[ExI] Hard Takeoff

[Brent Allsop]

Hi Michael,

Yes, it is fun to see you back on this list.

I'm still relatively uneducated about arguments for a "Hard Takoff".  
Thanks for pointing these out, and I've still got lots of study to fully 
understand them.  Thanks for the help.

Obviously there is some diversity of opinion about the importance of 
some of these arguments.

It appears this particular hard takoff issue could be a big reason for 
our difference of opinions about the importance of friendliness.

I think it would be great if we could survey for this particular hard 
takeoff issue, and find out how closely the break down of who is on 
which side of this issue matches the more general issue of the 
importance of Friendly AI and so on.

We could even create sub topics and rank the individual arguments, such 
as the ones you've listed here, to find out which ones are the most 
successful (ie, acceptable to more people) and which ones are most 
important.

I'll add my comments below to be included with Allan's and your POV.

Brent Allsop


On 11/14/2010 12:32 PM, Alan Grimes wrote:
> chrome://messenger/locale/messengercompose/composeMsgs.properties:
>> <mailto:stefano.vaj at gmail.com>>  wrote:
>
>> 1.  ability to copy itself
> Sufficiently true.
>
> nb: requires work by someone with a pulse to provide hardware space,
> etc... (at least for now).
>
Michael.  Is your ordering important?  In other words, for you, is this 
the most important argument compared to the others?  If so, I would 
agree that this is the most important argument compared to the others.

>> 2.  stay awake 24/7
> FALSE.
> Not implied. The substrate does not confer or imply this property
> because an uploaded mind would still need to sleep for precisely the
> same reasons a physical brain does.
I would also include the ability to fully concentrate 100% of the time.  
We seem to be required to do more than just one thing, and to play, have 
sex... a lot.  In addition to sleeping.  But all of these, at best, are 
linear differences, and can be overcome by having 2 or 10... times more 
people working on a particular problem.

>> 3.  spin off separate threads of attention in the same mind
> FALSE.
> (same reason as for 2).
>
>> 4.  overclock helpful modules on-the-fly
> Possibly true but strains the limits of plausibility, also benefits of
> this are severely limited.
>
>> 5.  absorb computing power (humans can't do this)
> FALSE.
> Implies scalability of the hardware and software architecture not at all
> implied by simply residing in a silicon substrate, indeed this is a
> major research issue in computer science.
I probably don't fully understand what you mean by this one.  To me, all 
computer power we've created so for is only because we can utilize / 
absorb / or benefit from all of it, at least as much as any other 
computer would.

>> 6.  constructed from scratch with self-improvement in mind
> Possibly true but not implied.
>
>> 7.  the possibility of direct integration with new sensory modalities,
>> like a codic modality
> True, but not unique, the human brain can also integrate with new
> sensory modalities, this has been tested.

What is 'codic modality'?  We have significant diversity of knowledge 
representation abilities as compared to the mere ones and zeros of 
computers.  I.E. we represent wavelengths of visible light with 
different colors, wavelengths of acoustic vibrations with sound, 
hotness/coldness for different temperatures, and so on.  And we have 
great abilities to map new problem spaces into these very capable 
representation systems as can  be seen by all the progress in field of 
scientific data representation / visualization.

>> 8.  the ability to accelerate its own thinking speed depending on the
>> speed of available computers
> True to a limited extent, also Speed is not everything.

I admit that the initial speed difference is huge.  But I agree with 
Alan  that we make up with parallelism and many other things, what we 
lack in speed.  And, we already seem to be at the limit of hardware 
speed - i.e. CPU speed has not significantly changed in the last 10 
years right?
>> When you have a human-equivalent mind that can copy itself, it would be
>> in its best interest to rent computing power to perform tasks.  If it
>> can make $1 of "income" with less than $1 of computing power, you have
>> the ingredients for a hard takeoff.
> Mostly true. Could, would, and should being discreet questions here.
>
I would agree that a copy-able human level AI would launch a take-off, 
leaving what we have today, to the degree that it is unchanged, in the 
dust.  But I don't think acheiving this is going to be anything like 
spontaneous, as you seem to assume is possible.  The rate of progress of 
intelligence is so painfully slow.   So slow, in fact, that many have 
accused great old AI folks like Minsky as being completely mistaken.

I also think we are on the verge of discovering how the phenomenal mind 
works, represents knowledge, how to interface with it in a conscious 
way, enhance it and so on.  I think such discoveries will greatly speed 
up this very slow process of aproaching human level AI.

And once we achieve this, we'll be able to upload ourselves, or at least 
fully consciously integrate ourselves / utilize all the same things 
artificial systems are capable of, including increased speed, copy 
ability, ability to not sleep, and all the others.  In other words, I 
believe anything computers can do, we'll also be able to do within a 
very short period of time after first achieved.  The maximum time limit 
between when AI would get it, and when we would also acheive the same 
abilities, would be very insignificant compared to any rate of overall 
AI progress.


Brent Allsop