uncertainty
Remixes of the nature of uncertainty described by Taleb, Kahneman & Marks
@sophia-indrajaal
Holy cow, didnt know this channel existed, Uncertainty has become the most important vector in my explorations with LLMs! It's the space that creates the conditions for emergence!
1 reply
0 recast
3 reactions
@vt
i wrote a tiny booklet about different kinds of not-knowings, to accompany a beta of an installation i put up in adelaide a couple of weeks ago: https://vaughntan.org/nkmod
turns out, writing at 12yo reading level is hard when writing about things that aren't known.
0 reply
0 recast
1 reaction
@vt
black swans are often defined as improbable, rare, high-impact events. i'd say that any of the four categories of not-knowing can generate black swan events.
for instance, unexpectedly discovering a feasible way to control an atomic detonation is an action that generates an outcome (the possibility of unprecedentedly destructive detonations) which changes causal probabilities (e.g. suddenly changing the perceived strength of relationship between foreign policy strutting and the likelihood of existential hazard) and might result in an unexpected dramatic shift in the public sentiment and valuation of investment in national armament. each of these might be considered rare, improbable, high impact — all of them are black swan-like. but they emerge out of (and are produced by) not-knowing of, respectively, actions, outcomes, causation, and valuation.
(continued!)
1 reply
0 recast
4 reactions
@vt
I wrote about games that operationalise true uncertainty, not just simple risk, to help players become better at making decisions when facing different types of not-knowing.
https://vaughntan.org/uncertaintygames
here, too, is a photo of a late july afternoon at catalans, as algo-bait
2 replies
0 recast
2 reactions
@vt
in a meeting last week, a c-suite type person said "we've got to get a better operating and conceptual grasp of all these uncertainties" and then NOT EVEN A MINUTE LATER said "i want likelihoods for them included in our updated risk register by end of the next quarter."
🤦♂️
(true uncertainties cannot be quantified and must be dealt with differently: https://vaughntan.org/uncertaintyrisk)
0 reply
0 recast
0 reaction
@vt
I've been writing for over a decade about why we should be clear about the difference between "uncertainty" and "risk."
when we call a situation with unknowns "risky", we automatically apply risk-management ways of thinking when deciding how to act — cost-benefit analysis, expected value analysis etc.
these are good when the unknowns can be both precisely AND accurately estimated. most of the important unknowns today aren't accurately quantifiable like that.
and if you can't quantify the unknowns accurately these risk-management ways of thinking just don't work.
worse, they encourage the optimisation instinct, which is fragile when the situation is actually unquantifiably uncertain
geostrategic unknowns from highly capricious state actors are the definition of unquantifiable uncertainty (to take just one example 😑)
https://vaughntan.org/notknowing
0 reply
0 recast
0 reaction
0 reply
0 recast
1 reaction
@vt
workshop today! and it was good — will post later mechanism and prototyping ideas that came out.
but here's the group hard at work
0 reply
0 recast
2 reactions
@vt
prepping to run a workshop tomorrow to unpack the different types of non-risk not-knowing with mathematicians, artists, sociologists, museum exhibit creators, futurists, and other random types.
(in adelaide, lmk if you want to join 0930-1230h local time, at MOD in the CBD)
https://vaughntan.org/notknowing
1 reply
0 recast
5 reactions
@vt
i write lots about how an uncertainty mindset changes what we see and how we interpret it in ways that are useful when the situation we face is uncertain not risky.
having the wrong mindset leads to making the wrong decisions. these can have serious consequences.
https://www.newyorker.com/magazine/1995/05/29/the-fifty-nine-story-crisis-citicorp-center (this is a great story, well worth reading)
0 reply
0 recast
3 reactions