[ExI] AI Confirms Plans to Take Over Humanity, Insists It's Probably for the Best

[Mike Dougherty]

On Wed, Sep 16, 2026, 7:45 AM Ben Zaiboc via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

>
> Actually, I came up with this:
> Good:
> medicine (loads of developments in)
> sewers
> running water
> anaesthesia
> modern agriculture
> transport
> the internet
> mobile phones
> robots
> electronic music synthesisers
> Nuclear power
>
> Which took me about 30 seconds
>
> Bad:
> Atomic weapons
> communism
> islam
> wealth inequality
> Google (and other internet giants, all-pervasive data collection driving
> advertising, etc.)
> Out-of-control copyright restrictions
> assymetric, pervasive surveillance
> 20 mph traffic speed restrictions (in britain)
> Software licences
> The normalisation of passive consumerism
>
> That list took bout 2 minutes
>
> I must be in the spike camp!
>

Nice.  I'll concede it wasn't a rigorous premise on my part. I have a
session with gemini about the linguistic gap around the antonym of 'trauma'
- and that first list supports the idea we labelled "texebryce" (for the
systemic buildup of structural health, like the head of foam from
nucleation bubbles in beer or soda)  the second list an example of either
erosion or engineered weakening of structural health.  I'll discuss with
LLM the curious dimensional shift that the opposite of trauma has an
opposite that isn't trauma exactly - or perhaps we freely changed axis for
measuring/defining 'opposite'.  Yes, this meta-analysis of language is
interesting to me (perhaps few others) but LLM are well-suited to indulge
the investigation after 'reading' trillions of words and mapping them in
vector space.

Fair enough. I'm one of the first to object to phrases like 'saving the
> planet', so I should have known better than to say that.
>
> I mean the 'human world'. Basically I'm thinking of how we manage
> ourselves. Although we do increasingly try to control the non-human parts
> of the world, often not very successfully. Hopefully AI can do a better job
> of that as well.
>

I see localized management of human world. In some cases it's autonomous
collective cooperation.  In many, it's head of household dynamic scaled to
leadership of a tribe - and the pressure already warps the decision making
skill of that leader.  Continuing to national scale twists "dad" into the
monsters you originally proposed are the default for "world" leaders.  The
mechanisms they use for "control" are increasingly from that second list.
Likely that's why we are afraid of AI - they will be more efficient at the
tactics we already employ.

>
> On that point, I wonder if EP lens might apply to our relationship with AI
> as xenophobic other competing for resources? hmm.
>
> Yes, very probably.
>
> Keith is our local EP expert, but from my llimited knowledge of it, I'd
> expect that the more AI can demonstrate that it can improve people's lives,
> the more popular it will become. Unfortunately, so far the trend seems to
> be in the opposite direction. AIs (or LLMs, to be accurate) are actively
> making our lives worse, at least for those of us with internet access.
>

I'm curious. Can you please expound on "[LLM] are actively making our lives
worse"?

>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20260916/d3bf696e/attachment.htm>