(Interview in The Sinclair Story, 1985)
I loved Question Time because you meet interesting people - it's great fun.
I asked Sinclair why he'd been chosen for this weekly television panel of punditry, run by benign, no-nonsense chairman Robin Day. I had thought that the balance was the three political parties and one industrialist Sinclair had been on the panel when the programme was broadcast from Scotland, with its Timex relevance.
No, they get two people thinking from the left and two from the right. And usually one of the two on each side is a known politician, and the other is less obviously political … they think l think from the right.
But in fact, one's opinions may vary from a set party line I think it such a pity that affiliation to a political party appears to tie one's thinking to a set of national party standards whereas so many decisions should be made at a more domestic, local level.
Exactly. You and I not being politicians are ill-suited to Question Time; I think it benefits the programme to have people who think for themselves, but the politicians are very vocal because they've thought out a stock answer for almost every subject, whereas l have to dunk up something on the spot because most subjects I haven't given any thought to.
[So sometimes Sinclair's answers sound a bit terse. But if you listen to a politician answering questions, you will often hear the automatic pilot being switched on, and the standard wording being fed out. Sometimes 'not answering the question' springs not from a desire to be evasive, but from an inability to switch off the programmed spiel. It happens to all of us; people tend to ask predictable questions, and we give our prepared answers.]
What is all this computer power going to do for us?
The fifth-generation machine is a design principally concerned with thoughts and ideas as opposed to numbers. What I'm interested in what excites me - is making a machine which aids us as humans in the mental sphere in the same way as motors have aided us in the mechanical and the physical sphere. We've done very well there - right from the start of the industrial revolution we've been developing machines to aid us, and to take the burden off our muscles in every sense - machine tools, means of transport, domestic appliances, hand tools - that's gone very well; now's the time for doing the same thing for the intellect.
So far we've replaced human labour at the very lowest intellectual
levels - no intellect's needed for washing up and so on -
and now we're aiming at moving upwards and replacing intellect
at the professional level. I think that what I'm doing is
making a machine which will in due course sit in the home
and replace - or supplement - the doctor, the solicitor, the
I get the impression that you're expecting Mr & Mrs
Everybody to have a higher intellectual standard than I believe
they actually have.
Oh no, not a bit; I'm not expecting that. I don't think that
my machine will demand anything intellectually; it itself
will have to have very great intellectual powers, but its
users will have whatever intellectual powers they choose to
have. The machine will be there, and it will advise them;
they can ask it questions and it'll give them answers perhaps
even tell them what to do. They can say: 'What's on television
tonight?' and they won't have to worry about how it will get
the information - it'll decide that; it'll ring up somebody
or look it up in its memory banks, or find out by whatever
means. Or they can say to it: 'What's the first train to London
round about midday tomorrow?' or: 'I've got a pain in my right
side and really haven't been feeling too well' - and it'll
recognise them; it'll know who they are when they're talking
TV programmes and trains to London - isn't this covered
by teletext - Ceefax and Oracle?
Oh no, it's going to talk to you. If you go to teletext you've
got to think: dial 100 or whatever - I meant it literally
as I said it; you go to the machine and say: 'What's on TV
tonight?' and it'll reply: 'What do you want to watch? Should
I go right through the programmes, just tell you ITV … ?'
I see. Will it actually tune itself to my intellectual
Yes, certainly. It'll know the person; it'll deal with each
person in the family as an individual. I hadn't thought of
it as tuning to intellectual level but it's a nice way of
putting it; that's what it will be doing.
Well, in a sense it is. The way I ask my questions and
the sorts of questions I can answer when it asks me will vary
according to who I am. How does the machine find out about
It's introduced to the family when it first arrives. You'll
sit down and it'll say: 'Hello, tell me your name?' and you'll
say: 'I'm Johnny' and it'll ask questions to find out all
But what then? Do you really foresee that everyone's going
to be sitting at home talking to computers?
I think it's a wonderful thing. I think it will remove a
lot of loneliness for old people, and it'll improve the standards
of education dramatically - because we'll be able to have
individual tuition … whether it'll be done at home, or
at school with each person sitting in front of one of these
strange teachers, I don't know.
As I imagine it, you'd look at this machine and see a
face there talking to you; it would have a personality. There
wouldn't be a face; you 'd just see one … I imagine that
people would generate an image of the sort of face they'd
like to see - it would certainly assume different personalities
for different people.
Of course - eventually you could pack it all into a body
and make it into a robot …
That's a different can of worms; how far away do you think
I think we can make a machine that does all this in the
early 90s, but it'll be too expensive for domestic use until
the turn of the century.
We've already got a problem of non-standardisation - computers
which don't communicate with one another because they use
different languages or formats or media. What about standardisation
for these machines which are orders of magnitude more powerful?
Ah - this is a machine you talk to, so its language is
Interesting thought - machines from different manufacturers
would have their own internal standards, but they could communicate
with one another in plain English - or Japanese. You could
shut a group of them in a room and listen to them having a
conversation. But what about the standards for storing the
information for what's on the telly or what time the trains
It will refer to whatever systems are available, just
as you do - or I do - now. It'll be so enormously intelligent
that it will be able to cope with all the different forms
It sounds like making a rod for your own back -
The machine's got to relate to the world as it finds it.
But if you have to alter the world …
No, you're trying to alter the world to fit in with this
machine that I haven't yet developed.
So you'll need to have optical character recognition and
something blinking over the times on the timetable?
I think it'll need to be able to read the papers … I
can see what you're saying and I don't deny that standards
would help, but you've also got to be realistic. It probably
won't happen to the extent that one might like - it would
be nice if character fonts were standardised to make reading
easier but that won't happen, so the machine's got to cope
What is it that you find so terribly exciting about this
I suppose it's because it'll be the first time that humans
won't be the only known intelligence in the known universe.
We don't have to wait for them to arrive from outer space;
we can build them here.
Why do you want that to happen?
I don't particularly; I think it's exciting though!
So you treat it as an intellectual challenge rather than
as an end product that you think will be useful?
On the contrary! I think it's a product which will be
immensely useful. I think it will change mankind - create
more wealth for mankind than any other development in history.
I think that early in the next century we'll be able to make
a robot with true intelligence and patience and all sorts
of qualities; it will be able to walk into the Third World
and advise them. These machines will be as the Greek slaves
were to the Romans.
One difference is that you're not actually creating a slave
class from existing people. In the case of the Greeks and
the Romans you had a means for soaking up those who were not
employers of slaves.
The people who employed the slaves were often the intellectual
inferiors. The Greek slaves were superior to the Roman slave-owners.
The Greek was often the intellectual - reading and writing
letters, teaching the children and so on - that's why I'm
using that particular analogy. So in that sense one doesn't
see a need to mop up people.
Aren't some of today's problems borne of the fact that there
is a shortage of menial jobs - or perhaps that many people
have been 'educated' to have expectations beyond what they
I don't know that they have been educated to believe that.
I think there's a shortage of jobs at the moment because jobs
have been shed from the manufacturing industry at such a rate
that society can't adjust. Once that job-shedding ends, as
it will, then I think society will adjust, and more and more
people will be employed in service industries. I think this
is happening in the States now. Even when we have robots people
will prefer to be served by other human beings.
I think my view of what is happening is different from yours.
I think unemployment started with the industrial revolution;
when we talk about unemployment there are people who might
have been farm labourers, or people who might have been looking
after the cab horses
In 1850, 60 per cent of the population were either employed
on the land or were servants, and it must have been unthinkable
that those people could be elsewhere employed because they
were the people who were thought of as only good for ploughing
the land or scrubbing floors how on earth could they be employed?
But they were.
That's what I mean about expectations. A servant who wanted
to rise in the world could do so - well, theoretically. Most
of them were content to do what they were doing and live from
day to day in a simple fashion. They didn't think the world
owed them a living. More benign employers and fewer potential
employees with big ideas would work wonders for unemployment.
I think we'll go back to full employment in the early
90s. And then I think that one day much further ahead the
robots really will be able to do everything that humans can
- with all human actions and so on - it may be that people
will prefer to be served by robots than by humans … in
which case there won't be any jobs, unless the robots decide
they don't want to bother with us! However, we won't have
to face the problem until the twenty-first century, and it
will depend on what people actually want to do; we might change
our views about unemployment. Macmillan said to me: 'I don't
know why people keep complaining about not having any work;
some of my best friends never work at all and look at them'.
People will be brought up to that way of thinking.
But they still have to have some means of support.
They need wealth, but that wealth will be created in abundance
by the robot slaves. Think about it - supposing you could
economically make a certain number of robots. They could make
many more, and they could make a larger number still - so
suddenly you've got lots and lots of them. And they could
all go out and work, generate all the wealth - do anything
you like; they're as intelligent as we are. They don't tire.
I'm not saying when - but one day.
I think there's either a flaw in your argument or I'm not
understanding it in the right way. I can't help wondering
what on earth I shall do when robots are doing everything
for me. I suppose they'll form their own bridge fours, and
I can carry on as before. How far do you think the home computer
has educated us for the future you describe?
I think that we've done the first and second stages of
the job - we've got millions of people out there now who play
with computers, who're familiar with keyboards; the next stage
is to make a machine that's useful to them. An awful lot of
computers have been bought by people who wanted to learn about
computers - which is the intention - and to play games on
them. But an awful lot of other people have bought them and
found them inappropriate for their use. And we're really not
winning if you still choose to use a pad to take your notes,
rather than a computer. You can type, so if you had a machine
which was sufficiently right for your needs you'd have it
with you and put your notes straight into it.
But then if I had a machine which was sufficient to my needs
it would actually transcribe what we're saying wouldn't it?
Oh yes, but the first sort of machine is just around the
corner. It's not there yet, but it should be. It's what we've
got to make; the opportunity is there. Instead of writing
in your notebook, wouldn't you prefer to tell your machine
where to file the data so you could retrieve it?
It all depends what one is used to. I find that, admirable
as your idea might be, spreading out all my papers and saying:
'I want this sentence from here, and that sentence from there,
and this paragraph in here … 'is something that you cannot
do very well with a computer. Of course I know what you'll
say about word processing, but I still need hard copy to be
able to take a synoptic view of my material.
So you've tried to educate people, but they may have found
that they can't do what they'd like to do with machines. Now,
is there a danger that these people will go away sadly saying:
'Computers are not for us after all'? Do you really think
you can take a second bite at the cherry; that you can come
back with another machine and say: 'This will do what you
I think the disillusionment is among a very few people.
The market that I'm concerned with is the millions of young
people who've played with computers and made them do just
what they wanted them to. Now they're going to think - if
we present the right machine to them: 'Ah yes, here's the
Yes. What tasks do you think young people want to do with
computers that they can't do at the moment?
I think that they can do all the tasks with the appropriate
machine, but to do all the tasks well is perhaps too expensive
- and portability is so vital. I think that the reason that
we cling to paper is that it's so portable.
Yes, and so fileable and so throw-awayable.
Yes, but it's the portability that's still lacking with
computers. You want a printout of the data so that you can
take it with you - not because you want the paper. If the
whole machine always goes with you without it being cumbersome,
then you can do away with the paper.
Even this small step needs a very great change in attitudes
and methods of working. How much greater will be the problems
of your new Periclean Athens?
Manifold. I don't want to think about it!