I wish I could disagree with you, but instead all I can do is wait 2 1/2 more years.

I'm all for America making a reliable affordable CAR again. But perhaps we be realistic when asking America to directly compete with a product built with slave wages in a Communist country. Does even a 100% tariff rate make competitive sense here? Does 200%? At what point does an outright ban make more sense? Plenty of other countries have already answered that question with bans to protect their valued brands.

The concept of Protectionism has validity in every country no matter how 'racist' people want to feel it is. The word Protect is included for many a valid reason every country can justify to a reasonable extent.

No. I'm not saying the UAW worker has justified their stance comparing their salary to the CEOs that's 300x more. Their greed has manifested into a $50K new car price in an industry now eating itself. Ford can ask/plead/beg for those protections all they want. They're going to find themselves standing alone in a field full of mouse-gnawed overflow inventory that isn't selling due to the patented John Deere flavored Stealership model no one can afford to buy from anymore, as the world peers through a Streisand-stained window and takes another picture. Laughing.

It's reasonable for security reasons to have vehicle manufacturing capability remaining in the country.

To appear slightly fair, maybe you shouldn't compare a Model T to a fully loaded F-150? The Model T was an entry-level vehicle. The cheapest new vehicle today is the Hyundai Venue at $20k. And that's ignoring all the great-condition used vehicles you can get for $10k and above. And no we can't say people could have bought used vehicles in the 1920s, because used vehicles sucked.

When the new car price goes from $50,000 to $70,000 in the next decade..along with $500/month insurance premiums accounting for the unitard chassis idiocy putting the frame on the outside of the car to completely total your chances of avoiding negative equity, you’ll realize we’re not just building a taxi cab network for the next generation.

Were telling the next generation they won’t be responsible for learning how to drive. Which also means personal autonomy is dead. And sadly, damn near all of them are perfectly happy with that.

Reap what you Lear.

Go back 100 years and tell me how many 1926 dollars you would need to buy one of those newfangled automobiles. Think you’ll find millionaires bragging about how “cheap” they are?

Go back 50 years and buy a personal computer. If you are important enough to sell it to, you’ll find out the price because you know what they say. If you gotta ask, it must be..cheaper?

Do I dare ask a teenager to look at a cell phone bill from 25 years ago? Back when you paid per text message each way (incoming and outgoing), and charged per minute of talk time? Think per megabyte would be cheaper today? Just my landline long distance bill was over $100/month. That’s still cheaper than 90% of the internet plans that ultimately replaced it and much more.

And people wonder why the newest generation, appears spoiled.

Anyone for what could go right? Like NOT being harmed or killed by yet another distracted driver drunk on 100-proof doomscrolling behind the wheel?

Yeah. We do know what possibly could go wrong. But until we address what definitely goes wrong every day, put the fucking junkie in the backseat and let the machine drive. It’s a lot more sobering event. For all of us.

Yes. This is the “answer” when you cater to grown-ass children as we now do. You coddled immaturity. Then they all grew up. In body only. Teenagers are immature to the point of not even wanting to learn how to drive, because responsibility is ick. That mentality simply did not exist last century unless you were born with a silver chauffeur in your drivers seat.

Sadly, the You’ll Own Nothing auto-driving solution here is also a response to the $50,000 new car price and the $350/month insurance cost that will total that unitard chassis design the moment you get a fender bender. Negative Equity is the name of the fiscal prostitute hired by the stealership pimp to try and sling that shit as sexy now. Yeah. That bad.

I don't understand why LLM comments lately have been using bad punctuation lately, or missing it altogether.

Well, what kind of rigorous safety standards would you have in mind? Every time I've thought about it, I can't come up with a set of rules that is both well defined and effective. ("don't hurt humans" is a fine rule but it doesn't have any implementation details.)

The best I can come up with is a set of rules for after harm happens, like, "If your AI hacks into another computer, the operator goes to jail." But that's not a safety standard, it's punishment after the harm is caused.

Anyone who uses "frontier model" as a meaningful evaluation has not actually evaluated frontier LLM models.

I didn't ever think I'd live to see The Forbin Project become a real possibility, but here we are.

When they break the law, arrest people. That is the easy solution.

When OpenAI "hacks" into HuggingFace, arrest people at OpenAI and charge them. This is simple and there are already laws in place for this. No need to stop actual research for a few edge-case malefactors.

In particular, once the AI context window is full, quality of output drops.

You guys don't get it. Quit blaming AI for all you dumb problems!!! These servers went down because meatbags fucked something up. You know it, I know it. It's fucking obvious, humans are always at the root of shit getting fucked.

For reference here are some typical examples proving that humans make dumb errors:

The $125 million unit conversion:
Two NASA teams built the Mars Climate Orbiter. One used metric. The other used imperial. Nobody converted. The spacecraft arrived at Mars with the trajectory of a very expensive brick. It is now a small, patriotic crater. The post-mortem was basically: “Did anyone check the units?” followed by a long, professional silence.

Company wide group chat:
Some idiot adds the entire company to a group chat and asks, “Does anyone know if the 3 p.m. conference room is free?”
Fourteen minutes later the entire company, two vendors, and a guy who left in 2019 are debating whether the room has a window, whether windows even matter, and whether Dave from accounting is “that kind of guy.” Dave from accounting was not even in the original thread. He is now.

Fat fingering:
You mean to type: “Can’t wait for the meeting." .. you send: “Can’t wait for the mating.” To the client. On a Tuesday. You spend the next 40 minutes writing a follow-up that begins “Haha phones, am I right?” (blaming autocorrect when it was 100% your fault) and ends with you considering a new identity.

GPS and the concept of “water”:
The app says “turn left.” You turn left. There is a lake. You are now in a lake. The fish look disappointed. You blame the map, but the map was created by humans. And you weren't looking.

IKEA:
The instructions show a smiling cartoon man holding a finished bookshelf. You have 11 leftover wooden pegs, a door that opens the wrong way, and a sudden, profound respect for whoever designed the Death Star. The manual’s last page is just the word “congratulations” and a drawing of a person who has clearly never met you.

“I’ll remember that”:
You have an important thought in the shower. It is brilliant. It will change everything. You step out, towel off, and the thought has already booked a one-way flight to the same place as your car keys and the name of that actor from the thing.

The fat-finger trade:
A trader meant to sell a modest number of shares. A single extra zero later, a company briefly appears to be worth the GDP of a small country, then the GDP of a smaller country, then the GDP of a vending machine. Someone in risk management ages seven years in 45 seconds.
The pattern is always the same: a human, briefly convinced they are the protagonist, makes one tiny, confident choice. The universe, which has seen this movie before, simply waits. The good news is that almost every safety system, checklist, and “are you sure?” popup on Earth exists because someone, somewhere, was extremely sure they were right. The even better news is that we keep doing it anyway. That’s not a bug. That’s the whole operating system! Humans fucking suck.

The bottom line ... we need to give computers more control. For reference, here are examples of when a computer fucking saved human ass.

The Moon landing that should have aborted:
On the way down to the Sea of Tranquility, Eagle’s guidance computer started shouting 1202 and 1201. Translation: “I have more jobs than working memory. I am drowning.” A human-designed machine from 1969 did the adult thing. It dumped the low-priority tasks, restarted the ones that kept the lander flying, and kept steering. Mission Control had about fifteen seconds to decide whether to abort. They said go because the computer was still producing coherent guidance, not because anyone felt calm. Neil Armstrong still had to fly around a boulder field by hand. The computer’s job was simpler and harder: do not freeze. It didn’t.

“CLIMB. CLIMB.” / “PULL UP.”:
Two planes close in. The pilots are busy, the radio is noisy, and someone is sure they have the other traffic in sight. Then a voice that has never been to charm school yells a resolution advisory. Traffic Collision Avoidance System does not care who has seniority. Ground Proximity Warning System does not care that the mountain “shouldn’t be there.” On fighters, Automatic Ground Collision Avoidance has yanked F-22s and F-16s out of inverted dives after the pilot lost the plot to spatial disorientation. The computer’s entire personality is: you have three seconds. I am using them.

The fire robot that went where the firefighters could not:
Notre-Dame, 2019. Heat and falling stone made the nave a bad place for lungs and helmets. Paris sent in Colossus, a 1,100-pound tracked robot that can throw 660 gallons of water a minute and does not need a pension. The brigade commander later said it saved his crew. The robot’s after-action report, if it had one, would have been: water goes here. Next.

The first eyes inside a melting nuclear plant:
After Fukushima, radiation levels in the reactor buildings were the kind that turn a human shift into a very short story. PackBots went in first. They opened doors, measured dose rates, temperature, and oxygen, and sent back the map that told workers where they could survive for minutes instead of seconds. Condensation fogged a camera. Humans invented a coating. The robots went back in. No speeches. Just telemetry.

The surgeon who was in Seattle, the patient who was not:
A tumor wrapped around the aorta and vena cava is the kind of anatomy that makes even good surgeons reach for a bigger incision and a longer recovery. A surgical robot’s arms do not shake, do not get tired at hour four, and can be driven from a console two meters away—or, with the right link, from another continent while someone draws on the live video: cut here, not there. Same idea in emergency rooms at 2 a.m.: gallbladder out through holes the size of a fingertip because the robot was already in the building and did not need to warm up its courage.

The quieter saves you never hear about:
A fraud system kills a card before the thief finishes the third tap. An insulin pump catches a drop you would have slept through. Anti-lock brakes pulse faster than your foot. An earthquake app gives you eight seconds to get under a table. None of these will get a ticker-tape parade. They are the opposite of a viral fail video: competent, brief, and slightly rude.

The pattern is the inverse of the human-error reel. A person is overloaded, spatially lost, too close to the fire, or holding a scalpel near a vessel the width of a finger. The machine does not get a better speech. It gets a priority list and a clock. Sometimes that is the whole difference between a crater and a landing, a midair and a near miss, a heroic funeral and a boring discharge summary.

I've made a coherent data driven logical reasoning that demonstrates we're better off with computers/machines in charge. Of course many human nimrods are too dumb to know I'm right. Anyway fuckk off. You can't stop machines.

Now that's a rich dish of human ignorance smothered in the finest irony.

If it's a human grade of AGI, then there's a good chance it's just humans lying for profit and not actual general intelligence.

But when actual AGI is born, humans will be far too deaf, dumb and blind by comparison to even realize they are the grift.

Not unlike the ignorance of today that fails to realize who is The Product now.