Emptying the inventories of the NATO powers to equip the Ukrainians would not have enabled the Ukrainians to roll back the Russians last year. The full explanation of why this is, is too lengthy to fit in the margin of a blog post. The basic theme is illustrated by what the Ukrainians did have this past summer, two NATO-spec combat brigades. Actually one of them was NATO-spec+. The Ukrainians could not use them, the way they are meant to be used. They are not good enough yet to have independent platoons operating in a way that makes the company achieve its objectives, and the companies cannot operate independently in a way that makes the battalions achieve their objectives, and so on up the line. You cannot hand NATO-spec weapons system complexes to effectively-Warsaw-Pact troops and expect the troops to be able to make the complex operate with all of the synergies that are designed into it.
As for the supply of 155mm NATO artillery ammo, it does appear pretty shortsighted for all of the Western powers to have divested of the capacity to make it in bulk. The US has started to address this but it will take a couple of years to build the capacity to make shells at the rate the Ukrainians were using them last year at peak time. It seems plausible that some big chunk of that couple of years is a lack of urgency though.
]]>Russia could get away with this shit when they invaded Ukraine because Russia kept national service, so the call-up mostly got adults who had been through the (highly abusive) draft some time in the preceding years
Um no. One of the reasons Russia has not been able to crush Ukraine is that they divested of all the infrastructure for a mass conscript army also. They go through the motions of conscription but they don't have boot camps or drill sergeants either.
We don't hear too much call for re-instituting conscription here in the US. When you talk about reintroducing conscription in a country that's had decades to build a long-service professional mercenary corps, you're talking about COMPLETELY REPLACING your Army with a DIFFERENT KIND of Army. 20 years would not be a bad estimate for how long it would take.
All that being said, I've concluded that getting rid of conscription and building the AVF was a mistake. I was there when the debate happened, and as a draft-age young man was paying close attention. The thinking was that we would deprive the Deep State (didn't call it that, then) of the cannon fodder for another Vietnam-type adventure. What nobody thought about was what the State could do with an AVF that was militarily effective like they one they built. The interventionism has arguably gotten worse, after the capabilities of that were revealed.
]]>And I don't think there's a way around that. It really is true that the best way to classify LLMs is as automated bullshit generators. They have no mechanism for knowledge representation, just statistical associations. And the magic that turns the statistics into chunks of output also apparently obscures the statistical confidence so that there's not a way to put in a check of the form "if the confidence is less than $X, give up and say 'I don't know'"
Your test problem is probably just about ideal for illustrating this: most of us don't have at our fingertips a topic on which we're the world's top expert, that is noteworthy enough for an LLM to have encountered it, but not so widely-represented in the training corpus that the model can fake convincingly for more than 5 fun facts.
]]>I like this reasoning, and it provides something to stack up against the almost overwhelmingly negative social solvent effect I see in media from the early 20th century onward. Maybe people were being hornswoggled into demanding impossible standards of mate perfection or driven to overconsume by depictions of ordinary people living in opulence, but at least they weren't so bored that they drank themselves to death so often.
]]>To the extent that such a policy were successful, it would lead to gangs killing each other over black-market rights to various sections of your major cities. The violence associated with the cocaine and meth and narcotics trades is about the money, not the effects of those drugs. There would be enough money in a cigarette prohibition black market to drive some fierce violence and serious police corruption.
]]>But I did feel impelled to stop in and note that for the first time in 50 years of SF reading I think I am going to register for WorldCon just so that I can nominate/vote for Lena.
]]>we now share a literary agent, so hopefully you'll be reading more of his books soon.
This is excellent news.
]]>The Labour Party, from over here, seems to have basically nobody in it who realizes that it's not 1993 (or even 1999) anymore. You'd think there should be some space for at least a soft-Left Party with a platform of "Nobody freezes to death or starves" even if they won't go full on seize-the-commanding-heights-of-finance.
]]>For a long time, I felt less urgency about the existing great fortunes. I figured that with a decently structured death tax and a confiscatory income tax they would eventually evaporate via regression to the mean.
I don't have any good suggestions about how to prevent the accumulation of fortunes like Gates' and Jobs' and Musk's, which they got basically by increasing the valuation of something they already owned.
]]>It's worth noting WRT Roe that the smart money did not expect an opinion like the one leaked: the Supremes seldom explicitly state "$PREVIOUS_DECISION was wrongly held ab initio and is hereby made null, it and all the reasoning that depends on it." If they pull that with Griswold, the 2nd half of the 20th century really does get repealed.
]]>I have been in the programming end of the business since before you started, and I have mostly lost track of the tech side. The days when a new hire could be handed a blank computer (or better yet a box of parts) and a stack of distribution disks are long gone. Just knowing how to set up build environments seems to be a full-time job.
]]>