A. AI is fundamentally a model of how humans think. It has to be, because the only example of ‘intelligence’ with which we are familiar is human intelligence. (The same can be said of the concept of ‘artificial.’) As a model, AI is going to tell us what we tell it to tell us. It simply can’t do otherwise. People who understand how models really work understand this limitation – it is obvious.
Concern over AI getting too intelligent and deciding it doesn’t need us puny humans any more is misdirected. The idea that an independent meta-human intelligence will arise, Athena-like, as an emergent property from anything we can build is fantasy. Our idea of meta-intelligence is as limited as our idea of Superman: just as Superman is, fundamentally, a man, just stronger, faster, and incorporating better versions of human tech (laser eyeballs, flight), an AI is – must be! – imagined to be fundamentally human intelligence, only more so – faster, able to process more data at a pop, able to draw connections and conclusions farther and faster. And even this remains fantasy – we have no idea how all this works, but since it does in humans, it must work in our model! The dogma that the human mind simply is a machine demands it.
Putting these two ideas together and acknowledging the limitation inherent in them: What AI may eventually produce is a very fast, very large process that will – must! – be a model of intelligence and the world as the model builders imagine those things to be. AI will produce what its builders tell it to produce.
What we need to be concerned with, then, is not some imagined mysterious, emergent power of AI that no one can control or predict; what we need to be concerned with is what the builders of AI believe and want. That’s what AI will give us. It will give us nothing else. The surprise will be for the builders, as AI demonstrates what they, the builders, truly believe and want.
B. In traditional, by which I mean, obsolete, warfare, an aircraft carrier is the bee’s knees: one modern carrier projects force like nobody’s business. Trouble is, those suckers are expensive: the USS Gerald R. Ford ran a sweet $13 billion to build. And, to make matters worse, a single cruise missile can sink one – Tomahawk cruise missiles, for example, only cost $1.9 million each. You could determine that you needed to launch 1,000 cruise missiles at the Gerald R. Ford to make sure one got through to sink it – and have spent only a bit over 10% of the cost of the carrier to eliminate it. And there are other ways of taking out carriers, such as submarine attack, which are similarly cheaper than building one in the first place.
Knowing this, no carriers go galivanting about unaccompanied. Carriers travel in carrier groups, which include destroyers, frigates, a guided missile cruiser, sometimes submarines – which, all in, will run you $20-$30 billion per group to build, and billions more per year to operate. The main goal of the carrier group is to keep the carrier from getting sunk. So, now, you’ve invested $20-$30 billion, plus billions more per year in operating costs, just to be able to project force along the world’s coasts.
If you wanted to sink a carrier, and had 1,000 cruise missiles at you disposal, and the carrier group was an astounding 99.9% effective in stopping those cruise missiles – you win. But it’s way worse than that:
“The exercise was called Millennium Challenge 2002,” Blake Stilwell wrote for We Are the Mighty.
It was designed by the Joint Forces Command over the course of two years. It had 13,500 participants, numerous live and simulated training sites, and was supposed to pit an Iran-like Middle Eastern country against the U.S. military, which would be fielding advanced technology it didn’t plan to implement until five years later.
The war game would begin with a forced-entry exercise that included the 82nd Airborne and the 1st Marine Division. When the blue forces issued a surrender ultimatum, Van Riper, commanding the red forces, turned them down. Since the Bush Doctrine of the period included preemptive strikes against perceived enemies, Van Riper knew the blue forces would be coming for him. And they did.
But the three-star general didn’t spend 41 years in the Marine Corps by being timid. As soon as the Navy was beyond the point of no return, he hit them and hit them hard. Missiles from land-based units, civilian boats, and low-flying planes tore through the fleet as explosive-ladened speedboats decimated the Navy using suicide tactics. His code to initiate the attack was a coded message sent from the minarets of mosques at the call to prayer.
In less than 10 minutes, the whole thing was over and Lt. Gen. Paul Van Riper was victorious.
Micah Zenko provided some context in a piece for War on the Rocks. “The impact of the [opposing force’s] ability to render a U.S. carrier battle group — the centerpiece of the U.S. Navy — militarily worthless stunned most of the MC ’02 participants.”from National Interest, Oct 15, 2019
So, in a war game, a Marine general was given the resources of an Iran-equivalent power and told to take on the combined might of a large chunk of the US Navy – and, using the few missiles at his disposal, plus suicide speedboats and civilian boats and aircraft, took them out in 10 minutes.
Lt. Gen Paul Van Riper. For real. Damn. My only issue with this: nowhere I can find listed among General Van Riper’s assets ‘armored battle goats’. Because – well, because. As hard as it is to imagine, he somehow won without them.
Um, oops. As Sun Tzu so aptly put it: to know your enemy, you must become your enemy.
No reason I’m thinking about this. What could possibly go wrong? I’m sure our current president, what with his razor sharp intellect and surrounded as he is by Top
Men Humanoids, has this sort of thing completely under control, no matter who the enemy might turn out to be in this best of all possible worlds.
C. Been under the weather due to circumstances well within my control that I, nevertheless, failed to control. Something about making sure prescriptions got filled before health plans flipped. Dolly Parton once quipped: “It takes a lot of money to look this cheap.” Does it take a lot of brains to be this stupid? No, I think I just have a talent for it.
But much better now! Will get back to the writing soon. No, really! Haven’t totally neglected it, but not going gangbusters, either.
D. Looking like we might have an epic fruit season out in the front yard micro-orchard. This past winter, I was better about clean-up, trimming, fertilizing, and spraying copper fungicide. Also watering a bit more, as we only had 40% of average rainfall this season:
- Fig tree has lots of breba figs on it
- Cherry tree has several times as many cherries as last year
- Pomegranate just starting to bloom, looking beautiful
- Our latest additions, two blueberry bushes, seem to be doing well – one is covered in fruit and blossoms, the other has less but is growing vigorously
- 4-in-1 pear tree, devastated last season by that loathsome leaf curl fungus, is now looking pretty good, with way, way too much fruit setting – I’m going to need to thin by about 80%!
- My two little peach trees are doing well. Last year, one caught the leaf curl from the pear tree next to it, and lost all its fruit and leaves, but recovered enough to put out enough leaves to survive – it actually looks good, and has a fair amount of fruit on it. The other peach, a dwarf variety, is insane:
- Apricots are doing very well, too
The nicest thing: the Minneola tree our late son Andrew grew from a seed as a child is, for the first time, covered in blossoms:
This tree is over 15 years old. Last year was the best ever – about a dozen fruit. Now, if even 10% of the blossoms set fruit, we’re looking at many dozens. The fruit is good, nice and sweet.
Andrew wrote a poem about it (it was presumed to be an orange tree at the time):
My Orange Tree by the Wall
by Andrew Moore
My orange tree by the wall
For many a spring and fall
Has grown and grown and grown
And done nothing much else at all
But then in spring one day
I shout ‘hip hip hooray!’
For blossoms it shows me
And oranges it grows me
For many a long summer day
E. Further updates as events warrant.