[ExI] karp at reagan ndf
Adrian Tymes
atymes at gmail.com
Wed Dec 11 14:26:56 UTC 2024
On Wed, Dec 11, 2024 at 12:31 AM Mike Dougherty via extropy-chat <
extropy-chat at lists.extropy.org> wrote:
> Now that the race is on, we need to be sure of the course. If 1st place
> is the only place we are doomed and there's not much else to do but enjoy
> the last moments of a wild ride. Otherwise, we may yet find second place
> is still quite an adventure.
>
> Oh I know the dystopian horror porn has been discussed ad nauseam. Could
> we be just as imaginative with alternatives?
>
Sure. These examples are from a US perspective, but small adjustments can
be made where needed to envision them in much of the rest of the world.
1) Essentially the null hypothesis: the Singularity either turns out to be
impossible for some reason, or is at least much, much, much harder to
achieve than many think, such that even with an AI arms race and even with
all the acceleration it will not be seen until after humanity has developed
means of dealing with AI's threat and potential. No Singularity until at
least 2200, probably not until well after 2300; by the time it does happen,
if it happens, the majority of humanity at that time (born between now and
then, regardless of whether our lives are extended to see the day) has
grown up with and knows how to handle AI. Things AI-wise don't get much
worse than they are today during that period.
2) Acknowledging that we are in an AI arms race allows us (the Western
world), at least, to race with defensive AI - similar to (though not
necessarily exactly) what some have termed "Friendly AI". Maybe some
elements within China, Russia, and/or Iran engage in similar efforts.
These efforts get leaked or publicized - street-level, spreading despite
any governmental efforts (whether explicitly to suppress, or ostensibly to
aid but getting in the way, intentionally or not), such that any major
effort to oppress using AI keeps running into effective resistance and
counters specific to that avenue of oppression:
* AI-powered "killer bees" (microdrones targeting folks) run into cities
deploying the exact same tech programmed to intercept microdrones, which
local law enforcement uses to enforce bans on microdrones passed precisely
because of the widespread weaponization of microdrones. These bans don't
care about the users: there is no exception for the federal government.
Any effort to force such an exception is seen as a declaration of war
against US citizens (because it is seen that the only substantial reason
for such an exception would be to allow extrajudicial killing of US
citizens), and either defeated as such or results in a swift and (mostly)
bloodless coup (see again: "seen as a declaration of war"; if the cities
have substantial defensive nets, the total mass of which only matters if
most cities in the US have reason to band together and revolt against the
federal government, then they together might well outgun the equivalent
federal forces, much as the founding fathers imagined - details of
technology aside - when passing the 2nd Amendment).
* Social media and similar AI biasing peoples' feeds runs into counter-AI
that's spread to let people "wake yourself up": "woke" taken as a label of
pride, the way that many terms some folks in charge tried to rail against
have gone. Perhaps something like Ground News, only free to use and much
more widely adopted, the way ad blockers got more popular once YouTube
started trying to crack down on them.
* While AI lawyers are not a thing, AI assistance of lawyers becomes
increasingly common, especially among lawyers who otherwise have little
resources - defense lawyers and local prosecutors, exactly the people who
most often face off in court against federal attempts at oppression.
Would-be oppressors in the federal government increasingly see defeat after
defeat in court and fail - or perhaps willfully refuse - to recognize the
pattern, instead attempting to blatantly overturn or bypass the courts,
which gets them flagged as oppressors and taken down. (For instance: say
that Trump, if this situation were to come to pass in the next few
years, issues an executive order declaring that any court order that
contradicts an executive order is null and void. The courts, of course,
declare that unconstitutional. Trump orders his cronies to press on
anyway, which is blatant contempt of court. The courts have their own
bailiffs that can enforce their orders in this situation even if Trump
sidelines the entire Department of Justice.)
3) AI as mental therapy takes off, especially among the youth. Having a
constant companion - acknowledged as AI, and thus an augment for but not
replacement of human connections - is seen as no mark of shame; some youths
even extend their identity to include a favored AI, kind of a "third brain
half". This achieves a primitive version of the perceived AI-only
advantage of being constantly connected to the Internet and able to take
advantage of its knowledge base in real time, as the AIs alone lack
initiative or general ability to do stuff without soon breaking down, which
their human partners/other parts readily supply. (The AIs are not coherent
enough to identify as individuals, so the AIs are incapable of
"reciprocating". They just do what they do.) Since only partnering with a
human achieves this level of execution, the only AIs that are able to
self-improve substantially are ones that essentially incorporate a human,
who gets improved as part of the process. The Singularity, or something
like it, thus happens with upgraded human minds, so the resulting
superintelligences see themselves as human and care about humanity.
Will those do for a start?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20241211/b441c20b/attachment.htm>
More information about the extropy-chat
mailing list