WEBVTT

00:06.100 --> 00:10.939
In between five and 20 years from now, there's a
good chance, a 50% chance,

00:11.219 --> 00:12.819
we'll get AI smarter than us.

00:13.000 --> 00:16.799
Oh, yeah. Yes. In fact, I have reduced my

00:17.319 --> 00:21.940
expectations for when we will get to AGI or
whatever, human level.

00:22.719 --> 00:24.239
Uh, I used to think it would be

00:24.840 --> 00:29.079
five to 20 years, and now I think it could be, you
know, two to 10 years t-

00:29.079 --> 00:33.979
Uh, and so as a result of that, I, I... Uh, you
can find it

00:33.979 --> 00:38.199
on my blog from about 2009 I think is the first
time I publicly talked about it.

00:38.619 --> 00:42.440
I predicted a 50% chance of, um, AGI by 2028.

00:43.180 --> 00:45.140
How did- I still, I still believe that today.

00:46.919 --> 00:50.840
But still, if you em- if you extrapolate this kind
of progress, we'll be in a very, very

00:50.959 --> 00:51.840
different place

00:52.700 --> 00:54.239
in five years, let alone 10 years.

00:54.239 --> 00:54.399
...

00:55.279 --> 00:59.380
has been added. If you just kind of, like, eyeball
the rate at which these capabilities

00:59.439 --> 01:00.239
are increasing,

01:01.020 --> 01:04.959
it does make you think that we'll get there by
2026 or 2027.

01:05.239 --> 01:10.059
In the essay, you actually say a really big thing,
which is

01:10.599 --> 01:15.379
ASI, superintelligence, is actually thousands of
days

01:15.419 --> 01:15.839
away.

01:15.860 --> 01:18.500
Maybe. I mean, that's our hope and our guess what-

01:18.500 --> 01:18.799
Yeah.

01:21.659 --> 01:24.459
Like, I mean, I think we're quite close to digital
superintelligence.

01:25.680 --> 01:26.720
It, it may happen

01:27.659 --> 01:30.739
this year, and if it doesn't happen this year,
next year for sure.