This is an image of the whiteboard in my study. In the top left hand corner are 4 numbers taken to a LOT of decimal places. What are these numbers? A good mathematician should get 3, and exceptional mathematician should get 4. How many can YOU get?
All posts by Greg
A beautiful clear evening (but with a bright Moon overhead) so that meant star imaging only. There, low on the southern horizon was beautiful Sirius blazing away – so that was the target! Managed to get 22 x 2-minute subs using the Canon 200mm prime lenses and the ASI 2600MC Pro OSC CMOS cameras. EXTREMELY pleased with the result.
5-frame focus-stacked mosaic of spirogyra in conjugation
The Real Dangers of AI
Up until today, I haven’t really been bothered at all about the hype generated concerning the dangers of AI. The “singularity” and Terminator-like scenarios are as far as I’m concerned just complete garbage and not worth thinking too much about. However, there is another side to AI that is far from innocuous.
When I was at University I received a phone call from a colleague who was seriously asking me whether Mars was going to appear the size of the full Moon in the next few days. He had of course read about this nonsense on the Internet, and typical of most people reading stuff on the Internet, he thought there must be some truth to this. Reading stuff on the computer screen seems to lower the brain’s disciminatory functions for some reason – and if that’s just from reading stuff, then what happens when you actually have a CONVERSATION with your computer (or at least the AI program that is interacting with you)? The answer to Mars being the size of the full Moon by the way is that even Stone Age Man knew better than that! Get a grip.
We are currently at a stage where an AI program can hold a reasonably good (in truth, frighteningly good) conversation with you. Now thanks to the way you let all your interests, likes, dislikes, prejudices, hobbies – whatever – freely disseminate on the Internet through all your Social Media connections – it is not too difficult to program the AI to specifically interact with YOU in a very meaningful way! Think very carefully about that. In the course of what appears to be a plain ordinary conversation, the AI could manipulate you, very easily, without your knowledge. Not possible? Well I wouldn’t have thought that it was possible that anyone would believe Mars was going to appear the size of a full Moon – and this use of AI is far more intrusive into your Mind than just reading about some nonsense.
Now let’s go one stage further into the real nightmare scenario. It’s all great fun to type in your conversations and have the AI program type back its comments. But we’re not going to be happy staying at that primitive level of communication, are we? Soon we are going to SPEAK directly to the AI program and it will SPEAK back to us through our speakers having gone round the invisible (to us) loop of speech recognition programs and text to speech translation programs. The invisble software tech behind the “meaningful” conversation you are having with the AI is immaterial. What is NOT immaterial is what can now happen during the conversation.
I was in two minds whether to write this bit as I didn’t want to give the wrong people ideas which could wreak havoc on the planet, but I’m sure those developing these systems are already fully aware. This being the case I think it is my duty to make YOU also fully aware of where this can lead.
All the AI software guys need to do is team up with a Paul McKenna or Derren Brown-like character, and THEN see the damage that can be done! Without your knowing, whilst having a totally (to you) innocuous conversation with this charming AI – you are “put under” by hypnotic suggestions from the AI to do/carry out – God knows what.
At a Government level you could all be made to conform to their requirements. For a terrorist organisation you could be programmed to go out and carry out attrocities. For a commercial organisation you could be made to purchase stuff you don’t really need/want. And of course there are MANY more possibilities too (again I don’t want to give away ideas that could easily be the end of us).
Bottom line. Think VERY CAREFULLY before you engage in the seemingly innocuous business of interacting with an AI. You don’t know the motivations behind the programmer!!
What is Consciousness?
Looking at some old emails from way back in 2015 now, 8 years on, the idea I had back then seems better than ever to me now. The reason I wrote it off back then, is that I mailed David Deutsch my idea, and he stated emphatically, that there couldn’t be a Quantum Computer side of the brain, even with my “modification”.
Now you might be aware of the decoherence time problem (if not Google it), and the real problem of having anything like enough decoherence time available when working with room-temperature wetware (the brain). Well, my way of overcoming the decoherence time problem in the brain was by invoking the Quantum Zeno Effect – the Quantum equivalent of a watched pot never boiling. Basically, if a Quantum State is being observed then it lives longer than if it is not observed. I reckoned in a similar fashion that the decoherence time of entangled states could be similarly increased by being closely “observed”. David Deutsch didn’t think so – and him being far more engaged with this Quantum stuff than me, I thought so to, and forgot about all of this.
But then I read a bit more about what I was thinking about back in 2015, and now I’m not so sure Deutsch has all the answers.
My “feeling” was that our brain is comprised of two main components – our CONSCIOUS brain – which is based on a massively parallel classical computer – AND – our SUB-CONSCIOUS brain which is based on a Quantum Computer. If our sub-conscious IS the quantum part of the brain, then that would go a long way towards explaining many of the (almost supernatural) capabilities of the sub-conscious. But how do you get a decent decoherence time? The conscious (the observer!) part of the brain is in intimate contact with the quantum part, and it is this that pushes up the decoherence time to (practically) useful values.
So, what is consciousness itself? I believe it is the product of a classical parallel computer working together (intimately) with a quantum computer. So, I don’t believe a “classical” computer on its own, no matter how big and complex, and no matter how big and complex the program, can ever become “self-aware” in the hard A.I. sense. I don’t know whether a sufficiently large quantum computer on its own could become self-aware, but my feeling is that it would be no more self-aware than a classical computer. However – put the two intimately together – and who knows?
Fractals from Complex Power Iterations
Complex Power Iterations
If you play white noise through a pair of headphones – then after a while, with a little practice, you can hear any tune you choose in full detail – it is quite astonishing.
So I wondered if you could do the same with vision. Here is a random array of dots. By staring at the image can you pull out a picture/shape of your choice?
Star Xterminated Pillars of Creation
Sirius – The Brightest Star in the Sky
Out of all my bright, single star images, I think this one is my favourite. Sirius – the brightest star in the sky. Two-frame (vertical) mosaic with camera in landscape mode. 3 x Sky90 refractors and 3 x M26C (non-Trius) OSC CCD cameras. Each frame is 90-minutes of 2-minute subs (so that’s actually 270-minutes of actual exposure time per frame).