We have profiled some of the most amazing and interesting people in this space. So we thought we'd like to show case them again - just in case you missed them the first time around... Today it is the turn of Anish Mohammed and his take on AI - worth reading!
What happens when machines become smarter than humans? This is not a what if but a when. Jillian speaks with Anish Mohammed and learns many things including how rare it is to have an obese genius. Who would have known? Zero Sum Games, Zero-Knowledge Proofs and Elliptic Curves are all mentioned.
Technological Singularity is when super- or ultra-computers become more intelligent than humans. It is a point in the near future that poses a huge issue: at the point of singularity or the emergence of this ultra-computer, will the newly more intelligent computer still respect humans? Will the intelligence be able to encompass human traits such as compassion, understanding, and empathy?
It is easy to program for right or wrong, but hard to program for humanity. It’s like the age-old question: if you could go back in time to 1933 and assassinate Hitler, would you? Should you? Would another Hitler, maybe an even worse tyrant, rise in his place? Would the lessons learnt from enduring Hitler in the last century have some value or even global significance for the future? Could the world be saved because of Hitler?
What decision would a computer make, a super-computer? And how can humanity be protected even as it creates the potential creator of its own demise? Then factor in how AI progressively creates better versions of itself, and ask how would technical singularity be revisited again and again as the latest, best versions of the super-computer inch forward out of the carcasses of their previous versions, like multiple butterflies from the same caterpillar.
If your head is starting to hurt with my largely unformed and naïve thoughts on Singularity, then I am glad. Ninety minutes talking with Anish had begun pleasantly enough but as we expanded further into his brain it felt like I had taken a wrong turn and found myself in an adult playground where language and form had been utterly transformed. He kept on sending me links to refer back to, but I was not keeping up.
Then we, or rather he, posited about the authenticity of humans. “We often refer to the fact that it’s because we have thoughts that differentiate us from machines. Thoughts make us authentic and human.”
Some 80% of all energy consumed in the body is by the brain. When Anish told me that last night I merely accepted it. Now, this morning, I am wondering if I can turn my gym workout into a brain workout and does that mean that all brainy people are mostly skinny? Right now I am not sure I can think of any fat geniuses. And a quick scoot around Google seems to uphold that theory (among other random traits of geniuses such as being taller, having blue eyes, being an atheist and having a high sex drive – maybe that last point might play into the skinny bit).
But what Anish is asking is: how do we measure brain cogitations? “Humans and machines are evolving together,” he explains patiently. “Consider the iPhone. The iPhone changes the way people think. Before, if you were travelling in a strange part of town, you’d get directions or take out a map and plan your journey. Now the default response is to load up googlemaps.
“This directly impacts on the neural pathways of the brain. This has been studied already using retinal implants in humans.
“Another aspect that we claim as uniquely human is our ability to dream and have memories. But both these unique attributes are incredibly fragile, and we are largely biased in how important these are. Memories or dreams do not relate to anything real and yet humans value the authenticity of memory and thoughts even if they are not true.”
How on earth had we strayed into this line of questioning? In fact, it was less my questioning Anish and more Anish making questions of my belief systems on humanity. Again this morning dreaming came into sharp relief as I watched the family dog, Bart, follow endless rabbits in his sleep, his legs whirling in chase. Are a domestic pet’s dreams authentic and valuable? Are animals and machines evolving in unison or are they too in danger of being left behind?
Well, it all began last night with a shared post-colonial appetite for changing place names. Anish had described to me his upbringing in the region of Kerala, in southern India. As a province, Kerala experiences many of the same characteristics of Northern European countries – high levels of literacy, low levels of infant mortality. It demonstrates an inversion of gender ratio leading to a decreasing overall population, another trait enjoyed by much of the old European landscape.
Anish attended the local high school and then went to medical school in the University of Thiruanthapuram. He had to spell it for me and then ultimately message it to me. Thiruanthapuram has changed its name several times from the colonial name applied by the British as they were unable or unwilling to get their tongues around such multisyllabic appellations. I pointed out we were the same in Ireland with many town names moving from English back to Irish names: Kingstown to Dun Laoghaire for example but I won’t ask non-Irish readers to hazard a pronunciation.
While Anish studied medicine he acknowledges it was largely for his parents’ sake. He had much more interest in all things computing and programming. His father was a deputy director in the department of technical education in the state and, as a result, Anish had access to the computer science labs where he spent much more time than over at the medical halls. Then, in 1994, Bruce Schneider’s Applied Cryptography fell into his hands and he was hooked.
“My first real job after getting my medical license was with Ericcson working on digital payments.”
Where did that come from?
“During my medical training I was self-teaching and self-reading,” says Anish. “Ericcson wanted me to build a micropayments system and they obviously believed I had the smarts. It helped too that a professor from the science faculty was also kind enough to give me a recommendation.
“This first job was very important in my future thinking as I was looking at cash over digital payments. Especially in the light of the Wassenaar Agreement from 1998.”
For those of you, myself included, who have not heard of the Wassenaar Agreement, it was an international accord between 33 nations to limit the export of cryptographic materials. It was created to prevent terrorism but acted as a barrier to collaboration between cryptographic sciences creating in-country silos of knowledge.
Accordingly, Anish moved to London where he began an MSc in Information Security where his thesis was on ‘Attacks on Elliptic Curves’. For the purpose of educating myself I sought out a definition of elliptic curves and came across this:
In mathematics, an elliptic curveis a plane algebraiccurve defined by an equation of the form. which is non-singular; that is, the curve has no cusps or self-intersections
Recommended for You
I must confess to teenage humour and chuckled at the ‘non-singular’ nature of these curves. However, as much more intelligent readers than I will already know, elliptic curves are used in cryptography and, in particular, is part of the Ethereum development.
Again let me dip into a definition to better explain:
Elliptic Curve Cryptography (ECC) is a type of public key cryptography that relies on the math of both elliptic curves as well as number theory. This technique can be used to create smaller, faster, and more efficient cryptographickeys.
Afterwards, Anish studied a Ph.D. where he researched authentication protocols.
Anish was a member of a newsletter mailing list where he first came across the Satoshi white paper and was immediately in thrall to the possibilities of the exchange of value.
“From my first job with Ericcson and now working with Cap Gemini, payment systems fascinated me. Cash is not trackable unless you take down the serial number of each note but that is laborious and not very practical. Payment systems, on the other hand, allow for no privacy. Credit cards are always traceable. I found the new cryptography system super interesting.”
In 2012 Anish joined a few of the other Hackerspace members to run a workshop on Bitcoin. “I have to say that despite my intrinsic and deep love for cryptography I am unbiased in that I hold very little coin. This allows me to be impartial and truthful in my thinking.”
In 2012/13 Anish got more involved. He attended Singularity University, an immersive experience where 80 carefully selected individuals merge for 10 weeks of 24/7 learning at NASA Research Park, the global epicentre of innovation, and while there met Stefan Thomas, CTO of Ripple. He joined the team as an advisor and stayed for three years.
Now, Anish is blockchain protocol agnostic and for the past two years has worked as a full-time consultant in the blockchain space. He also gives talks, attends panels and workshops, almost 60 alone across the UK and Europe. So far he doesn’t charge a fee for his talks as he is keen to educate as many people as he can as fast as possible.
He has four main areas of interest: he is fascinated by scaling in protocols, has interests in formal verification of smart contracts, is obsessed with zero-knowledge proofs, and he adores crypto-economics.
When he talks, he enjoys discussing the ethics of AI.
“I was recently in Berlin where I spoke at a conference looking at the intersection of Artificial Intelligence, Machine Learning, and humans. I believe a Singularity is coming soon. I am a Singulatarian. I am curious about the intersection of humans with technology. If ML can ensure that AI systems get smarter and smarter, this has implications for humanity. We really need to address this now and not at the point of Singularity.
“Is it possible for the peaceful coexistence of humans with machines that are far smarter than us?
“We seem to be stumbling into disasters and mistakes. Consider the rise of Facebook, the role of Cambridge Analytics in the US presidential election and even Brexit. There seems to be a complex lack of ethics in these considerations. Then look at the DeepMind, a wholly-owned subsidiary of Google. They have not had any disasters so far but I would not have the NHS transferred into a private ecosystem where something like DeepMind has access to all my medical data.
“I am worried.”
Anish points to the huge advances in previous years: the cloud, drones, big data and then blockchain.
“As we approach FinTech 3.0 I feel that blockchain will have a positive role in managing our Singularity. It may be the guarantor of AI.
“I don’t believe in the zero-sum game. If I win, then I want you to win too. It shouldn’t be a them orus competition.”
While Anish is worried about some aspects of the future and the role that bad actors may play, he is also confident that there is enough enthusiasm to work together. “But it is not enough to put one’s head in the ground and ignore the collision of humans and machines.
“If they [the machines]can be smarter than we need to be too.”
Explanation Point on Zero-Knowledge Proofs
Anish was kind enough to use a simple explanation of this. If I show you a room with only one entry point, the door, and I place you inside that room. And if I appear next to you, on the other side of the door, clearly no longer inside the room and not having used the door to exit, then you have the proof that I can escape from the room, while you have no knowledge of how I achieved it.
And he uses this in his startup to provide protocols which provide assurances of how assists are valued in an open market without compromising privacy.