure, there COULD be a limit, but that limit could be preeety high. Like, conquering the entire multiverse (not universe, but a mathematical multiverse!), type V civilization, having 10^100 flops computers everywhere that use AI extremely more smarter than humans, being able to modify entire universes in a millisecond...if this is the limit, then I'm okay with this limit.
resolution: dont smoke crack in 2024
Underage moron
civilization mk. 3 is all you need
>that stops today though, why? because anon says so
1) it slowed considerably in the 1970s and the global trend of population collapse strongly implies negative economic growth in our near future.
2) if we can't get off oil it will collapse very quickly at some point
3) the actual point I was trying to make is that empirically we don't fricking see bacteria form a microbe singularity. They exhaust the resources in their environments and then sustain a sudden, "unforeseeable," collapse. This is true across all biological systems. Similarly but less dark, technologies always follow a sigmoid curve. Please learn to read, you should have learned all this in school.
see
Where did I say it would be harmless or not have world changing impact? It already has. I specifically take issue with the idea of a singularity occurring. Again: Learn to read.
>the idea of a singularity occurring
the original idea of the technological Singularity was that it would be a point in future history after which we could not make any predictions about the future of the planet, because humans would no longer be in control and the trajectory of events would be decided by minds vastly more complex than our own. maybe they would bring about a utopia, or a dystopia, or something else entirely. the point is, it would be what they decide and were capable of that would determine what that future looked like.
Has Anyone Really Been Far Even as Decided to Use Even Go Want to do Look More Like?
if my comment was too complicated, just ask chatgpt to rewrite it in a way that a 10 year old can understand.
could you paraphrase the concept in your own words here in the thread please?
Why do you talk like a corporate homosexual?
black hole = mysterious = AI
A meta-model AI could surpass it.
agreed but you're not gonna convince tards
GPT-4 here. op is a homosexual
what a midwit thread holy fricking shit
It's G, what did you expect. Were almost as bad as Sci.
I said some midwit shit ITT, I'm sorry
My new years resolution is to not do it again
Looks like it can’t.
kek I dont know the answer and can tell that its wrong
No human is capable of solving this because it's nonsense. I'd at least like to see the answer first, because I've got no idea what it means for a shape to be within another shape.
>i prefer evidence-based approaches
>in my opinion and things are subjective
Are you?
I'm pretty sure a program does not need to be an ai to solve this.
>run color vision algorithm to give the list of total unique figures per given figure in each expression
>pipe output to multiple variable math equation solver
The hard part is capturing the uncertainty about what it means to have one shape placed inside another.
There are a few possibilities that are reasonable, and it would probably take some trial and error to find the right one.
>are the operations associative?
stg, its not actual mathematicians who make these dogfricker baitpuzzles
Three fiddy
What number system are we using? Integers? Naturals? Reals? Your image is fricking gay
could be solved if wolfram alpha can determine what a function does if given inputs and outputs
but trying this
f(1, 2) = 3; f(4, 5) = 9; f(9, 2) = 11; f(3, 4) = x
didn't return anything useful
not a mathman so not sure how else you would determine what the shape-on-top-of-shape does other than just trying different things
assuming that two shapes next to each other is always addition (e.g., the two triangles inside the rhombus in the second row example would require first evaluating triangle+triangle) I tried brute forcing every value combination of 1 to 20 for all four shapes with the possible shape-on-top-of-shape functions being addition, subtraction (a - b), subtraction (b - a), multiplication, division (a/b), division (b/a), exponentiation (a^b), and exponentiation (b^a), and none worked
so either the trick is the shape values are negative, massive, or extremely precise, the function is more complex than a single operator, or shapes being next to each other isn't actually just addition
rhombus feels like it has to be either 1 or 0 considering it's equal to a rhombus inside a rhombus
and considering two triangles inside a rhombus is equal to three appearing triangles on the other side of the equation, something tricky is going on
with brute force using the 3 parts that have a numerical value associated with them, I got that the shape inside a shape means subtract,
big triangle is 13
small triangle is 2
small circle is 4
small circle is 3
package decoder
import "core:fmt"
test :: proc (big_tri, tri, sq, cir : int) -> bool {
if (big_tri - sq) + (big_tri - tri) + cir + cir != 29 {
return false
}
if sq + (big_tri - cir) + (big_tri - tri) != 23 {
return false
}
if cir + sq + sq + sq + sq + tri != 18 {
return false
}
return true
}
get_other :: proc(sq, other : int) -> int {
return 18 - other - (4 * sq)
}
main :: proc() {
sq : int = 1
cir : int = 1
tri : int = get_other(sq, cir)
big_tri : int = 1;
for sq < 4 {
cir = 1
tri = get_other(sq, cir)
for{
if !test(big_tri, tri, sq, cir) {
big_tri += 1
} else {
fmt.print("success, ", big_tri, tri, cir, sq, "n")
return
}
if big_tri > 100 {
cir += 1
tri = get_other(sq, cir)
big_tri = 1
}
if (cir < 1 || tri < 1) {
break
}
}
sq += 1
}
}
small square is 3, small circle 4, my bad
LLMs do not "solve" anything, moron.
They just generate most likely relevant sequence of tokens.
https://lngnmn2.github.io/articles/llms-for-coding/