perm filename M.PL[1,VDS]1 blob sn#180333 filedate 1975-10-05 generic text, type C, neo UTF8
C00001 00001
C00015 ENDMK
\CPALO ALTO, CA. 94306
\F4\←L\-R\/'7;\+R\→.\→S   Telephone:
\F1\COctober 6,1975

Professor Marvin Minsky
Artificial Intelligence Laboratory
545 Technology Square
Cambridge, Mass. 02139

Dear Marvin:

	This past month has  been one of new highs, new  lows and new
concerns in  my life and the life of Vicarm.   I feel that you should
share in some  of these  developments and  advise me on  a course  of

	As you know, I started Vicarm  a bit over two years ago, with
the   purpose  of  producing  computer  controlled  manipulators  and
accessories.  Since that time, I have spent countless hours and large
sums of my own money  to attain these goals.  In  the meantime, I have given
up my staff position at Stanford  and have  chosen to try  for a  PhD
instead, with the hope  that I could follow both  paths in parallel. 
So  far, the  best I can  say of  this dream  is that it  hasn't been
proved wrong, but I have not  yet accomplished what I started out  to

	The recent highs in my life have been the reports I have been
getting in from some of my customers.  Both the M.I.T. model arm (the
one I designed  at Tech) and  the Stanford model  arm are being  very
favorably recieved.  The  few customers, other than M.I.T., have been
more than satisfied with what they got, and it looks like I may  even
be  getting  some  repeat  orders.    At  the  recent  symposium  on
industrial  robotics  in Chicago,  my  M.I.T. arm  demo  (just manual
control) created a favorable stir too. 

	On the other hand my  disappointments have been that to  this
date I have  been able to deliver only three  M.I.T. arms besides the
ones you have, and even the more well developed Stanford arm takes me
twice as long to  make as I expected.   To add to  my disappointments
has been  my inability to  divorce myself from  arm manufacturing and
concentrate on my PhD and future  designs and developments.  To  this
end, I have hired several new people with promising potential. 

	As you know, I  have hired David Silver as  a Vicarm employee
to complement  the growing Vicarm staff (now  4 people).  David's job
offer came as a  result of his approaching  me about wanting to  work
for Vicarm, and  my briefly checking out the opinions  of some of the
people he has worked with in the past.
Your letter of recommendation was well appreciated.  Dave has been with Vicarm for
about a month now.  And instead of helping ease  my delivery woes and
creating  a  new  company capability,  I  find  my  concern over  his
suitability growing daily.  Dave is a very energetic and alert person
with a very  strong self image as  a reasercher and developer  of new
ideas and gadgets.   However, I feel that this image has been created
as a result of spending the  last ten years at a university  research
laboratory, rather  than out in  the competitive and trying  world of
industry  and small  business.   Even his outside  consulting efforts
have apparantly been under the protective tutelage of this father. 
	As I  expected, Dave has  proved a  difficult person to  work
with, but not  as expected I find I am  spending an inordinate amount
of time  directing and  organizing  his efforts  at  Vicarm.   It  is
getting painfully obvious to  me that Dave is not a  good team effort
type  of person,  but more  of an  individual advanced  hacker type.  
Vicarm's immediate needs (at  least for the  next few months are  for
the group effort type of person)  rather than the hacker, and as such
I have grown concerned over the advisability of my decision to employ

	Marvin, I am  sure that you are  well aware of Dave's  assets
and his faults, having seen him grow from a high school dropout to an
A.I. out growth.  You are certainly  aware of the  effects of such  a
childhood on the development of the complete  person, and as such are
in  a very  good position  to advise  me  on my  near term  course of

	Vicarm has two major orders outstanding at the present time. 
Both are  on a fixed price-  pay after acceptance- basis.   A Stanford
arm  to General Motors and  an M.I.T.arm to  the Naval Research Labs.
The GM order has been in  hand for 8 months now, and it is  already 2
months late.   The NRL order is  not yet due.  I  currently have Dave
working on the electronics for the GM order.  This involves a bit  of
design,  some  documentation, and  component   and  card  layout  and
assembly.  As many details are sketchy, understanding of the complete
system is necessary, and the ability to organize the  various details
into  a reliable,  well  documented operating  product is  necessary.
Delivery  yesterday  is  called for  on this  item.  Our  next project
calls  for an  LSI-11  based  system,  with delivery  of  the  LSI-11
scheduled  for December.   I  had planned  to have  Dave  handle that
entire  project,  which   calls  for  both   hardware  and   software
development of a system for the NRL by next April.  He seems eager to
do  this, but I  have mixed feelings  on his  ability to successfully
complete  the  project  and  give  us  something  which  is  properly
documented  and fully  compatible  with hardware  and software  being
developed  at Tech and other places working  in parallel.  Dave wants
to do it all alone, and is resistant to the ideas of compatibility. I
have explained to Dave that I had planned to buy software consulting
(I know several ex-A.I. types who are more than eager to work out  an
arrangement of sorts) to handle the difficult parts in an elegant and
efficient manner, but he  seems to feel that he can do it all just as
well.  I  worry that I may  just end up  with an expensive  prototype
kludge suitable  only for  the Navy  and A.I.  labs., rather  than an
elegant, understandable, piece of well planned out hardware and code.  

	Dave  has repeatedly mentioned his plans  to do something for
M.I.T. through Vicarm.  Recently, I talked with Gordon  Oro (we still
haven't settled our account with  you), and he mentioned that part of
your current budget  request, as yet  unapproved, included some  work
with Vicarm.   Dave says that this  is probably for another  arm like
his, or  else it has something to do with  his ideas for an automated
assembly and inspection system.  I  feel that Dave would be eager  to
do something like this for you through Vicarm, and as such I would want
to  encourage an  arrangement by  offering to  do this  at cost.   By
letting Dave work  on this independently I  would be saved the  daily
(actually almost hourly) hassle of directing and interfacing with him
and we both  could benefit, and  with it the  A.I. community, as  the
work would be public as part of a  government contract.  In addition,
my current agony and concern would be eased and I could pursue my PhD
with a clear consience that Vicarm personnel  were all operating  in
best application and best effort mode.  

	There is  a certain urgency to  this letter.  I  have delayed
long  enough on doing my PhD.  I  fear that any furthur delay will be
fatal to its  successful execution.   To that end,  I must feel  that
Vicarm can  continue to  survive without my  daily attention.   I now
have total confidence in  everyone at Vicarm execpt David.  I must
make a decision in  the very near future with respect  to David.  So,
could you give me some thruthful answers to these pressing questions.
What do you feel David  should be doing NOW,  and in 4 months?   What
are the chances that M.I.T. will be dealing with Vicarm in the future
and WHEN?  Would A.I. be willing to take David Silver back should the
current arrangement prove too
difficult for me to handle?   Please let
me know your thoughts soon as  I must make a decision.  Vicarm is too
frail to  tolerate too  many unknowns,  risks and  uncertainties  for

	Should you care to talk rather than  write, please call me at
home- 415-493-3038 mornings 7-8AM or eves. 7-10 pm (my time).  



Yours sincerely,

Victor Scheinman