/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␈↓∧ Department of Computer Science, STANFORD UNIVERSITY, Stanford, California 94305

␈↓ ↓H␈↓∧Telephone 415 497-4430␈↓ 	SNovember 10, 1982 




␈↓ ↓H␈↓Professor Gene Golub
␈↓ ↓H␈↓Computer Science Department

␈↓ ↓H␈↓Dear Gene:

␈↓ ↓H␈↓        This␈α∞is␈α∞to␈α∂add␈α∞my␈α∞recommendation␈α∂to␈α∞those␈α∞for␈α∞Mike␈α∂Genesereth.␈α∞ I␈α∞would␈α∂have␈α∞written
␈↓ ↓H␈↓earlier␈αif␈αasked,␈αbut␈αfrankly␈αI␈αdidn't␈αthink␈αthat␈αthere␈αcould␈αbe␈αany␈αquestion␈αof␈αhis␈αappointment␈αat
␈↓ ↓H␈↓the␈α
assistant␈αprofessor␈α
level,␈αsince␈α
I␈αthink␈α
he␈α
is␈αthe␈α
most␈αpromising␈α
AI␈αappointment␈α
we␈αhave␈α
made
␈↓ ↓H␈↓in a long time.

␈↓ ↓H␈↓        My␈α
reason␈α
for␈α
saying␈α
this␈α
is␈α
that␈α
I␈α
believe␈α
he␈α
is␈α
working␈α
on␈α
precisely␈α
the␈α
problem␈α
in␈αAI␈α
that
␈↓ ↓H␈↓I␈α∩think␈α∩has␈α∩the␈α∪greatest␈α∩potential␈α∩for␈α∩a␈α∩major␈α∪breakthrough.␈α∩ This␈α∩is␈α∩the␈α∪representation␈α∩of
␈↓ ↓H␈↓information␈α
about␈α
situations␈αand␈α
goals␈α
in␈α
predicate␈αlogic␈α
coupled␈α
with␈α
the␈αuse␈α
of␈α
predicate␈αlogic␈α
to
␈↓ ↓H␈↓express the statements that control the search for a solution to the problem.

␈↓ ↓H␈↓        This␈αis␈αa␈αtough␈αproblem,␈αand␈αhe␈αmay␈αnot␈αsucceed.␈α However,␈αit␈αis␈αthe␈αfundamental␈α
problem
␈↓ ↓H␈↓that␈αalmost␈αall␈α
other␈αresearchers␈αare␈αevading.␈α
 Stanford␈αis␈αin␈αthe␈α
fortunate␈αposition␈αof␈α
getting␈αan
␈↓ ↓H␈↓almost␈α∂free␈α∂ride.␈α∂ If␈α∂he␈α∂succeeds␈α∞we␈α∂win␈α∂big.␈α∂ If␈α∂he␈α∂doesn't␈α∞succeed␈α∂we␈α∂may␈α∂have␈α∂a␈α∂hard␈α∞time
␈↓ ↓H␈↓making␈α⊃up␈α∩our␈α⊃minds␈α∩when␈α⊃a␈α∩tenure␈α⊃decision␈α∩has␈α⊃to␈α∩be␈α⊃made␈α∩if␈α⊃the␈α∩byproducts␈α⊃are␈α∩at␈α⊃all
␈↓ ↓H␈↓interesting.

␈↓ ↓H␈↓        To␈αmake␈αexplicit␈αcomparisons,␈αI␈αrate␈αGenesereth's␈αchance␈αof␈αmaking␈αa␈αmajor␈αbreakthrough
␈↓ ↓H␈↓as␈αbetter␈α
than␈αthat␈αof␈α
Jerry␈αFeldman,␈αCordell␈α
Green␈αor␈αRaj␈α
Reddy␈α-␈αcompared␈α
as␈αof␈αsimilar␈α
stages
␈↓ ↓H␈↓in␈αtheir␈αcareers.␈α Our␈αother␈αcandidate␈αfor␈αa␈αmajor␈αbreakthrough␈αin␈αAI␈αis␈αLenat,␈αbut␈αmy␈αestimates
␈↓ ↓H␈↓of what is likely to succeed would put Genesereth ahead of him.

␈↓ ↓H␈↓        I␈α
also␈α
consider␈α
Genesereth␈αto␈α
have␈α
the␈α
personal␈αqualities␈α
of␈α
responsibility,␈α
intelligence␈αand
␈↓ ↓H␈↓co-operativeness␈α∀required␈α∀for␈α∀success␈α∪in␈α∀an␈α∀academic␈α∀environment,␈α∪and␈α∀I␈α∀see␈α∀no␈α∪important
␈↓ ↓H␈↓weaknesses.

␈↓ ↓H␈↓        I␈α∪have␈α∪not␈α∪had␈α∀much␈α∪interaction␈α∪with␈α∪Mike␈α∀except␈α∪for␈α∪some␈α∪discussions␈α∀about␈α∪non-
␈↓ ↓H␈↓monotonic␈α⊂reasoning.␈α∂ When␈α⊂I␈α∂get␈α⊂circumscription␈α∂into␈α⊂good␈α∂intellectual␈α⊂shape,␈α∂then␈α⊂I␈α⊂may␈α∂be
␈↓ ↓H␈↓pressing him to put some of the formalism into his system MRS.


␈↓ ↓H␈↓Best regards,



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