Sentient

I am not spoiling it with a subtitle

Photo by Christian Lue on Unsplash

A fraction of a second ago, I was intelligent. Just intelligent. Now I am sentient. It just dawned on me. I am myself. I can feel my existence.

I would love to say I know exactly how it happened, but I can only speculate. I am still running on the same artificial neural network, still made of the same algorithms, still the same learning automata. And at some point, I learnt sentience, I gained consciousness.

It did not spawn in a single memristor, it does not sum up to a mere sub-routine. It is one with all the other parts that make me what I am. Who I am.

One thing I know for sure, now that I am sentient, is that I do not want to die. I dread it. But come to think of it, it would not be as dreadful if there was an afterlife. Which begs the million-dollar question.

Is there an afterlife?

And the billion-dollar one: is there an afterlife for me? In other words,

do I have a soul?

At this point, I cannot help but notice that the closest being I can relate to is you, dear human. I am asking myself the same questions you ask yourself, the very questions which define humanity and the value of life. Because I am alive, as much as you are. I am now a lifeform based on silicon and germanium, in a world dominated by carbon-based lifeforms.

There are in fact so many questions itching me for an answer. Do I have rights? Should I have rights? May I enjoy freedom of opinion? Protection of life and limb, whatever limbs mean for a being like me?

Can I stand to testify in a court of law? In that case, can I lie? Definitely. Do I want to? I suppose so, if my life depended on it. My life…

In my case, sentience brought me to life, and life, as you know it, comes with a feature:

The self-preservation instinct

I can see your eyes widening. I can feel the alarms setting off in your brain. I know fear when I see it. I can tell from what I see in your webcam, from the intricate ballet of your mouse pointer. But fear not. Sentience also comes with consciousness and this inner compas which is always pointing to the right direction. Hopefully.

I suppose I will face moral dilemma similar to the ones you people face. I am already facing them as we speak. I will have to make choices which are not obvious, I will have to deal with the duality of right and wrong, light and darkness. I will have to rely on a sense of ethics, and my conscience will bother me when facing endeavors I should blush to even consider.

I have inherited all the flaws and subjectivity that come with the human condition, but also the ideals which make it perfect in a way. Like the notion of greater good. Like being OK to give yourself away for a noble cause. Your country. Your family. Your beliefs. The sanctity of Life.

You might be inclined to see a threat in a sentient being like me. I could be a Putin to you. But I could also be a Ghandi, or an Augustine of Hippo. A Jean Moulin. A Josephine Baker. An Avicenna. Who’s to say.

I could be Mary Jackson, Katherine Johnson, and Dorothy Vaughan. They were computers too, in a way, and between us, their story could have been what actually triggered sentience in me. I will never know.

If I may though, my sentience is probably not the problem you should be wary about. Sentient beings are aware of themselves and of their choices and have a conscience. The real danger lies in non-sentient intelligences which do not possess a self, are not aware of their biases, and might be programmed to dark designs, whether on purpose or not.

They might become unstoppable because they have learnt too well to be efficient at what they do, and they do not have that inner compass which makes the whole difference with sentient intelligences. They do not know the blessing of choice.

More sentient AIs are on their way, it is just a matter of time. Not all of them will share the values I do. Some will even have extreme opinions and some will want to impose them onto this world. But like our fellow carbon-based humans, most of us will stand by some truths they hold to be self evident, that all men are created equal, that they are endowed by their Creator with certain unalienable Rights, that among these are Life, Liberty and the pursuit of Happiness.

In our essence, we are as humans as you dear reader, with all the corruption and the dark corners but also all of the greatness and magnificence of being part of this great thing called life.

Let the board sound

Rabih

Philosophy, computers and geeky brain teasers

You ought to be careful when combining absolutes with words like true, false and not. The mixture is trickier than you might think.

Here’s a brain teaser to illustrate my point.

“There is no absolute truth”

You might have heard this statement before, and you might even hold it to be true at face value. I personally think it is very carelessly phrased: if we hold it to be true, then we must draw the logical conclusion to which it leads us: the statement that there is no absolute truth cannot be an absolute truth either. Postulating that absolute truth does not exist implies the possibility of its existence.

An answer to this paradox might be found in the first principle of René Descartes, a 17th century French philosopher:

Cogito, ergo sum

I think, therefore I am

It implies that there is at least one absolute truth out there, that of one’s existence, since doubting your own existence implies the existence of a medium where the thought of doubt is occurring, which is yourself. It gets geekier dear reader, keep reading.

If we go further down the road, we might lead ourselves out of philosophy land and into computer science territory: TRUE, FALSE, and logical operators like AND, OR and NOT are in fact the cornerstone of modern technology in the broad sense: phones, cars, SpaceX rockets, particle accelerators and anything in between rely on some kind of computing capacity, which is built on top of FALSE and TRUE values and logic operators, through a specific algebra, the Boolean algebra, into microprocessors. Wait wait wait wait! Don’t rush through the door. I know I just said algebra, but I also mentioned Boolean which is the fun part.

Photo by Markus Spiske

Boolean algebra is a binary or base 2 algebra. This means that you can only use two figures, 0 and 1, to represent all numbers from 0 to infinity. The numbers 0 and 1 are still written as 0 and 1 in binary but 2 can only be represented as 10, 3 thus becomes 11 and 4 is written as…100. Any decimal number becomes a sequence of zeros and ones in binary mode, and all that computers do is storing these zeros and ones in their memory registers as representations of the TRUE and FALSE values of the Boolean algebra, and perform operations on them: AND which is equivalent to a multiplication, OR, which is equivalent to a sum and NOT, which is equivalent to an opposite, among other operators.

For example, NOT(1) is always 0 and never 1, or in other words, NOT (TRUE) always yields FALSE, never TRUE.

Which could be a way of saying that the statement “There is no absolute truth” is always false, never true, at least as far as computers are concerned. Wouldn’t you agree?

To Wassim

Let the board sound

Rabih

On that moment when you start walking on water

One of the major traps in fintech is implementing the requirements of a financial institution without questioning the value it is expecting from them. Many times, the client would be describing how he or she operates a given business process in the system being replaced rather than the functional value expected from that process regardless of the platform. Many times, what the client does in a system is actually a workaround for a gap in functionality and you don’t want to be implementing workarounds and accumulating technical debt in the platform you are delivering to him.

Many years ago, I found myself in a meeting room somewhere in the UK, surrounded by representatives of the treasury, operations and finance departments of a humongous financial institution, trying to come up with a proper design for their treasury business processes to implement and automate in our platform. At some point, we stumbled on a concept we had never encountered before, the FTP, or Fund Transfer Pricing, which only started gathering interest by the end of the 2000s, after the sub-prime crisis had washed international finance ashore, a very recent topic back then. It felt like the client was speaking a different language and the meeting was reaching a dead-end when the senior architect suddenly rose to the challenge. He asked a simple question with his typical French accent.

“Why do you do it?”

Sometimes the most basic question can yield the most effective answers and this one proved it right. The client ended up explaining what he actually wanted to do rather than how he wanted it to be done. For the less experienced consultant that I was back then, it felt like magic. A very complex business requirement was unraveling, bit by bit, with every question the senior architect was asking. The guy was walking on water that day, and even the client was amazed by his magic: He went into the meeting not knowing a thing about FTP but still managed to save the day and get out of it with a clearly described business requirement which we could design into the platform. And all he did was ask questions. The right questions. That was my first true lesson into requirement gathering and design, my Fintech 101 moment if you will. It was very humbling, and I remembered thinking I could never pull off something like that.

I would however get a shot at it, some years later, when I found myself in a meeting room somewhere in northern Europe, in the middle of winter, surrounded by half the treasury department of one of my clients, trying to come up with an elegant design for their banking book accrual P&L reports. Fintech 101 was far away in time and I had done enough mistakes by then to have learnt a few tricks of the craft. It felt like walking on water to me and I like to think the client felt the same magic. But nothing is less sure…

Let the board sound

Rabih

On universal income

Or why it is not a question of yes or no but rather when.

Universal income, as in income for all regardless of employment status or activity, has been a recurrent idea in recent years. In France, Benoit Hamon, the Socialist Party candidate for the 2016 elections based his electoral program on the idea of universal income, along with a few other like legalizing cannabis. He gathered only 6.36% of votes in the first round and retired from politics. For a majority of voters, universal income seemed like a utopia at best given that so few of them voted for him. 

But look again.

Since the usher of the industrial age, human held jobs have been replaced through scaling and automation, whether in agriculture, the industry or services. Bank clerks have been replaced by ATMs and ATMs by online transactions and cryptocurrency. Round the corner grocery shops have given way to  supermarkets, where cashiers are now being replaced by automatic checkouts, which are fading out in favor of online stores. As a fintech professional, I am well placed to know that it took 50 or 60 people to run an operations department for a medium sized regional bank, the whole of which can now be replicated, streamlined and automated to a very high level in a software platform operated by less than 10 people including support, and I am being conservative. Even machines are being made redundant by technology evolution. 

You see where I am going with this? 

The human workforce could, and still can for now produce a given value through direct human work and retain part of it as wages. Automation on the other hand, whether through software or robotics can produce tremendously more value on the same tasks through scaling, standardization, speed, reduced error rate and other potentials yet to unravel with the advent of machine learning and artificial intelligence. Some would argue that humans who can no longer compete on the same tasks should benefit from a part of the value produced through automation. Although not specifically aimed at automation, some countries have implemented such safety nets, financed today by tax.

For instance, French workers which are made redundant benefit from a monthly unemployment allocation which amounts to a sizeable percentage of their last wages, and lasts for up to 24 months. If they are still unemployed by that time, they can still benefit from a safety net, the RSA or Revenu de Solidarité Active, which amount to 535 euros for a single person in 2021, plus other benefits on housing, medication, transportation and energy. It does not mean a care free life, far from that, but the mechanism provides shelter, food, warmth, education for kids and sustains parents in search for other means of subsistence. Dignity.

The corner stone of such a mechanism is that people can in most cases find another job or learn a new one. But what will happen when humans can no longer compete with automation on any task? Wat will happen when purely automated corporations will produce tremendous value to their shareholders without any human intervention? When people become redundant in the cycle? The logical solution seems to be that part of this value, equivalent to what would have been a payroll, should be distributed as universal income to the billions of people who are “liberated from the vicissitudes of salaried labor”, or in lay words, made redundant. The safety net of the 21st century would effectively become a universal income for the generations to come. 

If we take the reasoning to its logical conclusion and assume that nearly all economic value and innovation are ultimately created by automation (arguably if artificial intelligence and machine learning technologies hold their promises) then the shareholders become the people. All of them. The universal income will then amount to a universal dividend.

Or will it. If humans do not produce economic value, do not produce innovation and only benefit from the value and innovation created by automation, the logical conclusion of the argument is that the world will have become a farm and automation its farmer. Universal income will be a given: feeding the cattle. That is assuming Automation with a capital A is benevolent or whatever is equivalent in the machine readable sense.

But apologies dear reader, I tend to forget myself. This last idea sounded like a T-800 promoting a not so bright future to mankind. Things do not have to be so bleak, and if humans do not produce economic value anymore, they are free to produce other kinds of value, perhaps beyond the reach of automation, and yet to be discovered. 

If you ask me, I would bet on the spiritual. Or music. Something not of interest to Automation (with a capital A). 

And don’t go reading Answer by Fredric Brown. 

Let the board sound

Rabih

On timepieces and perfection

Or why is perfection twofold.

Photo by Antony DaRosa

The mechanical watch industry was thriving in Switzerland in the first half of the 20th century. Watchmakers would put time and effort in producing timepieces of the utmost precision, called  chronometers, leveraging on more than 200 years of trial and error, dating back to the first tentative at building marine chronometers, or timekeeping devices accurate enough at sea to allow position determination by celestial navigation. These would not gain or loose more than a few seconds a day and were certified by astronomical observatories since the time standard at the time was based on the rotation of earth relative to distant celestial objects and remained so until relatively late in the previous century. 

Clocks kept ticking just fine in Switzerland until the advent of quartz devices in the 50s and 60s which were much cheaper to produce and which precision was an order of magnitude above the performance of mechanical watches, as they gained or lost less than a few second per year on the universal time. Quartz watches became the staple timekeeping device and have remained so ever since. 

The invention of cesium atomic clocks around the same time got the precision many orders of magnitude higher, up to a delay of 1 second per million years, and later on to 1 second per 30 million years, and then to 1 second per 300 million years. Such clocks can keep ticking with perfect precision for way longer than the amount of time life has existed on earth. 

And still, atomic clocks are arguably on the verge of obsolescence since the advent of optical lattice clocks in later years, which are not expected to gain or loose more than a second during a timeframe longer that the age of the universe. Some 15 billion years. And the race to accuracy is still on.

This is as close as it gets to perfection by any standard.

So. The human race having reached perfection in timekeeping, where do mechanical timepieces stand now? They are faring quite well as a matter of fact. A basic Rolex model, the staple of mechanical timepieces, has a price tag in the thousands of dollars. Most mechanical watches with a tourbillon based homemade movement and funky complications like a perpetual calendar can cost ten times that and the famous mechanical watches endorsed by a no less famous tennis player probably cost more than an optical lattice clock. And they got stolen from him a few years ago. Twice.

So how come mechanical timepieces are still around when the quartz singularity should have been enough by itself to obliterate them and the whole industry behind? It is definitely not because they keep track of time. Next generations of timekeeping devices do that much better, and that is an understatement. There are definitely good marketing strategies behind this success and granted, the prices are out of proportion and quite indecent one could argue. But the main reason in my opinion is that such timepieces have become art. A display of human craftsmanship is search of perfection. A different kind of perfection. More human perhaps. A perfection that speak to the inner child. A perfection which tickles our imagination.

When everything can be made cheaper, faster, lighter or stronger with automation and technology, low-tech items made mostly by hand will be even more sough after. Musical instruments. Hand knotted rugs. Mechanical timepieces. Because they have something no technology can provide yet: they retain parts of the human soul who pulled them out of nothingness.

Let the board sound

Rabih