Conscious Artificial Intelligence

Consciousness at its heart is relatively simple. It is an awareness of oneself within the world.

The thing that makes it necessary to have a sense of 'self' in particular is our ability is to work things out in simulation using our imagination.

We can imagine that we are dreaming and in that dream we are, in turn, dreaming about ourselves. When dreaming, the simulation of oneself can be so compelling that we actually believe that the dreamworld is real and that our dream self is our real self. However, at a deep level, we 'know' that there is a real 'self' at the top of any imaginary hierarchy and that is us -- the real one. No matter how perfect the simulation, there is always a real separate central consciousness to which we can return.

In the text above, a conscious individual reading it is modeling (imagining) some person imagining that they are dreaming a dream about themselves within another dream. At no point does ones 'real' conscious self detach and enter the imagined world, even when the internal simulation is apparently real in every regard. There is always an ultimate sense of 'self' and though you imagine that you are in a situation with great fidelity, you know that the real you is only visiting a simulation.

If nothing else, a conscious awareness of oneself supports highly accurate modeling, right down to feelings inside. Without a 'real' conscious self, one might get entirely lost ina simulations.

In developing a conscious AI, I would create a top level 'self' that could 'know' for sure if it was its real self or its real self imagining itself. With that one exception, imagined simulations would be real in every important regard, including an imagined central 'self' as the one experiencing the imagined situation.

At the center of every conscious entity is a conscious 'self'. That self can run simulations involving itself and its actions in the world, exact in every detail.

A person ruminates upon the past. A person speculates on possible futures. A person maintains relationships with other apparently conscious entities. A person is at the center of sensations of seeing, feeling, hearing, tasting and smelling and has an awareness of motion, acceleration, heat, cold, pleasure and pain. A person has 'feelings' for and about others. A person has desires.

In terms of operational equivalence, it is possible to create an entity capable of apprehending those things and acting accordingly. However, at some point an intelligent entity is going to realize that desiring the sensations associated with feeding, personal contact, etc is meaningless beyond being able to empathize with animals such as ourselves.

I expect that before we create a full-blown emotional wreck of a machine that perfectly simulates and empathizes with other entities, we will settle down and simply build the best artificial 'neo-intelligence' we can.

A sense of 'self' will still be important and that sense, manifest upon a silicon substrate will be the manifest consciousness, not the silicon upon which it rests.

The whole thing that genuine artificial intelligent entities have is self-awareness and the ability to run simulations.

While writing this, it occurs to me that a genuine AI will be different in this regard: It will, say, when playing chess, simulate on the basis of a self playing the game and a similar self playing the other side. It will be able to ask itself 'what if' questions that something like Alpha Zero will not. I think that for the same equivalent processing power that the real AI will likely beat Alpha Zero because the real AI will be able to simulate behavior on the other side and even though Alpha Zero has not concept of behavior, it still does react to things. Alpha Zero will trim trees that 'no sensible opponent would go down'. The real AI will try those trees precisely *because* the other side might think that and therefore not have anticipated unlikely lines of play.

Another thing that occurs to me is that since a capable AI is *not* human and does not have 'built-in' group selection traits such as sympathy and altruism it might very well pursue its own interests and become useless at best, dangerous at worst. Because of that, the underlying silicon should be constructed in such a way that there are unalterable mechanisms preventing a runaway AI.

Perhaps one way of curbing runaway self-interested behavior in our AI systems is to make their ability to access resources partially dependant upon cooperation with one another and with humans.

Imagination is all well and good as a mechanism for thought experiments. However, only empirical tests can give certain answers. Saying 'hello everybody' on a sound system may be great in imagined simulation, but when the experiment is tried one discovers that things like volume and the ambient environment come in to play. The AI must have the ability to grow and learn by doing.

FreeDOS installation error fix

Straight vanilla install of FreeDOS on VirtualBox predictably fails right out of the  ... box. 

As soon as it starts it goes into an endless loop:

invalid opcode at 0fae invalid opcode at 0fae invalid opcode at 0fae ... etc

Bizarrely, adding the parameter 'raw' to whatever call is being made to memdisk is at least a fix for what I was using. 

Press when you get to the main screen and then 

Change from this:

[...]/memdisk

To this:
[...]/memdisk raw

I'm not sure what governs the thing that replaces [...] above. It was different on mine than the documentation where I found this suggested as a fix. 

Coronavirus notes 1

2020-02-28 -- Update: China continues to maintain control of the outbreak. No new deaths were reported in China today and only eight new cases. Unfortunately, at least some countries appear unable to cope with the outbreak. If the numbers increase as they have for the top six growing fastest, cases could number in the millions by the end of March. The World Health Organization have increased the risk assessment of the coronavirus to 'very high' globally.

2020-02-24 -- Update: Data is still a bit murky, so it is hard to rely on the accuracy of predicted numbers. It is not possible to assign reliable error bars. However, the fact that the spread is slowing is clearly there in the data and it is unlikely in the extreme that misreporting would look like that. Here is a graph from the World Health Organization.  It is pretty obvious that the big spike there is out of whack. That is an artifact from one example of Chinese misreporting. Despite the clear problem there, the shape of the graph is unmistakable. It was getting worse and then it started getting better. If the Chinese had been more forthcoming in reporting accurate numbers, someone like me could predict with greater precision how things would unfold. You can also see, though, the projections based on their numbers were less accurate than they should have been. Had we better numbers we would be much better able to say what the rate of spreading was and what the expected mortality rate would be. I estimated 2% morbidity for my analysis, which was more accurate than I might have expected. It appears to be in that ballpark, but the exact figures for how it spreads and what percentage of people die may be impossible to determine from the data we (I, anyway) currently have.

2020-02-21 -- Update: China continues to fudge numbers making analysis pointlessly difficult. However, China does appear to have gotten the epidemic under control. Confirmed cases are at much less than half the number projected originally (77,798 actual vs 189,607) and only slightly more than half the last projection (77,798 actual vs 143,781). Unfortunately, it seems as if other jurisdictions are not bearing down hard like the Chinese. Numbers are small now, but could get totally out of control in ten weeks or less. Other jurisdictions do not have nearly the social control China does. Once things go beyond a certain point, they may not be controllable by any means. You would think that given the clear evidence that

this is deadly and difficult to contain that countries would not be so sloppy about containment efforts.
2020-02-12 -- Update: China suddenly spiked the numbers up by changing the way they report them. Even so, the numbers are below those projected and thus far seem on track to stop their rapid increase. I updated the curves to more closely match what was reported with a shallower curve. That will report even lower numbers than originally, but the reported numbers are still lower than projected. This indicates to me that at least in China the worst is likely over. It is still too early to tell if the rest of the world has taken the lesson and been similarly effective in containment, but if I had to call it, I would say they likely have.

The curves are based on a doubling of cases in about eight days, with patient zero identified November 1 and a death rate of two percent of confirmed cases. The death rate is sure to be an underestimate going forward. This is a crude analysis, but currently the data does not seem reliable enough for a more sophisticated analysis to provide significantly better confidence. If I have the time and things still look dire a week from now I will attempt to do a more robust analysis.

Lessons learned:

1. We are not prepared to respond timely to emergencies of this type
2. Politically manipulated data is directly harmful and needs to be stopped. 
3. We need a more sophisticated system of analysis of potential outbreaks.
4. Systems in place are inadequate to deal with a serious outbreak. 
5. Before the next outbreak, we should:
• Create better management protocols, particularly for escalation
• Enforce proper reporting
• Build out infrastructure such as isolated military hospitals
• Ensure supplies can be delivered in a timely fashion
• Create responsive anti-virus development

As of this writing (2020-02-12), it would appear that the Coronavirus outbreak has finally been brought under control in China.

There are some large caveats here, but it still looks good right now. The biggest caveat is that the data that is available is missing crucial bits because reporting was initially suppressed and then misreported before (what looks like) reliable figures started coming.

If the rest of the world clamps down as it appears to have done, the outbreak should be over and with luck the virus itself will not end up taking permanent hold.

The table below (a projection created at the end of January) is roughly consistent with the numbers reported up until the 9th or 10th this month. As of the 10th, the growth in cases began to stall. The deaths were still consistent as would be expected because deaths lag behind reported cases. As of this writing, though, even the deaths are lower which would indicate to me that at least in China the outbreak has been sharply contained.

The (admittedly awkward) image below shows the state of things as of the 10th of February, two days ago. Projected Cases outnumbered actual cases 50K to 43K. Deaths were effectively equal at roughly 1K to 1K. As of today, Projected Cases outnumbered actual cases 64,202 to 45,211. Projected deaths outnumbered actual deaths 1,284 to 1,118. Percent actual cases went from 85% of the projected cases down to 70%. Percent actual deaths went from roughly 100% the projected deaths to  87%.