I don't think theses LLMs are like any previous machine. You use a metal lathe or a cotton gin, you get reproducible results; near perfect results if you maintain the machine correctly and it doesn't have mechanical failure. LLMs for coding are not good deterministic machines. Tiny variations to a prompt or slight word ordering can dramatically change the output. Most models silently write away errors or do blank exception checking, making something seem like it works. It might be getting 50,000 records, but you're not seeing the 2,000 edge cases it's silently dropping. No one knows because no one reviews 10,000 lines of slop.

These machines are fundamentally worse than previous industrial revolutions. College students are booing people at graduations going on about AI because they've used it and the good students know just how wrong it gets things, sometimes 1/8 to 1/4 of the time! One day a bridge will collapse because a human check will be replaced by some shit engineering model and it will totally pass a design that has incorrect tensile strength. Than 4 ~ 10 people die and maybe, just maybe, we'll realize how dangerous these machines are ... NOT because they're intelligent game changes, but because they are next word predictive guessers with no intent and no real intelligence.

The machines won't kill us. Our trust in them will.

The entire economy is currently in a crashed state. Money printing went out of control, home compute is becoming unaffordable, no one wants any of these AI data centers, none of your OpenAI/Anthropic subscriptions even remotely cover the runtime cost of compute, much less the sunk cost of the infrastructure. It's all been run into the ground and we're all going through the motions like everything is fine. There was a term coined for continuing on in a collapse during the Soviet era: Hypernormalization.

We're distracted by irrelevant social issues, or people screaming about Orange Man (when really every single president in our lifetimes has been pushing towards this, unless you were alive during JFK, and we all know what really happened to him).

The data centers will get built, not by capitol raised through supply and demand, but because the governments of the world need the surveillance infrastructure to impose authoritarian, technocratic rule. That's the reason for "child safety" laws that are really digital identification laws designed to remove any remaining privacy anyone has.

Universal basic income only makes sense if there is zero resource scarcity. We don't have the dilithium crystals and replicators of Gene Rodenberry's universe. The Weighted Random Word/Code Generators are nowhere close to the level of Commander Data, nor are they "intelligent." They cannot reason, they just do a very fancy and compute intense predictive output. It's impressive, but transformers and attention blocks will never be an "AGI."

The technocratic class wants programmable money so you're only allowed to spend FOOD-COIN on x products and HOUSE-COIN on housing. Extreme control that can cut off anyone from their money should they say or think the wrong thing.

Universal Basic Income creates a permanent class tied to government gibs. It will be nothing like Star Trek and a whole lot more like The Expanse.

90%+ of "UFOs" are likely covert air force operations. The documentary Mirage Men interviews a CIA operative who intentionally drove a former Lt Col turned military contractor to a mental institution, in order to continue a narrative to cover up military operations.

Do meds make you stop thinking or analyzing things critically? Most of the psycho meds make people dumb as shit today. Here I have a whole writeup:

https://battlepenguin.com/poli...

There are sources for videos, analysis of the "bullet photos" and lots of other details. My arguments are sound. They might not be something you agree with because you believe everything you see on the Google and/or TV without question, but they are not an indication I need to "take my meds." Or have you forgotten how much the Epstein files have shown just how many "conspiracies" are true?

Operation Ajax. Operation Paperclip. Operation Mockingbird. COINTELPRO. All these are real, actual conspiracies against innocent people that took place under various presidents.

Keep taking your meds and go numb.

People don't take long haul trips in Europe. If they do, they'll often take a train and rent a normal car at the destination if they need one. America, Australia, etc. are really spread out. Every person I know who tried to do a long haul trip (like to the beach) in an EV said "never again." Especially in high vacation season when charging stations can have queues of 10~15 vehicles. If you wanted to have America on EVs for all trips, you would need charging stations the size of a Buc-ees everywhere, and they'd be mostly empty except for Thanksgiving, Memorial Day, etc.

Also, if you are dumb enough to do a long haul and find yourself in a 10 car queue for the charger .. go to an RV campground and rent a space instead. Keep some high voltage adapters and plug into the RV electric ports to charge.

EVs only work if you live in a city and never leave it. If you're in NYC, Chicago, you also need a garage (probably a space heater too) because you're not going to be able to charge those things at most charging stations when it's -10F outside. Most Americans can only have EVs as a secondary vehicle. The primary will always need to be gas or hybrid.

Exactly! I'm sick of these permanent cellphone modems installed on vehicles. I wish more consumers would fight back, but most don't even realize what's installed in their vehicles and how much tracking data is transmitted. There are aftermarket harnesses to disable them: yes harnesses because if you just pull the fuse, they often disable bluetooth and the drivers speaker (where onstar is injected from).

And now in 2027, cars will be REQUIRED to have computer vision to track you and make sure you're not drunk/sleepy. Yea I'm sure that won't have any false positives or leave people stranded in the middle of nowhere.

BUY OLD CARS. Fuck this dystopian surveillance-scape.

Wanting America to lose a war we had no business starting in the first place by the President of the United States of Israel isn't TDS. The swing voters in the middle were voting for the person who stood against wars and kept America out of it the first time. After the three entirely fake assassination attempts (his ear is 100% fine folks. He could have at least surgically removed a chunk to make it believable), it's clear this was a very long con game. I think Trump knew he'd go to war in his lame duck term for his handlers and the Epstein class that truly runs the world. It was just deferred by 4 years, and who knows, maybe that whole pile of bullshit was wagging the dog too.

Even if you can't hear it, the ultra low frequency sounds from these massive structures can fuck with people pretty bad: https://www.youtube.com/watch?...

There were people in the loop of that entire process and she was illegally detained for 6 months for a crime they had zero probable cause for. Humans don't mean shit if they outsource their brains.

It has to be on-site for that amount. Mixer trucks can't be on the road for too long. They get filled up and have to immediately be dumped out. If one gets stuck in traffic for too long, the entire concrete drum might have to be removed, and they're often buried at the destination site.

That's what happened to the "Killdozer" guy. He was using a buried cement mixer drum as a septic tank and the city was like "you gotta hook up to the sewer, shit is literally flowing into everyone else's land from your place." Then the guy went on and on about how the town was against him or some shit because he was fucking insane. Then he built his armored Kotomasu and rampaged through the town. The evacuation order was the only reason people didn't die. He drove through a library full of children!

The 2019 documentary Tred goes into a lot of the details. That guy is held up as some kind of libertarian hero when he really was a mentally disturbed psycho.

The elected officials in these places are absolutely shitty beyond any measure. Someone needs to audit them and see how much they're getting in kickbacks. In this particular case, thousands of locals were protesting and the elected representative tried to tell people the protestors were shipped in from other places:

https://www.youtube.com/watch?...

The guy is also bitching about people protesting at his house. Bitch you were elected and you're ignoring the will of the majority of your electorate for a colossal center company nobody fucking wants! If we went back a century and a half, you would be dragged from your house, and your family would be tied up and made to watch it burn to the ground.... and then you would have been voted out.

America has a dark history. The data centers are for the surveillance needed to stop the will of the people.

It benefits huge surveillance companies that want to invest in pre-crime and automating the criminal justice process .. like the girl from Tennessee who was extradited to North Dakota, a state she had never visited in her life, because an AI/computer-vision camera matched her via facial recognition to a crime she did not commit. She was detained for six months, and when her lawyer finally got her out, she was left outside the jail, with the clothing she came in with, no jacket, no winter clothing, no money and no airplane ticket back home. I hope she finds a more competent lawyer and clears out that fucking state for $12 million or more after legal fees and taxes! Fuck this AI bullshit.

When they first splurged this bullshit, I immediately though, "It's named 'Mythos'? Really?" Literally "Mythical"

I've read some post that show GPT-5 could find some of the same vulnerabilities if pointed to the same code, and the Mythos version that found some of these issues spent $2k or more worth of tokens on them.

I recently broke out Opus and Sonnet again on my personal projects (try to restrict LLMs to work where I don't care so much) and I found myself rewriting over a 1/3 of the output, even after trying to get the agent to fix issues. It's really a big quantity over quality issue still, with the latest and greatest models. Sure they can build things fast if you need unpredictable spaghetti code shit. Maybe great for one time migration scripts.

One of my managers showed me some MCP servers he setup and how he got Claude to connect to Grafana, examine his Pods, create a full dashboard and even automate alerts. It was kinda cool, but I was like "You used your read-only API keys for AWS/Grafana/etc, right?" ... He used full access, said you had to.

I worry about this level of dependence. I also have a feeling if I dug into those graphs, half of them would have bad queries or not be gathering the information they claim.

> quantum computers

These are such BS marketing fluff. It's been decades and IBM shut down their fun little Cloud Quantum experiment where regular people could run simple little sets of gates.

The record for quantum computer size seems to be 6,000 qbits right now, but they can only be held together for a few seconds at most. I don't think quantum computers will become a reality any time soon, and there is still a considerable debate on if some of these more advanced quantum gate based algorithms can really break older encryption in a reasonable amount of time.

Yea exactly. All it's probably doing is having a background prompt/process determine what seem to be the most repeated things in the context, or maybe the things with the highest degree of "frustration," and just save those to a file that will always get loaded as part of the initial context window for a given folder or project.

Spin some simple concept a lot of developers might already be doing by hand anyway with rules and skills files into some bullshit market speak of "Dreaming."