This seems like the most likely explanation.

The “memory palace” (AKA method of loci) inspiration is a plausible source for someone with a non-technical background, and there’s evidence that it’s closer to how the brain actually indexes memories natively.

(Although my understanding is that it bootstraps the hippocampus’s hard-wired ability to remember the layout of physical locations—I don’t know that an LLM would have a similar ability out of the box.)

A post intended to detract from the general context it’s part of.

The second law of thermodynamics.

Illogical fears require illogical solutions!

If you couldn’t understand anything unless you were literally standing underneath it.

In order to count anything, you have to apply some kind of equivalence relation to the world to decide which things belong to the category of things you’re counting—so there’s some degree of subjectivity involved, and you’ll most likely divide the world up into things whose quantities are manageable. If you’re counting the number of times something happens, and the number gets too big to keep track of, you need to be more selective about what qualifies as a countable event. And it’s been established that the point at which numbers get unwieldy is around 42.

That night when joy began
Our narrowest veins to flush,
We waited for the flash
Of morning’s levelled gun.

But morning let us pass,
And day by day relief
Outgrows his nervous laugh,
Grown credulous of peace,

As mile by mile is seen
No trespasser’s reproach,
And love’s best glasses reach
No fields but are his own.

—W. H. Auden

The title suggests that the government is pushing OpenClaw specifically, but the text says the opposite:

Yet as more ordinary Chinese get hooked, the government is pulling back. Chinese authorities have stepped up warnings of security and data risks and instructed government agencies and companies in sensitive sectors such as banking to curb OpenClaw’s use.

Yeah—but in theory you only need to train once, while inference costs are ongoing and scale up with usage.

I guess it’s ultimately a business decision by AI companies to weigh how often retraining is worth the cost.

TurboQuant, meanwhile, could lead to efficiency gains and systems that require less memory during inference. But it wouldn’t necessarily solve the wider RAM shortages driven by AI, given that it only targets inference memory, not training — the latter of which continues to require massive amounts of RAM.

I didn’t realize the RAM shortage was mostly due to training—I would have thought inference was at least a big a factor.

A burrito isn’t a real burrito if it doesn’t come wrapped in foil.

Same with tamales and corn husks.

Just a reflexive squiggle with the appropriate number of ascenders and descenders.

One attribute you want in a signature is that no one would be able to fake it from samples of your other handwriting—and for that, not using cursive normally is a plus.

It was “Ea-nasir”, until that one guy ruined the name for the rest of us.

Biodiversity.

A puzzle with one missing piece.

Having a cordwainer construct shoes directly on your feet without laces or other openings, then cutting the shoes off and discarding them after wearing them once.

4′33″

• Velvet Underground and Nico
• Bringing It All Back Home (Bob Dylan)
• Kind of Blue (Miles Davis)

They’re teleprompters crossed with Pepper’s ghost.

Humanity? Sure.

Human institutions? Less and less.