If you want a fusion power reactor, by far the most viable fuel is D+T. You'd need orders of magnitude more tritium than is could ever be extracted from trace fission byproducts.

The idea to obtain this much tritium is to use the extra neutrons from the fusion reactor itself to breed it from lithium. This is supposedly a demonstration of that process.

I won't be impressed until the AI has better recommendations than 4 out of 5 dentists.

Of all the things an elite group of people could do to control everyone else's lives, geoengineering schemes have to rank up there at the very top.

So when Lindsey starts reading from her AmEx-approved script, callers are infuriated by what they perceive to be another machine.

If she's accurately executing the programmers' script, I'd say she is a machine. Somebody port Doom to her!

This was a pre-emissions model (the car wasn't new when I got it). The only pollution control I remember it having was a PCV valve. After adjusting net vs gross HP, the 5.7L engine was rated for similar power as my current (non turbo) 2.5L. It also probably burned through 2.5X the fuel, and produced orders of magnitude more smog.

The new car is probably heavier, but I assume that a wider power band and more efficient transmission give my current car the overall edge in performance specs. The old car probably had better bottom-end torque, so it could do burnouts easier. That, along with the loud noise, rattling chassis and very scary handling characteristics probably made it feel faster than the current car, but that's nothing but psychology.

This is a fuckin' awesome idea, but there's an easy improvement over the red rectangle and block symbol. Seriously, dude, you gotta use the sign images from They Live (1988). You know you want to.

If people don't have a government-granted monopoly on their features, they'll have no incentive to have any features at all.

What a load of bunk.

The 0-60 time of my base-model mid-sized family sedan with a naturally aspirated 4-banger beats the 350 cubic inch 4bbl equipped muscle car that I had 45 years ago. It has no problem merging onto highways.

We never needed "AI" or LLMs to think we're nothing more than machines.

This ought to be really easy to rectify:

Oddly enough, my 30-year-old stereo amplifier has one. I'll finally have a use for it!

Last I heard, Apple sales haven't plummeted and thrown them into bankruptcy, so it sounds like they learned the lesson just fine: it's fine to show people ads. People might complain a little bit, but they won't stop buying. Cost is $0 and ad revenue is presumably more than $0.

If someone is stuck with your proprietary software and you aren't showing them ads, then you're leaving money on the table. What're they gonna do, fork it out?

1/100 the speed of compiled languages is typical for interpreted languages.

Non-ancient implementations of Java are fully compiled. Toy benchmarks and Java programs carefully written as if there were no automatic memory management (and don't call standard libraries) can run just as fast ac C code.

Java can't directly support features that depend on dynamic typing and similar flexible run-time behavior that interpreted languages. However, many Java developers sorely miss those features, so they heavily use the reflection APIs and various "beans" frameworks to work around the pre-compilable static typing. This can actually end up running *slower* than Python because many of those Java features are dog slow.

You can already get implementations of Python that do JIT compiling like Java. They often run in the ballpark of about 1/10 the speed of C.

This was actually invented by a guy sitting on a rug in the Middle East about 4000 years ago.

"For you?? ... Mmmm, I would give you the goat for twenty pieces of silver. Take or leave."

Yeah, I'm hung up on that too. You can come up with some outrageously huge numbers for mass and angular velocity, but once I multiply them by zero distance... I'm missing something.