Anyone can drive an F-22 into a ditch. Doesn't mean that it can't also be used to drop a 2k lb. bomb down your chimney from 40,000 ft.

That demonstration is interesting, but not really something new. Fooling very intelligent people into believing something completely absurd is incredibly easy. How many scientific papers have been retracted based on wholesale fabrications that fooled an entire review committee?

The question isn't "What is the dumbest thing I can do with this technology?" its "What is the most valuable thing I can do with this technology?"

The technology is so dumb can be easily made to believe there is a Google mushroom. We are way far from driving a F22 to the ditch...although I am sure with the same techniques, we could make the AI make the F22 bomb the Google headquarters....

A table saw is technology so dumb it can be made to chop off your fingers.

An air conditioner is technology so dumb that it can be used to kill an infant with hypothermia.

A human is a sentient being so dumb that it can be made to believe in things far more outlandish than a “Google mushroom”.

I can keep going. The point is that just about anything useful can do something dangerous or stupid. Most people can see that. Most people are more interested in how useful something can be, not how useless it is when intentionally misused.

Why is the comparison to someone being able to intentionally crash their own F22, to the above example of intentionally trying to get bad results from their cheapest AI, a bad one?

Every employee doesn't need $1k in token spend per month, either. That kind of spend makes sense for technical workers in r+d.

Most other workers are served fine by $20-30 worth of tokens on a budget model. You don't need Opus to help support write emails.

No, but you do want Opus-tier models to do desktop and office software automation (think about people who intensely use Excel and the like). Actually those might take even more tokens that coding in a lot of cases. Why do you think Claude Cowork is successful, and why do you think Codex is leaning so hard into Computer use?

I wonder if you will see app makers begin to open APIs (MCPs) up in ways that replace computer use. Computer use via human interfaces is pretty hacky IME, and if you can use an app that exposes spreadsheets in a way that reduces token costs by 90%.

I'm optimistic that the demand for AI accessibility will drive programmatic interfaces in places where companies were previously reluctant to.

The Dotcom bubble is an interesting comparison.

The general thrust that everything would be online was correct, it was just that the market mistimed and misallocated of capital by a decade or more. There was massive spending on infrastructure capacity that we wouldn't end up needing until the 2010s. There were hype driven valuations completely disconnected from business fundamentals just because a company was an 'internet' company. Things were going from cutting edge to obsolete in less than a year. There were breathless promises that this was business 2.0! Of course, none of that sounds remotely like what is going on today...

I'm optimistic about AI, but I also don't think that it is going to change everything as fast as promised.

The question you always have to ask is what problems does it directly solve. I personally think most of the current problems in software development and really the world at large are not time-bound problems but alignment issues, and all an LLM can really do there is be some 3rd party oracle that gives you an answer without needing other humans to agree with you.

> The question you always have to ask is what problems does it directly solve

Most directly, human labour. Labour is always a problem for capital. At a certain level of AI competence, businesses don't need to pay humans to complete the work they need doing in order to operate. I don't think anyone would dispute AI competence isn't growing steadily.

I agree with you. I think that if we're talking about actual reliable problem solving, we have to be discussing robotic / drone systems. Software is as complex as you want to make it, and always has been.

The majority of Deepseek providers on OpenRouter for v4 pro are in the US. Especially interesting is that they are in the same ballpark for pricing.

They are in the same ballpark for deepseek-v4-flash, but deepseek-v4-pro from deepseek is still around 1/2 of the alternatives.

I'm pretty sure that Deepseek said that pricing was promotional. Be curious to see if it lasts.

V3 pricing from them was right in line with what the commodity providers are charging.

They announced a few weeks back that the promotional pricing was permanent.

German tank and aircraft design and logistics had their own issues that made things worse, but largely, the biggest issue was just not being capable of keeping up with American manufacturing and Russian willingness to throw bodies at the fight.

Just for context Allied tank production was 276k to the axis 67k. Most other production categories show similar ratios. Your tanks can be perfectly reliable, and superior in every way, but it will be hard to win a fight when you are outnumbered 4:1.

Even now, the emerging doctrine from Ukraine, and now Iran, is to fight using asymmetric production advantages. Ukraine is taking out multimillion dollar facilities and ships with five figure UAVs. Iran has depleted US air defense stocks costing billions with a few million dollars worth of drones powered by motorcycle engines.

Exactly. Iran is beating us with balsa wood (paper airplane) drones. We're outspending them, but they can produce them faster and cheaper. We have to spend hundreds of thousands/millions per drone and missile. They are producing them for a fraction of that. We're being sold an expensive military industrial complex that actually is going to fail when put to the test of reality.

with the help of US industrial advisors who helped them set up factories / production lines.

https://www.youtube.com/watch?v=vrclztGCg6M is a decent watch, though really only tells the "US out produced everyone" part of the WWII story.

I recall seeing a better article that talked about WWII tank production but I can't find it right now.

The USSR was also able to protect most of their factories from bombing raids by moving them to the Urals. Germany and Japan never had that luxury.

You have some false beliefs about Detroit’s product lines.

Fords mustang mach e is unrelated to the gas one. Totally different platform. It was designed from the ground up as an electric platform. The electric F-150 was a shrewd bet that didn’t pay off, but a pretty defensible decision to try to electrify the most popular vehicle in North America without changing the outside look too much. Under the sheet metal it doesn’t share a lot of drivetrain parts with the gas truck.

Chevy makes full size electric trucks on an all electric platform with three truck/SUV models across three brands. The bolt was cancelled in 2022 with the last one produced in December 2023. They have restarted production of the Bolt for the 2026 model year. The Bolt is also built on a platform designed as electric only. The factory building the Bolt is being tentatively retooled next year to build higher margin cars as a result of Trumps tariffs. I don’t see how any of that has anything to do with appeasing Trump.

Stellantis is its own mess, and they suck at designing all cars regardless of energy source.

True, the Mach E mostly just shares the name. The E-Transit really is the same van with a different power train. It ought to be taking over fleet sales, but it only sells about 12K units/year, vs. 162K for the fuel powered version. The Ford electric F-150 was overdesigned and overpriced.

Comments on the Bolt indicate it's being killed due to "changes in the political environment."

> Fords mustang mach e is unrelated to the gas one.

That is right! Some might even go as far as to call the electric Mustang an abomination, a disgrace of the Mustang name.

Depends on the plastic, but even the most benign ones put out very minor amounts of fumes if YouTube testers are to be believed. Note that fumes in this case just means that they are seeing changes in AQI meters, which is something that can happen with too many people sitting in a room.

If it’s something you worry about, the enclosed models frequently either come with, or can be fitted with a carbon filter, and if you are really worried, you can use dryer vent to get it outside.

With the most popular plastic (PLA) at printing temperatures, there isn’t any data showing toxic byproducts. When you get into things like ABS and resin you have to think more about fumes.

Do not print a key fob out of PLA. It will not survive summer car temperatures. A better choice is PETG, or ABS and ASA if you have an enclosed printer that can vent outside.

You leave your key in the car? And you don’t run the air conditioner when you are inside?

> run the air conditioner

Look at Mr. Moneybags here with his car that has the AC add on and still functional!

The FBI isn't involved. A former FBI employee is.

My bad

> Oh, they also claim that they have "the world's largest community-driven public library of Adafruit products, including footprints, symbols, datasheets, and simulation models"². I wonder whether they designed these themselves or whether they use existing ones. Could not easily find licenses info.

Their PCB designs are mostly CC Attribution-ShareAlike typically.

What's funny is that most Adafruit products aren't exactly secret. Most of them have open source schematics and PCB layouts. Even when they aren't, they pretty much just a reference design from a data sheet. The kind of people that have the competence to be using board design software could replicate their designs pretty easily.

what about if I knock on the door (send an http request), and someone comes to it and hands me a bunch of documents (sends an http response with data).