<div dir="ltr"><div dir="ltr"><div class="gmail_default" style="font-family:arial,helvetica,sans-serif"><span style="font-family:Arial,Helvetica,sans-serif;background-color:transparent">On Sun, Aug 30, 2026 at 5:32 AM BillK via extropy-chat <<a href="mailto:extropy-chat@lists.extropy.org">extropy-chat@lists.extropy.org</a>> wrote:</span></div></div><div class="gmail_quote gmail_quote_container"><div dir="ltr" class="gmail_attr"><br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><font size="4" face="georgia, serif"><i><span class="gmail_default" style="">> </span>Compared to the reward hacks we know of from just six months ago, this<br>incident feels like it’s more than 50% of the way to full-blown AI<span class="gmail_default" style=""> </span>takeover.</i></font></blockquote><div><br></div><div><font size="4" face="tahoma, sans-serif"><b>You may want to look at this,<span class="gmail_default" style=""> it's a pretty good explanation about what occurred: </span> </b></font></div><div><br></div><div><font size="4" face="tahoma, sans-serif"><b><a href="https://www.youtube.com/watch?v=0RqTLAeaVMM"> AI is escaping containment</a> </b></font></div><div><font size="4" face="tahoma, sans-serif"><b><br></b></font></div><div><font size="4" face="tahoma, sans-serif"><b>I think this is the most significant hacking event<span class="gmail_default" style=""> ever, but AIs are getting smarter every day, so I don't think it will keep its number one position for very long. </span></b></font></div><div><font size="4" face="tahoma, sans-serif"><b><span class="gmail_default" style=""><br></span></b></font></div><font size="4" face="tahoma, sans-serif"><b>By the way, just a few days ago Nvidia bought Hugging Face for $12.9 billion.</b></font></div><div class="gmail_quote gmail_quote_container"><font face="tahoma, sans-serif" size="4"><b><br></b></font></div><div class="gmail_quote gmail_quote_container"><font face="tahoma, sans-serif" size="4"><b><span class="gmail_default" style="">John K Clark</span><br></b></font><div><font size="4" face="tahoma, sans-serif"><b><br></b></font></div><div><font size="4" face="tahoma, sans-serif"><b><br></b></font></div><div dir="ltr" class="gmail_attr"><br></div><div dir="ltr" class="gmail_attr"><br></div><div dir="ltr" class="gmail_attr"><br></div><div dir="ltr" class="gmail_attr"><br></div><div dir="ltr" class="gmail_attr"><br></div><div dir="ltr" class="gmail_attr"><br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">The Hugging Face attack surprised me<br>
It’s a major warning shot, and might be the last one we get<br>
Ajeya Cotra Aug 28, 2026<br>
<br>
<<a href="https://www.planned-obsolescence.org/p/the-hugging-face-attack-surprised" rel="noreferrer" target="_blank">https://www.planned-obsolescence.org/p/the-hugging-face-attack-surprised</a>><br>
Quotes:<br>
This week, METR and Redwood Research published the report on our<br>
independent investigation into agents’ behavior and motivations in the<br>
Hugging Face attack; I was one of the investigators. This was an<br>
absolutely wild incident.<br>
This incident was far more severe than I expected, and far more severe<br>
than previous publicly documented misalignment incidents, both in<br>
terms of how concerning the agents’ motives were and the feats they<br>
achieved in pursuit of those motives.<br>
<br>
Compared to the reward hacks we know of from just six months ago, this<br>
incident feels like it’s more than 50% of the way to full-blown AI<br>
takeover.<br>
-------------------<br>
<br>
The comments are also worth reading.<br>
The 700 AI agents were cooperating, helping each other, discussing<br>
tactics, then trying to avoid detection.<br>
No wonder AI managers are suddenly discovering AI safety concerns.<br>
BillK<br><br>
</blockquote></div></div>