> or in the cloud but way more expensive then it is today.

Why? It's widely understood that the big players are making profit on inference. The only reason they still have losses is because training is so expensive, but you need to do that no matter whether the models are running in the cloud or on your device.

If you think about it, it's always going to be cheaper and more energy-efficient to have dedicated cloud hardware to run models. Running them on your phone, even if possible, is just going to suck up your battery life.

> even though I cannot recall running into a problem that would have prevented by its presence in the before-time

I very, very much did. I was using a Python package that used a lot of NumPy internally, and sometimes its return values would be Python integers, and sometimes they'd be NumPy integers.

The Python integers would get written to SQLite as SQLite integers. The NumPy integers would get written to SQLite as SQLite binary blobs. Preventing you from doing simple things like even comparing for equal values.

Setting to STRICT caused an error whenever that happened, so I knew where in the code I needed to explicitly convert the values to Python integers when necessary.

Heck, once I cycled for half an hour with my iPhone in my pocket, and somehow the phone against my leg was in just the right position that it kept interpreting my leg movements as trying to enter a passcode.

Got home, pulled out my phone, and it had a message that it was locked for several hours due to so many failed passcode attempts. Incredibly annoying.

Still, only happened once in well over a decade of owning an iPhone.

I was mostly frustrated that there wasn't some alternate way of regaining access, like via my Mac or iPad logged in with the same Apple ID. Or that the failed passcode attempts didn't start eventually playing a loud alert sound or something on each failure.

> Those mac minis make a great little server box but losing 8GB to hundreds of processes

It doesn't matter because all the extra stuff just goes to swap. And you can't disable virtual memory anyways. So in the end you're not really losing anything. Those hundreds of processes are ultimately basically mostly just using up a little bit of your SSD, not your RAM, so it's not a concern.

And where's the control?

What about the other thousands of surges in pizza orders that had nothing to do with military missions abroad?

That's why Wikipedia calls it an "informal observation" and quotes the "potential for confirmation bias", asking "When else do spikes occur? How often do they have absolutely nothing to do with geopolitics?"

Not to mention, there's something like 25,000 people working at the Pentagon.

There are so many potential late-night work things happening that would need food, the idea that pizza orders can be used to identify high-profile military missions specifically doesn't make a lot of sense...