Douglas_Knight

Karma: 7,914

Douglas_Knight 27 Aug 2026 17:34 UTC
2 points
0
on: Romae Industriae
You seem to ask: if Rome had technological progress, why didn’t it have an industrial revolution?

I reject the premise. It appears to me that in about 150BC Rome destroyed Hellenistic civilization and not only produced a backwards shock to technology, but started 1000 years of regressing frontiers.

Why focus on the point event of the industrial revolution? If we have a grasp on general technological progress, we should study the broader question. If we don’t, we might focus on the industrial revolution as an undeniable event. But then we should be cautious about claiming that Rome made progress. You mention the Antikythera mechanism as a reason for uncertainty. But this was pre-Roman. If anything, it suggests that the Romans were behind the Greeks. You mention Archimedes, who was also pre-Roman. You mention Vitruvius, who was Roman, but was fairly early. In particular, you can see regression by comparing him to Frontinus, a century later.

You say that Rome had centuries of per capita growth. But your link is specifically about Britain. That was the most backwards Roman province. It did see centuries of technological advance. Rome did spread technology even as the height fell.

Why did Rome see technological regression? Actually, I think that is not such a difficult question. Rome had insecure property rights leading to consolidation and diseconomies of scale.

Of course, this leaves me with the question: Why didn’t the Hellenistic Egypt have an industrial revolution?

Douglas_Knight 26 Aug 2026 19:19 UTC
2 points
0
in reply to: aphyer’s comment on: AI Text Watermarking Is Free And Good
I believe it makes it more deterministic, but not much more deterministic. It knows the entropy budget and only consumes a small amount.

I do not believe that it is theoretically necessary for it to be more deterministic at all. If it were deterministic, it would be recording not just the watermark that it is produced by LLM, but also the fact that it was produced at the beginning of the session, which is not the goal and thus wasteful. The purpose is that the watermark can survive the prefix being truncated, so there is a potentially arbitrary context that the watermark should not be sensitive to. It is about correlations between successive calls to the RNG, but it could still be initialized randomly. That initialization is like the unknown context.

Douglas_Knight 16 Jul 2026 2:08 UTC
3 points
0
on: The mosquito bucket of doom works
Why does this work? Why does adding new bodies of water save you from the mosquitoes that were already breeding in the old bodies of water? My first thought was that if you add 10x as much water area, then 90% of the eggs will be laid there. But you explicitly say that if you treat big bodies of water, the mosquitoes will still breed in small bodies, which seems directly opposed to this strategy. Maybe this only works in a narrow range of conditions, but you might as well try it.

Douglas_Knight 1 Jul 2026 19:21 UTC
−1 points
−2
in reply to: Thomas Kwa’s comment on: Surprising facts about the slave trade
The clearest example is prewar Virginia where there was direction competition between free and slave labor. The free labor outcompeted slave labor at growing tobacco, which is why people expected slavery to die out on its own. Then the cotton boom caused Virginia plantations to switch to growing slaves.

You also make a theoretical argument. At most that suggests that the landowners should earn more. That is not the same as production. Moreover, free people put most of their surplus into growing the next generation of free people. Empirically, free populations expand much faster than slave populations.

Douglas_Knight 29 Jun 2026 14:25 UTC
9 points
−3
in reply to: lilkim2025’s comment on: Surprising facts about the slave trade
It is true that industry allowed England and the northern USA to impose their will on agricultural slave societies. But it is also true that big plantations aren’t efficient. Agricultural production increased after the abolition of slavery in both the Caribbean and the American South.

Douglas_Knight 21 May 2026 3:22 UTC
6 points
0
in reply to: Simon Lermen’s comment on: If AI is normal technology, history is not reassuring.
Just because the Holocaust was “industrial” doesn’t mean it was efficient. Slaughtering prisoners is physically easy. Genghis Khan set a quota of 300 executions per soldier per day. The death camps were about as efficient, using 25 people to kill a trainload of 2500 people in an hour. The trains were less efficient than the “Holocaust of bullets” that preceded them. The trains and death camps were not about physical efficiency, but about concentrating the responsibility on a small number of people who were willing to kill. The gas was about saving even those elite killers from the psychological cost of looking at their victims. Indeed, an early version of the death camp was to shoot through a window into the back of the head, clearly about psychology, not even saving a bullet.

Douglas_Knight 18 May 2026 17:48 UTC
2 points
0
on: James C. Scott: Seeing Like a State
The Green Revolution was like mechanization of Soviet agriculture in that they both involved the obvious belief that technology would increase productivity. And specifically the belief that they would lead to consolidation of family farms, as had happened with both technologies in America.

The Green Revolution was not top down. The governments of Mexico and India did not impose the new crops and petroleum fertilizers on their people. Indeed, they would rather their people starve than consolidate farms and create an urban proletariat. The government of Mexico (and the Foundations, essentially the USG) ordered Borlaug not to do it. The government of India fought the new technology for years.

Maybe central planners want consolidation so that there are fewer people to tax. They individual is not more legible, but he is more productive so you can devote more resources to understanding him. I think this is orthogonal to much of what Scott writes. Maybe these technologies produce more predictable outputs. That would be exactly the legibility Scott talks about, but I am not sure it is true. I am pretty sure that is not why Stalin did it. He merely demanded numbers, rather than engineering Scott’s legibility.

Douglas_Knight 25 Apr 2026 2:52 UTC
2 points
0
in reply to: rsaarelm’s comment on: Why no new notations since 1960?
How about linear logic? Maybe it’s a just small extensions, comparable to modal logic, but it’s new.

Douglas_Knight 23 Apr 2026 19:50 UTC
6 points
0
in reply to: kbear’s comment on: An Angry Review of Greg Egan’s “Didicosm”
What do you mean by “explicit”? That you can give a citation? Would you do so?

Adam Becker claims that his lifestyle was his goal from before starting his PhD.

Douglas_Knight 20 Apr 2026 19:27 UTC
2 points
0
on: Let goodness conquer all that it can defend
Was Truman the man in the arena? Did he take any actions? My understanding is that he was a scapegoat, both about the bombs and about his presidency in general. The buck stops here for responsibility, but not decisions. He didn’t make the decision to bomb the city of Hiroshima, but specifically wrote that he chose it because there were no civilians. He didn’t weigh the options and decide that it was worth the cost, but rewrote history to pretend that he did. No, he shouldn’t wring his hands about his lying subordinates, but he should do something about them. I hold him responsible for complicity in this charade.

Douglas_Knight 15 Apr 2026 17:21 UTC
4 points
0
on: Some takes on UV & cancer
Sometimes you say cancer and sometimes you say deadly cancer. How careful were you to distinguish these? My understanding is that sun exposure is positively correlated with melanoma and negatively correlated with melanoma deaths.

Douglas_Knight 15 Apr 2026 16:29 UTC
2 points
0
in reply to: mruwnik’s comment on: Annoyingly Principled People, and what befalls them
Semmelweis and Wegener are not so good examples because they were met with explicit rejection. I don’t like the example of Serpico, but I have trouble articulating my objection. Did he ever doubt himself? If he started at the top of a slippery slope, he explored all the way to the bottom and stopped doubting himself. He said he was subject to ridicule. I think it would matter whether he was condemned for having impossibly high standards or whether people simply defended bribery as a perk of policing.

Douglas_Knight 9 Apr 2026 18:33 UTC
2 points
0
on: Am I the baddie?
What is bad about this? You only get to that in the last paragraph. I can’t quite tell what you mean. There seem to be two topics. One is doom and the other is job loss. These are totally different issues. If AI use is contributing to doom, that is bad. I doubt it, but let’s leave it aside. The rest of the paragraph seems to be about job loss. This seems like a general argument against technology. Do you really endorse the general argument? If not, what is the difference with AI?

Douglas_Knight 27 Mar 2026 19:58 UTC
2 points
0
in reply to: Gurkenglas’s comment on: Dispatch from Anthropic v. Department of War Preliminary Injunction Motion Hearing
The Pentagon said in court that they did it because of Anthropic’s actions in the press. It would not be surprising if before they said that in court they wrote memos saying that. But if (as I predict) these are very recent memos, they tell us nothing about the real reason. The judge is taking them at their word and interpreting tweets (eg, “master class in arrogance”) as being about that. I don’t think those tweets were about the public conversation at all. In any event, they were after the designation was mentioned in the press, so are not evidence about the Pentagon’s motivation for the designation.

I think it is probably true that some people at the Pentagon saw unfavorable press treatment of their own leaks and assumed that they were Anthropic leaks. Possibly because other people at the Pentagon lied to them. But they were already talking about the designation before that.

The first mention in the press of the supply chain risk designation is on 2026-02-16 in Axios. It was explicitly named as punishment “for forcing our hand,” which is unclear, but doesn’t sound like public scrutiny.

Douglas_Knight 26 Mar 2026 20:55 UTC
3 points
1
in reply to: Zack_M_Davis’s comment on: Dispatch from Anthropic v. Department of War Preliminary Injunction Motion Hearing
You got the impression because Michael wanted you to. But he is a liar, so don’t attempt to extract any sense in which there is a kernel of truth.

I addressed your parenthetical in my comment. I think the negotiation began after the election. The Biden officials could have shaped the deal, but Palantir and the generals knew that they were temporary.

Douglas_Knight 26 Mar 2026 20:52 UTC
5 points
0
in reply to: Gurkenglas’s comment on: Dispatch from Anthropic v. Department of War Preliminary Injunction Motion Hearing
That is ridiculous. It was the Pentagon that went to the press, not Anthropic. And the Pentagon explicitly told the press that they were punishing Anthropic for refusing the contract, not for talking about it.

Douglas_Knight 26 Mar 2026 4:44 UTC
20 points
0
on: Dispatch from Anthropic v. Department of War Preliminary Injunction Motion Hearing
Anthropic signed its contract with the Government in July 2025, the current administration, not the previous administration. It announced its partnership with Palantir a couple days after the election, probably to identify the new administration before beginning negotiations for this very reason.

Douglas_Knight 13 Mar 2026 18:13 UTC
2 points
0
in reply to: JennaS’s comment on: The first confirmed instance of an LLM going rogue for instrumental reasons in a real-world setting has occurred, buried in an Alibaba paper about a new training pipeline.
Good call. The follow up is not entirely clear, but it looks like the teacher was trying to simulate an attack that the student could investigate.

Douglas_Knight 3 Mar 2026 18:16 UTC
6 points
4
in reply to: Max H’s comment on: Epstein and my world model
It’s one thing to say “I know X is controversial, but I want to assume it and talk about the consequences.” But saying that there “this article is for people who know X and not the ignorant” leaves it fair game to attack X.

I have a different complaint. Saying “Epstein stuff” and not being specific could create an alliance between people who believe contradictory things. This is a common pattern. For a crisp example, consider anti-carb as an alliance between people who think glucose is fine and fructose is poison with people who think the opposite.

Douglas_Knight 28 Feb 2026 22:29 UTC
13 points
2
in reply to: Kaj_Sotala’s comment on: Kaj’s shortform feed
I am skeptical of the framing “this episode” and the relevance of the Venezuela raid. The ~~Pentagon~~ White House has been complaining about Anthropic since September. There is a Pentagon memo dated ¹⁄₉ that talks about rewriting contracts to use the phrase “any lawful use.”