[ExI] it won't work, but i have a better idea

[Adrian Tymes]

On Sun, Oct 12, 2025 at 4:54 PM <spike at rainier66.com> wrote:
> Something you said has me thinking overtime.  Elon loves wacky ideas.  He has a memetic virtual barrier around himself to block ideas from his attention.  I have been thinking of Elon memetic virtual barrier piercing technology, designed to somehow reach his attention with kinda fun wacky ideas like feeding shrimp to maturity on whale-chips and that kinda silliness.

That idea might have legs, if generalized.  Elon is far from the only
person with such a barrier.  How, in general, can good ideas be
brought to the attention of those with means and interest to fund
them?  I know the DOD in particular has quite a few efforts trying to
figure out how it can become aware of better ways to do what it does.

Consider, for example, its counter-drone strategy.  A recently
published tactical manual suggested a tank squad should, upon
detection of a hostile drone, stop, have its commander pop the hatch
and hand signal the other tanks, and then they line up their cannons
to shoot it.  This was derided by Ukranians who pointed out that the
tanks would not have nearly enough reaction time for this, the drones
would often be at higher angles than the tanks could pivot their
cannons to, and that the hatch-popped commander would practically be
inviting the drone to slip through the hatch and attack the tank from
inside.  The DOD knows it needs to do much better than this, but
sufficiently lacks the ability to get good ideas to those who need it
that this proposal to counter drones got all the way to official
publication.

There are many efforts that amount to, "they should just AI their way
to a clue".  That requires that:
1) the decision makers be aware that there is this sort of problem
(which itself is often an issue),
2) they be willing to accept solutions that are, if necessary, wildly
different from their current understandings (pride can be a barrier),
3) they know generally how to phrase the problem in ways that will
prompt usable solutions, and
4) often times, they get buy-in from others with even less clue than
the alleged decision makers, which requires understanding why the new
solutions work better than the old ones (see "explainable AI").

This problem set does not appear to be purely solvable with "just AI".
Nor will a focus on just the parts that AI by itself can solve, be
useful at all at this point - not with how much focus other people are
already spending on said "just AI" portions.

Can you solve the remaining portions?  If so, find the DOD
solicitations looking for this, write up and submit a good answer, and
you could make yourself - maybe not bajillionaire-grade wealthy, but
quite a bit more financially comfortable, without much physical
exertion.