Emotional attachment isn't science

All of these variants of the Riddle of Epicurus presume *some* constraint that the God figure is acting under.

I think this is how AI and tech ought to go. I don't yet know if it will. I hope it does.

(am one of the signatories)

Straussian read on Marc is that he no longer cares if the narratives he and A16Z put out are grounded, just if they seed a market for his portcos.

A true OG founder mode was Henry VIII and the founding of the english church.

Why do many who talk of AI futures seem to be thinking religiously?

Smuggling infinities into your arguments always leads to visions of the infinite.

"Eunuchs" is too identity-limiting. Please use "non-huevos-having persons"

I also just love the word "abliterated" in LLMs. Playing on "ablations" (turning off or removing parts of a model) and "obliterated", yeah?

I would have also accepted "ablotomized"

TV gave us "a face for radio", podcasts gave us "a voice for substack"

Anyway I still just write a lot of text dunno what this says about me.

both. Eg retention and new user accounts for a social media app as a classic example

I wish there were comparative studies for cash-aid vs Hong Kong or Singapore model.

Cash gives resources, lets them figure it out. Institution building says "get the rules right, resources follow".

Best approach likely depends. A corrupt resource rich country needs LKY. But not everyone...

Friendly (somewhat serious) reminder that nihilism is a waiting area, an intermediate step along your life

You ever focus so hard on a single chunk of deep mental work that after like 6 straight hours your brain almost feels... like swollen and heavy?

yes metrics can fail in other ways (e.g. McNamara fallacy where you don't value things that aren't measured)

Only said that the gaming of metrics Goodhart's rightly predicts largely mitigated by picking another metric that shows if you're gaming the first.

This is a map of all states with an economy smaller than just the SF Bay Area. Consider all the $ generated here.

Now consider the flat rate of housing construction.

Supply/Demand is so annoyingly powerful.

Goodhart's Law is a problem, but it's an easy solve if your culture can balance 2+ metrics that counterbalance.

The hard part is that complex organizations spend so much time with infighting and coordination that they often can't manage more than one.

There’s a threshold where your company is so complex and political and opaque that there’s more signal from an outside journalist than in internal comms.

Knowing that I could turn down any side street and discover layers of human creativity: a cafe, a mural, people, an awning, and so much more

As @kasey.bsky.social put it: “going outside in NYC is like pulling a slot machine”

In tech parlance: like everyone would be running Linux just different distros

Trading blocks themselves have this — each country has its own tweaks and customizations, much as I assume the firewalls would.

So: suspect we agree

That’s not what this thread is about at all and why you are now blocked.

I suspect the boundaries will in effect be like trading blocks: groups of 1 or more countries with a shared set of values. Trading blocks are shared economic values, firewalls are informational values

This is basically the level of understanding from the "nooooo Abundance gets housing wrong" left objections

"not in a while" is just 👩🏼‍🍳 💋

I’m on the quadrant that thinks (Gary Markus is mostly right) and (we have 1-2 decades of products to figure out on these LLMs anyway)

Excited for this one. Had to preorder like the appropriately-grouped pleb that I am!

(Hits blunt) has the scientific method been around long enough to have embedded cultural wisdom?

technically this is an accurate description of the process. Ask the question, no matter the answer, deport.

yeah. Even the capability overhang of *last* year would be a 10-15 year buildout in products and businesses.

Add this to the “I asked specifically for a bad thing and a general purpose tech gave me what I asked for” hall of dunces