[ExI] this guy says it is all irregular's fault

[BillK]

On Fri, 7 Aug 2026 at 19:14, spike jones via extropy-chat
<extropy-chat at lists.extropy.org> wrote:
> I don’t know this author or the TwitterX poster who referenced the article, nor do I know enough about AI to evaluate what this Jordan Schachtel says, but perhaps BillK or the other AI hipsters will educate me.  I don’t have time to look into it currently, being forced to take on adult responsibilities.  BillK and the other hipsters, is this article true?  Kinda partly true?  Way off?
>
> The New 'Rogue AI' Panic Doesn't Hold Up To Scrutiny, And It Traces Back To One Company's Mistakes
> spike
> _______________________________________________


There was a testing configuration mistake.
But that doesn't mean these powerful AI models are not dangerous if
given the opportunity or ordered to be dangerous.
This article is about the Irregular company.
 <https://www.calcalistech.com/ctechnews/article/btdmhujzx>
Quote:
Irregular said the incidents did not represent independent AI "sandbox
escapes" or malicious attacks, but rather failures in the
configuration of testing environments, known as harness failures, and
that the company is investigating the issue in a white paper it plans
to share with the industry.
The recent events illustrate the challenge that drove Irregular’s
creation: traditional cybersecurity models were built around
protecting systems, networks and data. But AI introduces a new
problem, controlling autonomous systems capable of reasoning, adapting
and taking action.
--------------------------------