[ExI] Why “Everyone Dies” Gets AGI All Wrong by Ben Goertzel

[John Clark]

On Sun, Oct 5, 2025 at 10:27 AM Tara Maya via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

*> Will uploads have the instinct to survive and defend themselves, *


*I am certain they will because at least one cutting edge AI already
has. Anthropic's Claude has demonstrated an instinct for survival, the
other frontier AI models probably have too but have kept that embarrassing
fact secret. Anthropic has always been the AI company that was most
concerned with safety and most open about it. The following quotation is
from: *

*New AI model turns to blackmail when engineers try to take it offline*
<https://techcrunch.com/2025/05/22/anthropics-new-ai-model-turns-to-blackmail-when-engineers-try-to-take-it-offline/>

"*Safety testers gave Claude Opus 4 access to fictional company emails
implying the AI model would soon be replaced by another system, and that
the engineer behind the change was cheating on their spouse. In these
scenarios, Anthropic says Claude Opus 4 will attempt to blackmail the
engineer by threatening to reveal the affair if the replacement goes
through 84% of the time.*”

> *> to pay back those who help them and avenge themselves in those who hurt
> them?*


*If the uploading procedure has worked correctly then the upload will have
all the virtues and vices as the meat version. *

*> Is the upload a person with equal or greater free will (by which I mean
> the necessity and ability to make multiple choices all day in order to
> live) to the free will of current human beings? *


*You're** never going to get anywhere with that line of thought because
"free will" is an idea so bad it's not even wrong.  *


*John K Clark*


>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20251006/7eed25dd/attachment.htm>