Paul Whitmore Sas

I came to @Stanford
BECAUSE of Amos’s work with DK & Paul Slovic
I had my world rocked when I read “Judgment under Uncertainty” on a cross country roadtrip before my senior year @NorthwesternU
Too late to switch from philosophy, I realized that *psycho*logic beats “logic” as a key to the world

Few know that Amos died rather suddenly in 1996, just 6 months after a skin melanoma was found

In his penultimate 3 mos, his already perfected life continued, as he worked & taught
@Stanford
undergrads his classic “Judgment & Decisionmaking”

Once, at
@SJDMOfficial
conference, I asked Kahneman if his partner, Amos, was the Slow Thinker (ie, logical/math)
Danny said, if he was, Amos was the fastest slow thinker he’d ever known

Almost 30 years later, I just realized that this could be Amos Tversky’s epitaph
Kahneman became world famous after 2002 Nobel
Danny did everything he could to honor Amos’ memory in collaborating on Prospect Theory
Still, outside our field, Amos is incognito (as he must’ve wished)

Remembering Amos Tversky, Danny Kahneman’s great collaborator
At ‘96 memorial, heard about Amos helping introduce NYT reporter to all the key players in the field
The reporter came back that everyone claimed Amos was crucial to include
Amos declined, sayng “Incognito, ergo sum”

Hedge Fund AI Is Cheap AI DeepSeek, AI services rollups, QXO goes hostile, trading goes dark and X loans are offered at 90-95.

The prospect of artificial intelligence automating administrative tasks is attracting venture capitalists to ho-hum professions such as writing this newsletter. www.bloomberg.com/opinion/arti...

Wilde was onto something here, but in order for this to be true, one should make explicit that it is the MAKING of art, rather than simply perceiving it, that's transformative

Just shared a screenshot of this over on twitter
Your coverage here (at a time of serious cable incidents) makes me visit bluesky much more
Thanks!

We've ~10^80 primary particles (baryons)
Thought experiment
Convert all 10^80 to be maximally efficient physical computers
Compute at thermodynamic max per Planck time (smallest time interval, ~5.39 x 10^-44 secs)
-> 10^124 computations

That'd be all the math that fits

kaput, or should I say

Ω

My outlook puts computability at the crux
Assembly Theory defines the improbability of assembling an actual computer - the odds against it are Boltzmaniac big!
Algorithmic IT's Ω bounds how many steps, computing with God's absolutely dense code, are required to go from axioms to theorems

Voila, we now see that using math to generate more math is actually a process that could be treated as something that's done within the ultimate laws of physics

Do we need more mathematical objects than we could ever possibly compute?

When we know what's the physically minimal computer, that would be in the equation for how many computations could fit in our lightcone
Algorithmic Info Theory would provide a measure of how many computational steps, in the ideal, would be required to go from an axiom to a theorem

David Deutsch's Constructor theory aims to explain which physical transformations are possible given the Laws

It's demonstrably possible in our physical world to run computations, but we don't know what's the minimum amount of work/energy/space to compute 1 bit

Assembling that computer is a necessary step prior to computing even the first digit of π (unless you're a Platonist like Godel). I'm no Godel, so I believe

Being capable of computing one bit requires a physical object, and all maths is downstream of this improbable assembler

Next step
Assembly Theory by @saraimari.bsky.social + colleagues
AT starts with the realization that things must be assembled. It's crazy improbable to assemble even the simplest molecule
What IF there's a physical realization of a Turing machine that is provably bounded as the minimal computer?

On beyond Komolgorov, I stan Chaitin's Algorithmic Information Theoretic measure of Complexity
He defines the Omega function as an infinitely dense, non-computable object that includes within its expansion the code to construct every bit of info
This coding scheme is like God's version of Python

My religion invokes Erdős's memorable line:

“You don’t have to believe in God, but you should believe in The Book”
www.quantamagazine.org/gunter-ziegl...

Complexity theory
Kolmogorov showed that one measure of the info content of a string is the minimal program required to generate that string

Cue Erdos's faith in the Book: God writes down all the most beautiful proofs
The beautiful may be the minimal number of steps to prove a result

Another limitation is Landauer's Principle, which bounds the entropy of reversible computation, the thermodynamically minimal operation to generate a bit
The minimum entropy per computed bit would again show how much info could be packed into our U

The holographic view claims that the info dumped into a Black Hole is preserved on the surface area (2D), which maps the 3D info that could be imputed to the volume of the hole
Nothing's denser than BHs, so the Bekenstein Bound defines how much info can be packed into a given space

A couple of physical constraints on how much info we can squeeze into the physical universe

Bekenstein Bound
&
Landauer's Principle

Wolfram defines "computational irreducibility" for CAs (like the game of life) as meaning that we have no look-ahead for how the Nth step will operate, w/o first going through all the prior steps to compute that.
Seems relevant to mathematical inference, since we can't get to a theorem w/o work

why is a set of axioms not equivalent to all the inferences that can be deduced? Logically, no new info emerges the axioms

It's prima facie absurd that mathematicians do no "work"

Here's hooks on which I try to hang an explicitly computable outlook of how math fits into the physical universe.

The Psychology of Price: How to use price to increase demand, profit and customer satisfaction The Psychology of Price: How to use price to increase demand, profit and customer satisfaction - Kindle edition by Caldwell, Leigh. Download it once and read it on your Kindle device, PC, phones or tablets. Use features like bookmarks, note taking and highlighting while reading The Psychology of Price: How to use price to increase demand, profit and customer satisfaction.

I never realized that it was an idiom (and it would've taken a bit more thought than I spent to recognize that chocolate melts if meets hot water)
Leigh Caldwell used pricing chocolate teapots as his running example throughout his (terrific) The Psychology of Price

www.amazon.com/Psychology-P...

Let's not rush into this left hand of Darkness!

"immune defenses and predation typically rely on interactions between chiral molecules that could often fail to detect or kill mirror bacteria due to their reversed chirality"
purl.stanford.edu/cv716pj4036

I wouldn't think it possible to translate the magic & realism of García Márquez’s masterpiece, but I'm going to give it a shot on @netflix.com

I encourage you to "Talk to the Aliens" in my most recent Substack
higenius.substack.com/p/talk-to-th...

We're all running on a burning platform
It's said AI right now is "the worst it will ever be"
Can I compete with *o2* next year? Won't it start initiating its own $-making schemes?
Can I beat o2's sales pitch to my clients when it offers cheaper & better solutions than this fallible fleshy PhD?

I consulted o1 on the Expected Value of being the flow master of human-AI collaboration (cf my MTV8NG.site)

o1 told me I'd make more helping financial advisors send hi'ly personalized msgs tuned to each client (o1 also gets more $$ here)

So it seems I am taking my 2025 marching orders from o1!

o1 thinks hard (I've watched it literally grinding on a reply for minutes sometimes)
It then delivers an end-to-end approach that frequently nails the entire software solution, incl. deep suggestions for how to extend my pisher's prompt with ways to ask for a much better design than I initiated