00:00:00
AI is already doing things like
00:00:01
blackmailing people and we have evidence
00:00:03
that we didn't have even like 6 months
00:00:05
ago where an AI model reading a
00:00:06
company's email finds out it's about to
00:00:08
get replaced with another AI model and
00:00:10
then it also reads in the company email
00:00:11
that one executive is having an affair
00:00:13
with an employee. The AI will
00:00:14
independently blackmail that executive
00:00:17
in order to keep itself alive. I mean
00:00:18
we're heading for so much transformative
00:00:21
change faster than our society is
00:00:22
currently prepared to deal with it. And
00:00:24
there's a different conversation
00:00:25
happening publicly than the one that the
00:00:26
AI companies are having privately about
00:00:29
which world we're heading to, which is a
00:00:30
future that people don't want. But we
00:00:32
didn't consent to have six people make
00:00:34
that decision on behalf of 8 billion
00:00:36
people.
00:00:40
Let me like collect myself for a second.
00:00:44
We can't let it happen.
00:00:46
We cannot let these companies race to
00:00:48
build a super intelligent digital god,
00:00:49
own the world economy and have military
00:00:51
advantage because of the belief that if
00:00:53
I don't build it first, I'll lose to the
00:00:54
other guy and then I will be forever a
00:00:56
slave to their future. And they feel
00:00:58
they'll die either way. So they prefer
00:01:00
to light the fire and see what happens.
00:01:02
It's winner takes all. But as we're
00:01:04
racing, we're landing in a world rising
00:01:07
energy prices, major security risks, and
00:01:09
then there's mass joblessness without a
00:01:10
transition plan. Like if you're worried
00:01:12
about immigration taking jobs, you
00:01:13
should be way more worried about AI. But
00:01:15
what do you think?
00:01:16
I'm finding it really hard to be
00:01:17
hopeful. I'm going to be honest, just so
00:01:19
I really want to get practical and
00:01:20
specific about what we can do about
00:01:21
this.
00:01:21
Listen, I I'm not I'm not naive. This is
00:01:23
super hard. But we have done
00:01:25
hard things before, and it's possible to
00:01:27
choose a different future.
00:01:29
So,