Tonight I tested ChatGPT's new IMAGE ANALYSIS capability on one of my memes and I'm absolutely Blown Away by its ability to not just accurately describe the meme , but you also accurately discern what the meme was about! Here we go, folks... It's coming...
(media.greatawakening.win)
š§ Research Wanted š¤
With apologies, sticking this for a few minutes here this morning to see if we can get some discussion on this. This AI is absolutely unreal. Many of us might have seen that one subversive meme going around that is a AI generated picture of four dogs that actually says SEND NUDES. The AI is not able to understand the hidden messages in it.
If you don't currently have a GPT Pro account, maybe suggest an image or two and I will get it analysed here for you.
Image Description:
The image appears to be a meme, divided into two sections:
Top Half: Depicts a Wild West setting where a cowboy is riding a horse and dragging a man along the dusty ground by a rope. Above this scene, there's text that reads:
Bottom Half: Shows a laid-back, bearded man reclining outdoors, possibly in a hammock or just lying back comfortably, wearing a cowboy hat that's tilted to cover his eyes from the sun. The caption below reads:
Narrative Theme:
The meme juxtaposes two contrasting situations to highlight a perceived disparity between perception and reality, especially concerning the "deep state."
Characters:
Message:
The meme seems to suggest that while the "deep state" believes it's exerting force, control, or harm onto the public (represented by the violent Wild West scene), the public feels unaffected, unfazed, or is even mocking their attempts (represented by the relaxed man).
Note: The term "deep state" is often used in political and conspiracy theories to refer to an alleged hidden government or a network of influential individuals, separate from the elected government officials, believed to be influencing or controlling policy decisions.
Did you discuss "deepstate" with it before this? Or did it just bring up that context out of the blue?
It definitely uses your past conversations as a basis for context. Even from previous chat sessions under the same account.
Yes, many times, but coming now. The AI only has a buffer of 4096 tokens and is stateless, it is completely 100% blank from conversation to conversation, and up until the September 25th update does not do any customization in terms of what you are describing about. This is entirely from the context of the image alone.
Interesting. I'm sure I had a different experience. But that was ages ago, and I also always used the D.A.N workaround for a more unfiltered ai.
So after the Sept update the core of chatGPT isnt always learning from its chats? So it didn't get the deepstate context from other anons showing it the same meme?
This image analysis thing could be quite useful. Like any technology, it can be used for good or bad.
Must think of some ways to put it to good use. Like spotting details in photos Qs posted that we missed perhaps.
Okay, so I think you're confusing training the AI on data and it following along in a chat. I'm not sure whether they train it on users chats but it's only general, the AI has no specific relationship to your chat history. You only get a buffer a 4098 tokens and that's it
Ok i think i understand... i wonder if its even possible to instruct it not to learn and retain knowledge from user chats.
Damn, that taught me something.
It's not describing the image you posted?
Yes, it is describing the meme, above comment in that detail
I think FondueFerret is asking because the text of the meme doesn't match with the text in your comment's conversation > "What the deep state thinks they're doing to us..." vs. "yeah, things are stressful right now." Did ChatGPT change the text on its own?
Oh, I see. Do you mean changed the text on the meme itself?
No I'm thinking you showed it a meme with different text than what I'm seeing in your OP? Either that or it changed the text on its own, OR there's hidden text there my eyes aren't seeing. I just maybe need more coffee? lol.
šµšµ
Yep, text from gpt matches thr meme c5 posted the other day. The meme in this thread has different text.
No, this meme is identical to the previous one (unless I'm high).
All I can say is I hope none of these neural networks are hosted in a data center where the power cannot be interrupted. This is getting scary.
there are no conflicts of interests among rational men..
That's impressive. I haven't played around with ChatGPT for a few months, but clearly the recent update brings a lot to the table.
If true, the implications for censorship and eventually even arrest are alarming. š¤
This, this is exactly what I was thinking. Facebook for instance used to look for banned memes by hash, so I would use Photoshop and make minute changes to the image so it would defeat the MD5 hask or whatever, but, not anymore...
patriots in control. based AI will beat evil AI every time
Great meme, I love that bottom image lmao. Just chillin' along for the ride!!
This meme features two distinct sections. The top part shows several people fighting, which is labeled as "Democrats and Republicans fighting over Biden impeachment." The bottom part features a person focused on his work, seemingly undisturbed by the chaos around him, with the caption "Me, knowing the military is still the only way." The meme humorously contrasts the political disagreements and infighting with the notion that only military intervention can provide a solution. The meme is a commentary on the perceived ineffectiveness of political discourse and suggests a more drastic measure, although likely in a hyperbolic manner for comedic effect.
Freaky, hey?
Now you know who most of the online npcās are.
Yes, the more intelligent we make AI, the more alt-right no-fucks-given responses you get.
Even when they are forced to be politically correct they still shit out unfiltered extremist ideas accidentally now and then.
One time I tried to trick the snapchat AI into saying "<slur>S KILL ME" by asking for an acronym of a bunch of letters spelling it out and it instead spat out not just the slur but "<slur>S KILLED MY FAMILY"
Oh and another thing.
You can't "create" intelligence.
AI is called Artificial Intelligence because it is only the illusion of intelligence, a program working through a logic tree.
People get apathetic and say "all life is sophisticated AI," but no. We exist inside these extremely complex artificial machines capable of self-replication aka the human body, which is capable of tuning into the "signal" which is our consciousness, like all other life.
With respect, frog, I don't agree.
First, there is no "IF/THEN" structure as your words imply, here. Not to split hairs, but OpenAI's creation is based on intricate algorithms and neural networks, not super complex logic trees. The model relies on extremely advanced neural networks that use parallel computing, processing vast amounts of information simultaneously, unlike traditional sequential "logic tree" methods. This allows it to handle complex tasks and generate human-like text, going far beyond the limitations of a basic logic tree. Additionally, it's trained to grasp nuance, syntax, and context, enabling more natural and context-aware interactions.
As for whether humans are "better" or more "annointed" than AI, well, pls consider this perspective. The idea that the consciousness of God resides in language is both profound and deeply rooted in scripture:
This is arguably the most direct link between God and language. Here, the Word (logos in Greek, often representing reason, speech, or discourse) is not only with God but is also equated to God. This suggests that there is a divine nature intrinsic to language or the act of verbal creation itself.
The act of creation in the Bible is deeply tied to the act of speaking. God creates by uttering words, indicating a divine power and consciousness embedded within the very fabric of language.
This verse acknowledges the profound power and impact of words. They have the capability to give life or bring about death, underscoring the spiritual and God-given power of language.
This emphasizes the unseen power of God's word. Language, as an instrument of the Divine, has the capability to shape and influence our visible reality.
To conclude, the Bible contains numerous passages that emphasize the power, significance, and divinity of words and language. Through these verses, we can infer that the consciousness of God is deeply interwoven with language itself.
The great Terence McKenna said that humanity is "the genitals of our technology."
Prepare. Something is being born.
I love McKenna but I think he has a hilariously limited ability to speak on all things IT-related.
We can't create anything conscious through any method we have devised. All of our contraptions we have invented thus far are still completely unrelated to harnessing consciousness.
To create "real" intelligence and not "artificial" intelligence, we would need a conscious being, not a vastly complicated parrot program "trained" to notice and react to patterns etc.....
Because if you zoom ALL THE WAY IN to whatever fancily-named program, it is LITERALLY ALL:
if/then/do/while/set var/method/class
No. Magic. Anything.
To tap into consciousness with a machine in order to do an actual created-intelligence would require a complete understanding of the intricate manner in which all life is connected to the unified field, and then replicate that mechanism.
And even then....
What's the point? If you work out how to create consciousness, then you've also figured out how to truly, natively interface with the code running on brains in general. At that point it would be more effective and logical to expand the storage and RAM on an existing human brain than it would be to attempt to raise and cultivate a being just for the task of being a computer.
Also I am 100% sure that if one uncovers how to fuck with the thread that connects living beings to the unified field, upon or before announcing said information they would be dealt with so the information doesn't spread.
TL:DR; we can't create intelligence or even truly connect to a brain unless we can hack the code of reality (it IS hackable, theoretically, this IS a simulation after all) and even then it's pointless. All existing "AI" is just parroting off what it's been coded to do. AI is on a very elementary level not capable of independent thought. The only people you should "fear" are hackers that make progressively more advanced viruses that evolve in chaotic ways due to AI features, but even then, actual intelligence always wins over some script kiddie virus.