Granted I have more memory than can fit in currently-practical LLM context windows, but RAG mostly solves that. When an AI is thinking about math, it can have relevant math memories in context without needing all the other stuff.

[1] https://www.forbes.com/sites/markfaithfull/2026/05/07/heres-...

An AGI test should be black-box; we shouldn't impose require requirements on internal components. As long as the overall AI is capable of learning and remembering things, it shouldn't matter if there's a stateless LLM internally.

I think this is the real reason, while the others don't make that much sense when we really think about it.

Large datacenters are normally limited to industrial zoned land, so they're "competing" for space with other industrial projects, and their externalities are typically mild compared to other industrial projects.

Most datacenters create little traffic, minimal pollution, and no bad odors. They do use substantial power and water, but they pay for it (though in certain cases the economics aren't perfect).

I'd much rather live near a datacenter than a dump, sewage facility, refinery, etc. etc.

They may be noble causes, but it's not how many Wikipedia donors expected their funds to be used, after seeing the donation banners begging for funds to keep Wikipedia running.

https://meta.wikimedia.org/wiki/Knowledge_Equity_Fund

> I'd rather the outlet covers be original and clean than the new, usually oversized covers (with a slight color mismatch) I often find.

Any task can be screwed up if someone is careless enough, but this is extremely easy to get right. If a standard cover plate fit before, it's trivial to buy a replacement of the standard size and appropriate color.

Cleaning works too, I personally just wouldn't bother cleaning a part that costs ~$0.50.

> When I see sloppy caulk I first think of the time I'll need to spend to dig it all out.

Would you really rather see cracked caulk, which means (a) the original contractor probably used whatever cheap rigid caulk he found in his van, (b) it's been letting water through instead of sealing, and (c) noone cared enough to fix it?

In Vitalik's words: "If someone puts a gun to my head and tells me to write a hard-fork patch, I will definitely write it. If I publish a patch to delete a bunch of accounts, how many people here would download and install the update and switch to that chain? I see relatively few raised hands. This is called decentralization"

- For better or worse, Ethereum has more centralized leadership (mainly the EF) which helps coordinate upgrades. Hard forks happen pretty regularly, and don't typically risk fracturing the community, unless the changes are particularly controversial (see Ethereum Classic).

- PoS wasn't that controversial, since it was planned from the beginning (see the difficulty bomb). Some users still resisted the switch - see the EthereumPoW fork - but for the most part the community was on board.

- The Bitcoin community OTOH values ossification. It rarely accepts proposed hard forks as "still Bitcoin"; that has only happened a few times in its history. The community's response to most proposed hard forks is "interesting, but that's not Bitcoin, call it something else."

- The Bitcoin community is also more ideologically and economically invested in PoW, so a proposed PoS fork would be extremely controversial, and would fracture the community (at best, if it gained any traction at all).

Autoclaved aerated concrete is worth a mention as well - it's also a decent insulator and somewhat environmentally friendly, while additionally being, light, easy to cut, and somewhat usable as a structural element.

There's no novel problem here that any particular industry would need to find novel solutions for. Just some states ought to tweak their rules about the costs of distribution infrastructure upgrades.