<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">On 2016-08-19 18:25, William Flynn
Wallace wrote:<br>
</div>
<blockquote
cite="mid:CAO+xQEZGPWnS2_-SR+9nTB3_FL2mOH0fR9DGAZwd-CWj+ZSnAg@mail.gmail.com"
type="cite">
<div dir="ltr"><span
style="color:rgb(34,34,34);font-family:arial,sans-serif;font-size:12.8px;line-height:20px">Just
how can Anders group parse that? How do you put all that into
equations? Seems pretentious or maybe over-simplifying human
behavior. (now I could get an even more complicated response
from Anders that I will fail to follow in its entirely)</span></div>
</blockquote>
<br>
Mostly because our equations are about AI. The designers can build
AIs to have a particular utility function, and that will be it.
Humans are messy and might not be well described even by complex
utilities. But at least some AI can be analysed in great detail by
this formalism.<br>
<br>
(I'm actually more interested in neuromorphic systems that borrow
their structure and hence also messiness from biology; they will of
course at most have approximate utilities.)<br>
<br>
Note that lacking a utility function does not mean you are
irrational. It is just that you may have messy, probabilistic goals
that are not known entirely to yourself.<br>
<br>
<pre class="moz-signature" cols="72">--
Dr Anders Sandberg
Future of Humanity Institute
Oxford Martin School
Oxford University
</pre>
</body>
</html>