/LMAR=0/XLINE=3/FONT#0=BASL30/FONT#1=BASI30/FONT#2=BASB30/FONT#10=BAXM30/FONT#11=ZERO30/FONT#3=STA200/FONT#4=NGB25
␈↓ ↓H␈↓␈↓βS␈↓∧ Artificial Intelligence Laboratory, STANFORD UNIVERSITY, Stanford, California 94305

␈↓ ↓H␈↓∧Telephone 415 497-4430␈↓ 
⊗June 12, 1979 




␈↓ ↓H␈↓Thomas Hoover
␈↓ ↓H␈↓109 Waverly Place
␈↓ ↓H␈↓New York, New York 10011

␈↓ ↓H␈↓Dear Mr. Hoover:

␈↓ ↓H␈↓        You␈α∞would␈α∂have␈α∞to␈α∂pay␈α∞me␈α∞money␈α∂to␈α∞get␈α∂me␈α∞to␈α∂read␈α∞an␈α∞article␈α∂at␈α∞this␈α∂time␈α∞of␈α∂the␈α∞year.
␈↓ ↓H␈↓However,␈α
I␈α∞did␈α
≡nd␈α∞the␈α
reference␈α∞to␈α
my␈α∞own␈α
opinions,␈α∞and␈α
it␈α∞is␈α
correct,␈α∞except␈α
possibly␈α∞for␈α
the
␈↓ ↓H␈↓word␈α
"concede".␈α
 I␈α
have␈α
always␈α
felt␈α
that␈α
AI␈α
had␈α
a␈α
long␈α
way␈α
to␈α
go␈α
to␈α
achieve␈α
interesting␈α
general
␈↓ ↓H␈↓intelligence, and recent achievements have been insu≠cient to change this opinion.

␈↓ ↓H␈↓        One␈αcomment␈αon␈αthe␈αlast␈αpage.␈α I␈αagree␈αthat␈αif␈αhuman␈αintelligence␈αcan␈αbe␈αachieved,␈αthere␈αis
␈↓ ↓H␈↓no␈αbarrier␈αto␈αgoing␈αbeyond␈α
it.␈α However,␈αthe␈αmotivational␈αstructure␈α
of␈αa␈αcomputer␈αprogram␈αis␈α
just
␈↓ ↓H␈↓what␈α∞it␈α
is␈α∞programmed␈α
to␈α∞have.␈α∞ Therefore,␈α
it␈α∞is␈α
misleading␈α∞to␈α∞ask␈α
␈↓↓"would␈α∞it␈α
also␈α∞then␈α∞choose␈α
to
␈↓ ↓H␈↓↓make␈α⊃itself␈α⊂increasingly␈α⊃smarter␈α⊂..."␈↓.␈α⊃ It␈α⊃would␈α⊂do␈α⊃so␈α⊂if␈α⊃it␈α⊃was␈α⊂so␈α⊃programmed␈α⊂or␈α⊃if␈α⊃that␈α⊂was
␈↓ ↓H␈↓generated␈α
as␈αa␈α
subgoal␈αof␈α
some␈αother␈α
goal.␈α
 Naturally␈αpeople␈α
would␈αbe␈α
inclined␈αto␈α
program␈α
it␈αto
␈↓ ↓H␈↓make itself smarter.


␈↓ ↓H␈↓Sincerely,



␈↓ ↓H␈↓John McCarthy
␈↓ ↓H␈↓Director
␈↓ ↓H␈↓Professor of Computer Science