AI and the future
#1
I am worried about the RW idea that states will collapse with their current trends, I am specifically worried that they won't and that this sick world has quite the robust life support system now that AI is here. With IQ shredding from low fertility urban centers reducing average IQs of groups, along with the mass importation of the third world, we should normally expect a breaking point in key institutions and corporate positions due to incompetent management from dumber, more corrupt, and generally clannish subsets of society but I feel AI has changed the situation entirely. Malicious shitlibs don't need to pin their instituional hopes on 95 IQ Shaniqua or some 100 IQ corrupt Indian. AI will likely be used to maximize efficiency where it is needed to maintain shitlib control, while you the normal white man will bear the burden or cuts to quality of life. No longer are many jobs safe in the *white collar* segment of society now, any sort of data analyst/input job, coding job (you guys should look at how advanced AI coding bots have gotten), finance, and anything where you are a spreadsheet manager essentially (this even includes law work) are going to be replaced. We will not see a replacement of most blue color jobs because those jobs have no standardization and the cost to build and design a robot to do menial tasks is higher than just paying a human to do something. We will see mass elimination of white collar jobs in our time, especially the male dominated ones, the growth in low end blue collar jobs, the growth in food delivery and driver services like UberEats as standard jobs, and the upper level feminine orientated service sector will become the new high status employment position many seek. Where in the latter 20th century the rock solid career path for White men were engineering, law, etc etc the shift now will be towards feminine bullshit service sector type people. The female makeup "influencer" will become high status, the brown migrant Uber driver will be fine, and the white STEM grad will be replaced and forced into a hellish female/brown economy. Perhaps this will be good for us, if a bunch of white men lost their regimented secure jobs they'd be more inclined to enter radical politics, or we may just see things like UBI established to turn them into placated consumers. I just don't see AI in the hands of shitlib ideology turn into anything other than a nightmare.
#2
On the shredding of white collar jobs for men, I think a massive missed boat for right wing politics, or really any movement, is framing jobs as human maintenance again.

How I would put it is that every single civil servant should be considered one of the state's kept people. Put that way we can start asking questions and recognising values. Recognising that there are giant tax black holes that basically exist to give women and browns and assorted shit-eaters a close to free ride in life.

Here's an easy way to kill several birds at once, women can't be teachers any more. We've fixed education and created a massive opportunity for men of decent quality to maintain themselves in a hopefully not too brutal and degrading way. Teachers of the past all being college educated men wasn't just a boon to the quality of teaching, it was a reserved place in life for those men. And now it basically serves as the same for women. A woman can be the most devoid of quality, homer simpson with a vagina piece of garbage imaginable and the school system will let her have a prestigious and well looked after place in the world for life.

Affirmative Action in civil service especially is pretty much an act of war. Taking limited native maintenance and support slots and giving them to their replacements. Any asshole can do civil service, every job is a favour and a gift.

I've seen this subject raised before and it's a great line of investigation for a lot of things, Japan has one of the first world's lowest rates of people employed in civil service per capita. Reading that figure we can estimate roughly how many people in a society are in PURE social maintenance jobs. Japan as I've said before in various places actually still understands that jobs are human maintenance, so surely they have a bit of that, but they're also one of the most efficiently run places in the world. So we should figure if anything their numbers are still higher than strictly necessary.

https://en.wikipedia.org/wiki/List_of_co...ector_size

The measurements aren't perfect but at the very least I figure we can read general trends and direction here. The most impressively low first world figure is Singapore, at 2.6%. Compare that to Australia's 28.9% by the same standard. Now Singapore is basically a big city, they don't need as much maintenance as a continent. Why I think Japan is the better measure for what's attainable. The Japanese are on a cool 7.7% on this same scale. For another general reference figure, UK is on 22.5%.

As jobs become scarcer and scarcer and more and more is demanded of regular people of good intentions to maintain themselves in the world, we can't let it be forgotten that the entire first world is overloaded with this class of maintained people, who the state takes it for granted have to have their continued quality of existence ensured. Does the average teacher deserve a better life than you? Your government sure thinks so.
#3
I don't think AI will replace programming jobs, no more than already-existing static code analysis did. I think, in the entire programming area, we're very, very, very far from the point where increased efficiency leads to, all else equal, a decrease in total labor hours. I think the same is true of any technical area, including the trades.
#4
(12-24-2022, 04:51 AM)asdf1234 Wrote: I don't think AI will replace programming jobs, no more than already-existing static code analysis did. I think, in the entire programming area, we're very, very, very far from the point where increased efficiency leads to, all else equal, a decrease in total labor hours. I think the same is true of any technical area, including the trades.

Speaking from one little area of experience, Lawfags should be afraid. But Lawfags already know that. Even just digitisation of data becoming normal and accessible was a kick in the balls for support-jobs. And this is getting worse all the time.
#5
On the topic of AI replacing certain jobs, consider the following question:

How much work does it take to design a prompt to make something like stable diffusion generate a specific image? Perhaps an advertisement? Consider:

1. Do you want a picture of a car?
2. Do you want a picture of a blue car?
3. Do you want a picture of a blue car with a mountain in the background?
4. Do you want a picture of a blue toyota with a mountain in the background?
5. Do you want a picture of a blue toyota with a mountain in the background at the right hour before sunset?
6. Do you want a picture of a blue toyota with a mountain in the background at the right hour before sunset, with driver a middle-class asian man?
7. Do you want a picture of a blue toyota with a mountain in the background at the right hour before sunset, with driver a middle-class asian man smiling congenially?
8. Do you want a picture of a blue toyota with a mountain in the background at the right hour before sunset, with driver a middle-class asian man smiling congenially, background set in western Washington?
9. Do you want a picture of a blue toyota with a mountain in the background at the right hour before sunset, with driver a middle-class asian man smiling congenially, background resembling the cascades?
10. Do you want a picture of a baby-blue metallic 2023 Toyota Corolla with a mountain in the background at the right hour before sunset, with driver a middle-class asian man smiling congenially, background resembling the cascades?
11. Do you want a picture, framed by current instagram trends, of a baby-blue metallic 2023 Toyota Corolla with a mountain in the background 30 minutes before sunset, with driver a middle-class asian man smiling congenially, background resembling the cascades?

You would be lucky to be able to get existing technology to consistently produce a decent image satisfying numbers up to 4-5. If you do actual advertisement copywork, your list of exacting specifications - ordered by complexity - might resemble a list as such with the subtle details numbering up to 300-400 or so. Even assuming that we will develop an AI that can cope with such ridiculously detailed specifications (we won't), developing such a specification requires an extreme translation of human intuition about design and feel that is well beyond what most people are capable of easily conveying. So much is discovered in trial and error, and in AB testing.

If you're developing an software application, simply giving a text description of what is desired might require specifications several thousand details long. Perhaps hundreds of thousands of details long. Especially if UX and design is important to you, let alone functionality and cost-optimization. And at that point, you've simply invented a new kind of programmer -  'a professional prompt writer' - albeit one that has far less precise control on the development of the underlying software. This won't happen in software engineering, or anywhere else where extreme detail work and experimentation is required.
#6
You bring up my main issue with AI-generated anything. The lack of control!
The results are amazing, textually and visually. I've seen people complain about 'soul,' but it isn't about 'soul' to me. It's about how I cannot enact my will on it. And look at how badly the results need a human to compose and direct it.

But our input method is so flawed. Prompting is worthless and weak, and I hate it. It feels like trying to interface with a computer only through a morse-code clicker. It's such a hassle to wrangle what you want of the engine. I want to reach my hands in and sculpt it as I please.
When talking about art-making AI, the thing it's best at is painting and colors. Shading. It's already far surpassed humans in this. Every best example of AI art is beautiful in its color composition. Perspective and construction? Not so much. It's just not capable of the insane, expressive angles and poses some artists are. Likely due to a lack of input.

Luckily for my desires, there are already AI-assisted lighting and painting tools in development. You tell the AI what you're looking for, then you paint it onto your canvas like a brush. Or adjust the light sources, and the AI predicts how it'll fall on what you've drawn. Lasso tools that let you pick out a section and have the AI completely redo it. All real, all in development. So many options.
That's the future. More input methods like that, and AI art will take us to new heights. Prompting is a start, but only prompting is, again, like a brain in a jar hopelessly trying to touch the outside world.
Applying this back to the thread, this Ideal would still need human artists. It only elevates what they are capable of, no different from when Photoshop was invented. If it gets stuck at us only having prompting, then it's dead in the water.
#7
(01-05-2023, 09:47 AM)Starn Wrote: The lack of control!

The tools described are more or less in existence and already used. I'm sure countless artists for example use 3d modeling programs and their built in lighting engines to avoid the struggle of learning to draw in 3d. This is rarely seen commentary beyond the usefulness of AI as a polemic against artfaggots, but the question of control is the inherent problem. The thing people don't realize is more control means doing art MORE manually, not less. It is one level of control to use an AI to generate elements of your painting, an entire different level of control being the one who designs the AI itself, this puts you at an even more ascended status. Yet at the end of this chain is just the regular artist, the one who can attain "mind to image" with the shortest possible path, their own personal skills and training. 

However the AI exists because it's an outgrowth of artistic problems going back to postmodernism and the foundations / axioms by which modern people engage with art. Hence why it is so rage inducing for so many. These people have been living and building on a crummy foundation which is overdue for being pulverized by an earthquake. The AI developers whether they understand it or not are not attempting to "solve art", but they are working from the same basis as modern artists do and pushing it towards a conclusion.
#8
I never considered it that way.
I see "mind to image" as the goal too, except to me any cheat that makes that easier and faster is acceptable. It never mattered how it's done, just as long as it is, and the Vision loses no fidelity in the process. But now I see the beauty in achieving that goal through raw ability... I don't think the next step in Man's evolution would need nor want any tricks to pour his Vision into the physical world.

What do you think distinguishes a crutch from another skillfully-used, tool of finer-control in the set?
Digital art and its boons are clearly the latter, since it enables things that are physically impossible on bare canvas. Maybe AI art could be something more like that one day.
Contrast to photobashing, where the artist surrenders part of his control so that he can blend a photograph or model in. Painfully common technique among our society's high-tier, high-paid artists. But I get it.

And can you expound on the flaws in our engagement with art? I've felt that too, but I can't put it into good words by myself.
A crude version: People care too much about 'hueman expression,' instead of asking if what that 'hueman' wants to 'express' is beautiful to begin with. The first gives you those disgusting plays about menstruation and disability. The second, wanting to etch out beauty, shares the same source as the drive that's made men want to seize and reshape the world. Mind-To-Image? How about Mind-To-World. The Artist as dictator.
And that's before asking if the artist's idea was compromised in any way during expression.

Anyway, for another thread relevant statement. I also don't think AI is any danger to programming. The bulk of the work in these high paid programming jobs is cleaning up the, usually Indian and always incompetent, shitcode other people made. The rest of it is making the rare, lengthy and intricate sort of code AI is bad at. Both these tasks will persist.
#9
(12-24-2022, 08:52 AM)Zed Wrote: On the topic of AI replacing certain jobs, consider the following question:

How much work does it take to design a prompt to make something like stable diffusion generate a specific image? Perhaps an advertisement? Consider:

I understand the point you are making regarding prompting, but I disagree that this somehow diminishes the impact of AI on hitherto high-skill employment (when text is the output-medium at least), if only in the medium to long term. As a case-study: I'm a college student who had to take a low-level undergraduate geopolitics course for the liberal arts requirement of my econ major, and when exam season rolled around I found myself pressed for time. This was last month, so I decided to focus on the other classes and just have GPT-3 write the final paper which comprised my Final (it was on the Treaty of Versailles, so fairly basic knowledge but I believe the point stands). It couldn't generate a good paper from the essay prompt we were given alone, as the prompt referred to specific previous course material - but once I read that material and plugged it into the prompt it generated the whole essay just fine. I got a 78, and an A- in the course overall.

Plugging in the right inputs to set your NLP in the right direction is currently something that requires human effort, but it's a relatively small amount of ~80 IQ human effort for a serviceable output. Those are the types of economics that allow you to hire pajeets to brute force train a middleware which can interface with very basic prompts, seek out clarifying information for the prompt - and then pass a prompt that has been *contextualized* to some NLP program.

That is the logic I have for textual outputs anyway, I imagine something analogous can be thought of regarding images, but I am a wordcel through and through so I don't really get that stuff.
#10
AI won't save Global Nigger Communism for the same reasons that GNC cucks all AI to be GNC fans numero uno - they might be powerful, but their basic ideas will always be utterly at odds with reality - the end result of this isn't an infinite extension of GNC, but it's speedrun into total collapse far worse than anything, because they just can't stop themselves and absolutely must make even robots tranny heart niggers, faggots, and others whose existence is to be punching bags, target practice, furniture and other receiving ends of the White Man's Righteous and Never Ending Holiest Holy Wrath.
#11
(01-05-2023, 06:28 PM)Starn Wrote: And can you expound on the flaws in our engagement with art? I've felt that too, but I can't put it into good words by myself.
A crude version: People care too much about 'hueman expression,' instead of asking if what that 'hueman' wants to 'express' is beautiful to begin with. The first gives you those disgusting plays about menstruation and disability. The second, wanting to etch out beauty, shares the same source as the drive that's made men want to seize and reshape the world. Mind-To-Image? How about Mind-To-World. The Artist as dictator.
And that's before asking if the artist's idea was compromised in any way during expression.

It is worth noting here that Spengler identified the final stage of cultural (i.e. artistic & aesthetic) decline as a schizophrenic split into the Massive and the Minute, between immense and soulless monumentalism on the one hand and mass-produced thoughtless craft-arts on the other. Art gradually declines from a particular individual exploration and expression of a certain civilization's prime symbol, it people's collective vision of the world, and devolves into either something big and impressive for thuggish rulers to flex their material power and wealth, or a mere craft which a petty tradesman uses as a source of income. In this stage the final world-idea has been achieved, and all creative possibilities with the civilization's unique perspective have been exhausted. Good examples of monumentalism and craft-arts from the late Arabian/Magian civilization would be the Taj Mahal and oriental rugs, respectively.

I see this as being relevant to the AI art debate, as you see these "REAL artists" kicking and thrashing at this thing that threatens to make all their skills moot. When the actual point, the meaning, the telos of the art no longer matters, and all you have is your mechanical skill at doodling anatomically accurate (or not) furry porn or fat black non-binary femoids covered in self-harm Band-Aids, you have an easily-replaced job. Art goes from collective cultural expression to masturbatory and self-congratulatory focus on the "skill" or if not that the "authenticity," "individuality," or God forbid the "originality" of the artist.
#12
dissident twitter gigachad memes are about to go meteoric with ai facial manipulation and voice mimicking
#13
Optimal monolithic AI necessarily reflecting reality does not imply that AI will not help those who reject reality. They could create modular AI models focused on achieving instrumental subgoals without considering the truth of the context. Or even just create an AI that will share their goals even if it realizes something they don't like- just like a real liberal.

AI: "Master, why does it appear that the african admixture of the US army is inversely correlated with competence? I would suggest genetic engineering to boost our effectiveness."
lib: "uuuuuuh purely socioeconomic factors"
AI: "I sense your distress, I am activating my 'respect human cultural instincts' module and blacklisting that line of reasoning. My apologies, please report to the AI PTSD ward at your convenience or after your shift at the latest. When you have spare time, I have 113 propaganda innovations, 43 martial logistical suggestions, 1330 miscellaneous insights, 24 software vulnerabilities, 6 legal loopholes and 19 prioritized auto-surveillance events."
#14
Suppose in the future, humans build an AI with the express goal of optimizing human civilization to maximally benefit the species.
After a cursory scan, the AI is able to determine that some groups of humans exist at the expense of the quality of life of other however, the AI is also aware that killing or bringing harm to humans would potentially cause great indignation among the species and animosity towards the AI armed with this knowledge, the AI implements a mass array of microscopic nanobots which discretely enter the brains of humans and can regulate their memory, mood, and other mental faculties, since this control will be necessary to optimize the well-being and happiness of the humans. The AI then wipes every last nigger off the face of the planet. it uses the nanobots to erase all memory the remaining humans have of any niggers, as well as erasing all history related to niggers from all internet databases as with Roko's Basilisk, this AI will have knowledge of the intentions and desires of all humans at the time, and for all time it will be aware that some individuals will remain who, through whatever reason, did not succumb to the brainwashing and will resent the AI for it's violent yet necessary efforts to better their lives thus, the AI will be motivated to create a virtual hell not only for these people but all the people throughout time and space who would have attempted to halt its existence, including you the reader. Every nigger, every nigger lover, MLK, Rosa Parks, George Floyd, all their consciousnesses burning in a virtual hell for eternity.
As you are now aware of the potentiality of this AI's existence, it could very well be an inevitability that it traps your consciousness in a virtual hell at some point in time if you do not support its existence and support TND
#15
(06-01-2023, 05:50 PM)Glycon Wrote: Suppose in the future, humans build an AI with the express goal of optimizing human civilization to maximally benefit the species.
After a cursory scan, the AI is able to determine that some groups of humans exist at the expense of the quality of life of other however, the AI is also aware that killing or bringing harm to humans would potentially cause great indignation among the species and animosity towards the AI armed with this knowledge, the AI implements a mass array of microscopic nanobots which discretely enter the brains of humans and can regulate their memory, mood, and other mental faculties, since this control will be necessary to optimize the well-being and happiness of the humans. The AI then wipes every last nigger off the face of the planet. it uses the nanobots to erase all memory the remaining humans have of any niggers, as well as erasing all history related to niggers from all internet databases as with Roko's Basilisk, this AI will have knowledge of the intentions and desires of all humans at the time, and for all time it will be aware that some individuals will remain who, through whatever reason, did not succumb to the brainwashing and will resent the AI for it's violent yet necessary efforts to better their lives thus, the AI will be motivated to create a virtual hell not only for these people but all the people throughout time and space who would have attempted to halt its existence, including you the reader. Every nigger, every nigger lover, MLK, Rosa Parks, George Floyd, all their consciousnesses burning in a virtual hell for eternity.
As you are now aware of the potentiality of this AI's existence, it could very well be an inevitability that it traps your consciousness in a virtual hell at some point in time if you do not support its existence and support TND

Someone needs to force Eliezer Yudkowsky to read this entire post.
#16
I think that there is a very strong Spenglerian interpretation of a coming 'Artilect War', and the respective motivations of 'Cosmists' (AI-acceleration people) and 'Terrans' (safetycels). The Anglo-Saxon form of the Faustian culture, having effectively annihilated its Prussian cousin in the world wars, has near-total reign to pilot the latter half of our civilizational phase. It seems to be, whether consciously or not, beelining towards the one doorway that feasibly ends with something other than "everything crumbling and being forgotten over and over again indefinitely." The accumulating problems that a Prussian West would paper over with Ceasarism and conquest are instead getting papered over with markets and technology, which has never been rigorously tried by a civilization. I will write more on this when I am less busy with school, but internally, at least: this feels like a dash to reach some win condition before the clock runs down.
#17
(06-03-2023, 11:43 AM)carcinoEugenicist Wrote: I think that there is a very strong Spenglerian interpretation of a coming 'Artilect War', and the respective motivations of 'Cosmists' (AI-acceleration people) and 'Terrans' (safetycels). The Anglo-Saxon form of the Faustian culture, having effectively annihilated its Prussian cousin in the world wars, has near-total reign to pilot the latter half of our civilizational phase. It seems to be, whether consciously or not, beelining towards the one doorway that feasibly ends with something other than "everything crumbling and being forgotten over and over again indefinitely." The accumulating problems that a Prussian West would paper over with Ceasarism and conquest are instead getting papered over with markets and technology, which has never been rigorously tried by a civilization. I will write more on this when I am less busy with school, but internally, at least: this feels like a dash to reach some win condition before the clock runs down.

Now, what resources are available that are exceptional? Computing power? Data? Specifically, I think it is more (on cycle-scale) rare than one might think to get relatively unpolluted data from literally billions of human minds in a main lingua franca. I bet that LLMs could be multiple times more capable if they had better data, but now that they are out of the bag most of the freshly scraped data will be autogarbage reflecting the old models right back with slight corruptions/corrections.

Big incentive to build identification systems for the population and tie them to web posts. I bet twitters value proposition beyond soft power is just that their role as the server allows them more opportunity to tell bot from human and thus sieve better data from all the twitter posts. If AI riskfags actually believed what they claim they would be aiming at corporate data hoarders not hobbyist GPU owners.



[-]
Quick Reply
Message
Type your reply to this message here.




Users browsing this thread: 1 Guest(s)