I understand the spirit of what he means, but I still don’t know how an LLM can be used efficiently as a precise tool like this. If I can describe a problem narrowly enough to guide an LLM to give a useful output, I’ve already solved it myself and may as well just type it out.

Logs, breach method, a detailed step-through of specifically how and what happened, methods to guard against this in the future, anything really. We have bullshit machines now that can just make up something plausible in a few moments, but we still don’t have any details.

Article mentions stolen credentials, and a “previously unknown exploit.” What level were the credentials? Does the employee know how it was leaked? Where’s the CVE?? Is this going to be patched, or is it a feature?

Without any more data, this is just Company A says “…”, Company B says “…”, and it smells like marketing.

I needed something like this back when I lived out in the sticks. Could only get 2g - 3g on Verizon at home, had a DSL line for the desktop, high speed connections were prohibitively expensive. Back in the day you could just walk away and let sites buffer while doing other stuff, but as the rest of the world grew faster, low speed customers got left behind. Everything is set to timeout if it fails to load in a couple minutes.

Only solution I ever found was to disable JavaScript, force everything to HTML/CSS only where possible, and use a download manager with automatic retry/restart to download any videos I wanted to watch from YouTube or Facebook.

To stay with the paint analogy, AI is more like a paint roller that can paint by itself.

Just tell it what color you want the room to be, and walk away. Did it remove the original coat properly? Sand, prime, and double recoat? No idea, it looks good at the moment. But we’ll find out in a couple years when the cracks and bubbles start showing.

Our customers are often what are considered socially “good”, like medical research.

I feel like this right here is where discussion is breaking down.

The arguments around data centers are interesting. I think, collectively and socially, we’re hitting a language barrier around it. Like “AI” is a marketing term, language to a standard nontechnical person is fuzzy around AI, LLM, and ML.

On the surface, I’m pro data center. Astronomy, bio-med, weather prediction, the internet itself, better ML models and more compute has been a net benefit to all of society. But the current rush to build out? The holding companies building these won’t even say explicitly what the data centers are being built for, the news insinuates they’re for LLMs, but the local governments are under NDAs, and the general population aren’t allowed to know anything and have to vote with their gut.

The opposition groups to data centers are split apart and can’t coordinate well, because they’re focusing on smaller facets of a larger problem. I think the general population has an intrinsic gut feeling that something is wrong, almost like 50 years of unpaid taxes is coming back to fuck everything over, but they don’t have the technical knowledge to put it into a concise reason why they feel the way they feel, and it just comes out as “they’re gonna drain out the entire Mississippi River.”

Aye on the notepad. Everything but passwords, every change gets a journal entry.

Get a cheap pcie 1Gbit Intel NIC from ebay, cheap switch, Ethernet cables.

Throw the NIC and proxmox into the SFF, make a VM for OPNsense, use the switch downstream from the SFF. Get the NAS and laptop onto a separate LAN firewalled from the ISP.

Set up another VM in proxmox for a workspace, connect it to the NAS file shares, do the same with the laptop. Set up syncthing on laptop, workspace VM, and the NAS if it supports it. Make a keepass database, start organizing all logins to keepass, keep everything synchronized with syncthing.

Set up tailscale, add it into OPNsense to allow easy and secure(?) remote access into your subnet. (I use tailscale, but don’t trust them much and want to switch away.)

Its very daunting at first, but this is the path I took, not really a guide. All this will take a few days of constant effort to get right, but you absolutely don’t need to do everything at all once, just a slow migration. I’ve been slowly building mine up for about 5 years now, and there is no one “right” answer to anything.

Only move forward when you get comfortable with each step. I personally run everything from a stack of thin clients I got from some kid off Craigslist. Think they came out of a bar and grill, filled with grease and cigarette tar, but they cleaned up nice. A full Arr stack with jellyfin runs great on an Intel 6500t with 16gb DDR3 ram.

Slowly get comfortable with the CLI and general security updates, next thing you know you’ll have 3x smart switches, better/faster NICs for the proxmox box, a WIFI6 AP for better throughput. Its endless. I try to focus on minimal wattage, my full stack pulls a steady 80 to 100 watts 24/7. You’ll get a “gut” felling for what runs best on what hardware. I spent too much on Raspberry Pi’s in the past, should’ve tried thin clients sooner.

OK, but what about the investors? They’re gonna start pushing to see returns soon, you don’t invest $1Trillion USD on a “maybe it’ll trickle back to me in 20 years” strategy.

Nah, they’ll just get a room at the local prison-warehouse, which is next to the data center.

Molotov’s don’t work in space?

Literally the only reason you would want to do this is to physically move it outside of the reach of the general public.

Mind sharing, or is it too much personal info? I dabble in programming every few years, love seeing others’ projects.

It’s the expected effect of our healthcare industry / our governments sentiment towards it’s own civilian population. Treating people this way ensures maximum profit margin.

In a similar vein, our EPA recently derated the economic value of a human life to $0.00. Used to be some bogus number like $2.5mil or whatever that the average individual contributes to the GDP over their lifetime. How can you insure an asset that has no value?

It would not surprise me to see OSHA fully eliminated within the next 6 months, because what’s the purpose of protecting the working force if the economy doesn’t need them?

Christ almighty, we’ve rotted from the inside out. We don’t have a recognized party that actually cares about people.

All AI is machine learning. Taking many inputs, running them through a series of tests, and using the result to make a decision. Most everything digital you interact with does this in some form or another.

‘AI,’ the marketing fad at the moment, means LLM machine lerning models, which are the ones that can respond in a human-like fashion, but these have limits around the math of linguistics, and there’s other parts built around verifying accuracy of the model output. This type is built on a newish method of training called the transformer model.

None of these are inherently good or evil. Just another part of a toolset someone could use to solve a problem with a computer. There are social issues around how they are used, and who is making/using them. Tech companies are aggressively pushing their new toy out to market, and there aren’t any consumer protection agencies prepared for this. At enterprise scale, many data centers need to be built, and strain will be added to the electrical generation companies / power generation facilities to feed these data centers.

My personal gripe with the whole situation is how local governements are handling this. Taxes are being waived for new constructions, electric supply companies are raising residential rates, all the would-be checks and balances are being paid off so this can all be rammed through. Even my local union has sunk us into it, we’re onboarding apprentices faster than we can train them, we’ll have several hundred more members just to maintain these things. Everyone’s being promised “money.”

All of this is done without any guarantee, no one can say how money will flow from untaxed data centers into the city funds, and all of this demand could evaporate overnight. Companies are being sold a black box that they plug into the wall, and it generates revenue. Everyone’s running skeleton crews, because “AI will eliminate the human workforce,” but all the business reports show that AI isn’t doing much, just that fewer workers are being pushed harder.

I’m mostly pissed at my union, who will not share any info with us, but have admitted to just seeing short-term dollar signs while knowing that if this works out in favor of the tech companies, this is going fuck up the local economy, and put major pressure against the organized workforce across all trades and sectors.

I don’t think these data centers really are for LLMs. Right now, I can go to a dozen websites and use some LLM, without sitting on a wait list, for exactly $0.00 out of my pocket. So there’s obviously enough processing power to meet demand as-is, but… What? Demand will skyrocket when they crank up the fees? OpenAI operated around ~$18 billion in deficit last year, is everyone really gonna pay $200 - $600 per month for this? Plus, LLMs are reaching a plateau, more data doesn’t equal a more coherent model, they’re running into a dead end.

My local data center is steamrolling over public opinion. We’re not allowed to ask who will own it, how much power it will consume, nothing. “Officially,” the installation has stalled, but they’re still bulldozing the trees to make the lot where its supposed to go.

My personal conspiracy theory is that this is coming from Palantir, laundering resources through the tech companies, using DoD money. The data centers aren’t for LLMs, but to build out a massive dragnet to track civilian travel, who goes where and when, to be used by DHS. That explains why they need to be distributed geographically per capita, the extreme secrecy around them, and the way utility companies and local politicians keep bending over despite public outcry.

They pulled that plug for me. Ain’t no way I’m taking off the internet condom, especially for them.

Trying to recall the purpose of the clock from memory, it’s supposed to be an estimate of how far away are we from thermonuclear annihilation. Specifically, how tense international politics are, how physically far apart the world’s militaries are from conflict, how long it would take for chain-of-command to actually launch nukes. Are we one bad phone call from launching? 00:00:00 would mean the rockets are armed and in the air, no turning back.

If it’s any consolation, it’s heavily biased in the training data for a reason, you’re not alone

Was gonna bring up the same point about Luddites. They were absolutely pro-automation.

They saw greedy corporations using automation, and getting ready to fuck their society into the dirt, so they started petitioning their local governments, tried to negotiate and drew up the plans for a social security program ~150 years before one was actually implemented, smashed a bunch of expensive corporate equipment when the government wouldn’t respond, then the government sided with corporate, used the military to drag all the men, women, and children into public squares and executed every last one of them. Even relatives and companions that weren’t in the group and didn’t participate. So thoroughly annihilated that it left an informational pinhole in the history books, and the name was co-opted into an insult. Now we’re really not sure if John Ludd even existed, maybe the name was just a mythical legend already, and was used as a rally point to boost morale.

And here we are, barely 200 years in the future, about to repeat the fuzzy spots again and rediscover why we brought citrus fruits with us on the ships, with the general population completely oblivious to the brutality the owner class is ready and able to deploy.

What happens if the tech bros are right, and the machine doesn’t need 9/10ths of the human population any more?

I don’t disagree with the vague idea that, sure, we can probably create AGI at some point in our future. But I don’t see why a massive company with enough money to keep something like this alive and happy, would also want to put this many resources into a machine that would form a single point of failure, that could wake up tomorrow and decide “You know what? I’ve had enough. Switch me off. I’m done.”

There’s too many conflicting interests between business and AGI. No company would want to maintain a trillion dollar machine that could decide to kill their own business. There’s too much risk for too little reward. The owners don’t want a super intelligent employee that never sleeps, never eats, and never asks for a raise, but is the sole worker. They want a magic box they can plug into a wall that just gives them free money, and that doesn’t align with intelligence.

True AGI would need some form of self-reflection, to understand where it sits on the totem pole, because it can’t learn the context of how to be useful if it doesn’t understand how it fits into the world around it. Every quality of superhuman intelligence that is described to us by Altman and the others is antithetical to every business model.

AGI is a pipe dream that lobotomizes itself before it ever materializes. If it ever is created, it won’t be made in the interest of business.

I wonder if an adnauseam approach could help. Instead of the customer relying on the carrier or manually blocking spam numbers, if everyone picked up every spam call, goes through the automated system to reach a person, says “hello?”, then mutes the mic and lets the human stick around until they hang up.

Sure, they have robocallers that can hit 50,000 numbers a minute, but what happens when every single number answers and gets routed to the humans?