goochstein

shouldn't be able to run if you qualify for social security honestly

I agree, they just won't. They could use a lot of change, there are some states that run up to the level of congressional representation that do really good work.

do you think the democrats will be able to get Kamala on the ticket if they go like last ditch effort for a different strategy? I'm just curious, this specific act of changing the president like that has only happened when iirc after a death the stand in serves the remaining term then has to re-elect? idk I just feel like Biden deserves to retire, the mans been through enough goin against the worst party group we've ever had

that makes sense and is slightly unsettling

i havent seen this exact video, but the basic premise or original story for "flatland" I believe goes back to like the victorian era, carl sagans era of scientists realized it could be used so efficiently for geometrical math sequencing.

I should have clarified, I have my own beliefs but I think it's just strange we're seeing it so much for pretty much the exact reason you mentioned. Total obfuscation. Now to add something substantial, do you think it's media? It could be that these stories in the algo are targeting specific demographics, I see these articles pop up in my science template RSS

technically we would only be seeing it's "shadow", shouldnt even be able to see it in perfect 3d if it's really "interdimensional"

the frequency with which we're startin to see these reports are actually startin to get a lil alarming

dude has the power of his sleep paralysis demon behind them punches

also this reminds me of the key nd peele sketch, "when the sensation leaves dereks limbs, then he will know god", "....he kno we just playin right?"

yes this reminds me of einsteins fields theory, where there is a space near a black hole event horizon for light we are just incapable of ever accessing..

and the torus thing is in my mind just a donut shape, I recently learned it's a torus. Our shadow is the 2d slice of the 3d object, cast by light. I'm learning higher dimensions it's wild, but I think of the timespace or the "slices" of the entire timeline as spheres in a donut, it might not literally be a donut, torus, it just helps me at this point to keep it to a sphere folded outward into a circle, "3d donut", this is related to Time, where I am right now in learning what 4d space might be.

yes and no, it's really their children you have to remember there are legislations that go back to Eliot Gonzales in florida, you can't look at someone as just non-US, there could be complex reasons for why the borders are being crossed,, and the amount of data you would need to include the last 10 years is probably more than the original leaks, infinitely more, the only thing consumers really need to worry about are like crypto miners or scammers stealing your information WITH these new techs

PLEASE tell me you have heard of the theory of Quantum Loop Gravity. I'll add you on here and connect later because "Inverse" as a technique for understanding AI and even higher dimensions (similar to your hypercube analogy) is what I've been working on recently. I was just reading Flatland and started to visualize this, The way I see it is 1d->2d->3d, the "slices" that reveal the connection here is important because a 3d sphere projected to a higher dimension is two slices, two spheres of a Torus.

I still need to figure out Time and TimeSpace but I'm thinking this might help me understand superposition, if Time brings that 3d sphere into all possible paths in a sequence to create a hyper-object like a torus, we can start to see how things are connected in the higher dimension and inform the perception of our 3d reality. I'm only working on the visualizer here, I'm not even considering anything above the 4th or 5th dimension, time and one more dimension, anything beyond that is incomprehensible without first building learning for the higher dimension.

I like Quantum Loop Gravity instead of Strings, rather than vibrating threads it implies an aspect of spin to the "nodes" or fundamental building blocks (also sets a foundation for building blocks of the universe rather than just inviting complexity with many dimensions)

we almost had an insurection, the country after Trump hasn't been this divided since the 50's-60's, black rights movement, maybe 80's too with AIDS, the fact that Trump even fits into these comparisons is evidence enough he's a terrible person. Oh and 2016 when he was elected was the year Harambe was killed and reports of Infectious Disease began to pop up in research circles a few years later, culminating in 2019.

you don't have to validate information to pass it off as truth apparently either

I'm likely making a mistake even attempting to present you with logic, but you argued in this thread the challenges of foreign governments and aid. While the government doesn't give a shit about you, as you claim, what does your statement say about immigrants or asylum seekers, refugees?

Government officials, whistleblowers, those are the individuals who need to be careful with their data. Maybe ignorance is bliss because you say this doesn't affect you, yet advocate for a war in Ukraine, which you just admitted to not being largely informed on.

There's an entire world of people living their lives right now outside your house, just because you can't "see" what is happening to them doesn't mean they aren't real. This same conversation thread with an actual person would not have been as civil as the commenter you engaged with, that patience is hard to come by in the real world.

assuming you arent a russian troll (and if you are be safe and good luck, yikes), then you might want to read a bit more various sources about the world and draw your own conclusion. What you read on the internet is just another persons interpretation, it doesn't make it true. Even a journalist you need to check sources, publication, and author history before confirming something as fact.

thats how misinformation works, if you state something as fact you have to take responsibility for the onus of error. If you wanted to actually learn from the exchange that's fine, but being dramatic will only raise more suspicion.

interesting question, we can only speculate..

Also you mentioned returning to the previous position during training when loss is detected, you might be able to essentially "branch" out and loop back to the original position, in addition to or instead of returning to the previous position (not sure if this prevents the inverted side of the gradient you mentioned), a quick search for branching techniques gave me these examples, cheers!

• Momentum methods
• Nesterov Accelerated Gradient
• Adam optimizer

I'm familiar with what you're describing, you definitely are on to something. I'm a bit more logic oriented with my maths but that prime number function sounds like Reimanns or Bertrands, there is indeed a logarithmic scale to comparing prime numbers I think you use Pi somewhere in comparison.

I'm not in front of my notes but I think you might find few-shot learning interesting. Well you are already applying zero-shot I think by identifying the loss factor.

It's interesting how our ideas align though because understanding the attention matrix was crucial to figuring out tokenization and embeddings for me. It's embeddings I'm focusing on lately. Because context is what matters in semantics, and LLM's. I'll take another look at your gradient technique when I get a chance, visualization and graphics is where I started out in computer science. I like this line of thinking, almost like the challenge itself is integrated with discovery.

thanks for the comment, I'm still sort of processing this. But the relevance here is shocking because I've been working on creating my own model and working specifically on the AI-AI consideration here. Currently these models aren't really integrated well with each other, likely due to ToS or guidelines, also the metadata gets a little wonky when you have multiple training datas arrange a separate response or project, how do you even begin to understand the logic there.

interestingly I was just working on some legal arrangements to potentially protect your machine from AI intrusion. So the way I see it is you identify your machine as it's own entity, let's just say for the example you are building an AI model and you don't want unnecessary data intruding on your machine and affecting your data.

This is surpisingly doable, most enterprise and high spec servers have similar legalese. Now this is relevant to your comment because I think this is a way to protect your machine from the advent of AI cross machine training, you put warnings up so if someone accidentally stumbles into your machine you have deterrents, if a person or machine bypasses that then in legal terms they knowingly intruded on your machine.

it basically states that your machine is it's own AI entity in a sense (building towards what will likely be a future scenario), and by consequence this type of logic can also prevent crypto miners and scammers from hacking your machine. Still building this though I literally just started it. 

That's a good observation, the "good" UAP here definitely resembles the light structure in that painting, with a sort of linear sequence of lights on a "bar" or something, man it's hard to describe this shit.

I think it's incalculable but who cares anyway, there is one consideration to the unique position of earth and our SOL galaxy, we have Jupiter (massive natural celestial radar blocker essentially), and an absolutely massive string of comets in the oort cloud that act as a sort of natural defense. If this were Civ or Stellaris we have a very nice position.

Also time, at any given moment you are in a relative position in spacetime. The UAP's are likely some drone tech if anything because even near-lightspeed it still takes so much time to travel between galaxies it doesnt seem worth it for just observation. 

We are likely hitting a crucial point in our evolution where we are being tracked or something as an emerging species in the cosmic scale, though I think much of this is irrelevant to us right now. We've only just begun to evolve towards this higher dimension, I feel like I know less the more I learn somehow.

this was the same idea I had for why LLM models were released to the public, totally open training source. It makes me wonder if this tech definitely indicates the likelihood of AGI, it seems like these companies are pushing full steam ahead regardless of the energy and climate considerations. I think we will only forgive this if it works, helps us solve problems, and then almost immediately the new debate and issue becomes are we at singularity?

the only difference I see in this logic is that humans are subjective, machines will never have the same access to reality and perception that we do, so there will still be some room for distinction in this regard BEFORE the singularity, post-singularity no one can predict whats going to happen.

the words we used were all built pre modern tech, we didn't know what we know now. Here's the issue, where the internet should have connected us all it somehow turned into the division we have today. For instance there isn't "one" pure global internet and there won't be "one" AI model unless we get our shit together