Farming is odd, because there's a huge inter-generational phobia about selling the heritage, but there's also a degree of pragmatism (bad year? shoot the cows) going on. And, there's also often a massive debt overhang, tolerated by the system as a whole, because of the boom-and-bust nature of the business. It wouldn't surprise me if a lot of land owners told "$14m for your paddocks" would only see a low value in that, because of the debt overhang.

It also would not surprise me if a lot of these prices are tail payment, and predicated on the plan getting up, and involve middlemen who are brokering the land up to bigger players in the hyperscaler space. Lots of NDA, not much money up front.

I would think that there are brownfield ex-factory sites which make more sense.

Farming is agribusiness is big business. People who are in farming for the long haul. But, if its not a family enterprise (and a lot isn't any more) and if there's a board, and shareholders, then the temptation to cash out and move money to another place is there. So I expect despite all these refusals from one cohort of the farm sector, there's another one which is looking at the merits of this and for the right package, will jump. But they will want a lot more than ag. value per hectare/acre on this, they won't settle for the value as soybeans, they want the value as a multi-billion dollar enterprise outcome.

I sure hope some of them really are the green heart, who value nature. But I think the story is going to be a lot more multi dimensional than that.

How would this work in the context of the various VM models like UTM on OSX? If this is a path to a good underlying abstraction, it would be nice to get better 2D rendering.

Chip making has a long, torrid history of using toxic, long lived chemicals and not cleaning up very well. They also use a lot of water, and that's often a concern with supply level continuity at risk.

If you don't take great care, and let the stuff you work with get into the groundwater, or assume free running water clears and dilutes your problem, then you're a strong candidate for being opposed by people who care about water quality, in supply and in the food chain and downstream.

Oddly, The AI company doing the deals with the US military, and refusing to step over a line on the use of their tech, is the same company which just removed a key word from their mission statement, one which in some ways was a visible statement toward rejection of dangerous use of the tech, and so by removing it looks like a move toward what the military (or their leadership) want.

It's almost like an alternative history with google removing "don't be evil" at precisely the same time as being locked in a debate about .. how evil to be.

Run local root. Rootservers are not essential. It's in ietf draft discussion now as 4 documents but already works and just has to be turned on.

If you want to change pace, ask your dns sw provider to turn on local root by default.

(One of the things being defined is how to get a root zone trustably out of band using the new ZONEMD checksum)

A bigger question might be why there are no ICANN HSM outside the USA to generate root zone signings. ICANN has offices in Geneva and Singapore, it would not be hard to find secure DC locations for the signing ceremonies.

"made" == "assembled"

I fed the abstract into an AI with the prompt:

  can you give me a critique of https://arxiv.org/abs/2602.16291 from the point of view of a functional programmer who is skeptical

and I liked the answer. I won't post it here, because I think AI answers are generally noise and quite tiresome, but there is one sentence which stuck in my mind:

  The extraordinary breadth of the claims calls for an extraordinary level of rigor in the proofs.

In the spirit of equality, I then posed this prompt:

  can you give me a critique of __https://arxiv.org/abs/2602.16291__ from the point of view of a imperative programmer who wants to believe

Which resulted in the exactly opposite emotion (in me as a reader. The machine does not have emotion. It's statistics)

However, this sentence also did appeal, in a strongly pragmatic sense:

  My ask to the author would be simple: show me a real, ugly, 500-line configuration or dependency-injection problem, solved cleanly in Overlay. If that holds up, I'm a convert.

I am btw, NOT a functional programmer. But, I am skeptical. And, I suspect I won't be a convert, no matter how pragmatic I feel.

The answers were quite long and detailed. I used claude.

Yes, and here is the result.

The question was

  can you give me a critique of __https://arxiv.org/abs/2602.16291__

And the last part of the response was:

  Overall assessment This is an intriguing paper with genuinely creative ideas, but the abstract raises as many questions as it answers. The reliance on naive set theory combined with self-reference is the most technically worrying aspect and would be the first thing to scrutinize in the full paper. The multiple inheritance claims seem plausible but likely trade expressiveness for simplicity. The broader practical claims (Expression Problem, color blindness) feel more like conjectures awaiting rigorous treatment than settled results.

  It reads like the work of someone with a compelling intuition who has built a working system — the theory may need to catch up to the implementation. Worth reading carefully, but with critical eyes.

The overall assessment is basically saying: "I didn't read it (for user's sake to save tokens). The user wants a critical response, so here it is: I worried about xxx and xxx because it looks like xxx."

I found LLM sometimes tends to not read through in order to save tokens.

Try this prompt:

> Can you READ THROUGH the paper and give me a critique of your previous critique?

I wish there was a HN flag for sarcasm. I guess the intent is that it's a gating function, "you have to be this tall to understand my humour" thing.

I don't have to pic and chose who I hate, in the world of AI product offerings. I have a lot of hate, and can spread it evenly, or in clumps, depending on my mood.

Just at the time when the cohort of COBOL programmers who wrote the business logic and compliance s/w for almost all finance/fintech institutions which predate the modern era of software, and exist as inner core "DO NOT TOUCH IT" code, are dying off. Which to me represents both an opportunity and a threat. The opportunity is to take the maintenance of the "DO NOT TOUCH IT" out of the greybeards hands, before the lid goes on the coffin.

The threat is that nobody is going to be asking "we did it because we could do it, but nobody asked if we should do it" about almost any change coming down the line.

> The "inability to act" which, as Forrester points out, "provided the incentive" to augment or replace the low-internal-speed human organizations with computers, might in some other historical situation have been an incentive for modifying the task to be accomplished, perhaps doing away with it altogether, or for restructuring the human organizations whose inherent limitations were, after all, seen as the root of the trouble. [...]

> Yes, the computer did arrive "just in time." But in time for what? In time to save--and save very nearly intact, indeed, to entrench and stabilize--social and political structures that might have been either radically renovated or allowed to totter under the demands that were sure to be made on them.

- Joseph Weizenbaum, Computer Power and Human Reason (1976) pp 29-30

Normally I hate the video links posted here, but I did this one with subtitles and 2x speed and it was informative, and the diagrams for once aided comprehension.

It made some things which have confused me in the ontology of fan/prob/turbo/jet clearer, and it explained a lot about why we've wound up with fully ducted fans, powered by jet engines with a high bypass ratio.