· the fast track towards algorithmization ·

Fighting the hype by understanding the basics.

Welcome to this insightful area of our website. Beware, you are stepping into the dark side of tech, ML and specific domain knowledge. All together - i.e. utterly dense. That's the bad news. The good news is that we are mixing them all at a newbie level - i.e. once you have read a few of these zoom-outs you'll be able to eventually click concepts in a natural manner and start challenging both your business standards and the hype. You can then revisit the posts in order to understand the whole thing - there is a lot to process behind-the-scenes. This is, spend some time with each paragraph but just let it flow.


Let's start with this eloquent post from Sergio. He first introduces the inputs - a chunk of concepts based on his own hands-on experience across ventures and pro-bono. It is an excuse to open up the scope. Then, he goes for the output - the order, the overall relationship across the hypes behind the former (digitalization, algorithmization, federated algorithmics, blockchain and the metaverses)..

Zoom-out, it all makes sense: vol. 1

The order, the overall relationship across current hypes

Vol. 1

The order, the overall relationship across current hypes.

April, 30th, 2022

I've been radio silent for a while. Utterly busy with the many things we are doing across ventures, pro-bono and academia. Doing so much at the same time has its cons - you can literally go crazy - but also its pros - you have quite a cross understanding of what is going on for real in the world around tech advances. Let me try to explain some conclusions through a probably eclectic and surely lengthy article.

First, here you go some of the conceptual inputs extracted from each of above tasks. 


INPUTS

[From my ventures]

1. SciTheWorld: it is what I like calling my Centre of Excellence (restructuring the company at last to isolate innovation efforts). Some people like buying luxurious cars or houses when they do well professionally. I went for this toy though. It is far more expensive but also quite an asset to (try to) keep moving the frontier of knowledge (i.e. I bought a brain's gym). And, when you have a badass team like ours, far more rewarding. What have we been doing for the last 6 years in here? We did set out a roadmap of the technology that we wanted to create and we boldly went for it - hey, bootstrapping our own resources which is even bolder in innovation! We are fairly ambitious when it comes to innovation so we demanded a lot to our platform: to cope with bleeding edge algo trading strats; to include cyber security by design; to be transparent to experts and regulators; to be resilient to employees rotation (quite a massive issue in the tech industry nowadays); to be fractal (maximal synergies exploitation) and, here the thing today, to be hybrid in terms of IP. Why the latter? We believe that, going forward, companies can only expect to be competitive by teaming up. Through strategic partnerships. They are rapidly understanding that they have to become algorithms (web3, blockchain and the metaverses are merely yet another boost towards their Algorithmization) and the most advanced digital professionals are currently pushing for this. They want the best possible partners, often niche, to be fully on boarded in their processes. But, for that to happen, in order to see full collaboration across companies, one has to be able to grant that each of the partners' IP remains confidential. The future is built on partnerships not on mergers. Why? Often, the value of a company is larger as a provider of different companies than as a fully dedicated resource to one of them - we ourselves turned down promising proposals due to this; due to the fact that we can produce more value free than locking ourselves within a laser focused company, no matter how much of a leader of its vertical it is. We expect partial buy outs, if anything, to be optimal as a way to align interests and improve trust yet not complete ones. But back to the tech: the thing is that, without noticing it, with our approach we were reaching the forefront of Federated Algorithmics in a very convenient plug-and-play way. It was quite an investment but, just as when we created Avatar Calibration back in 2012, we got the tech ready before the hype. Federated Learning is a branch of Machine Learning that, for a number of years now (UCL being a big player), has been focusing precisely in letting algorithms to be broken into counter parties that can't share, mostly, data (where, pretty much, data = IP). Wouldn't be surprised if this field becomes the next big thing. But, beware, probably it'll happen behind-the-scene for a long time. Why? Not many companies can do this. It requires more judgement than budget when deploying it ad-hoc to different clients' idiosyncrasy. And, as a result, marketing won't be devoted as aggressively to it as to, say, Digitalization. Quite a place for underdogs. "Interesting but... how can I be so pumped by this data protection and by blockchain at the same time? Isn't blockchain precisely underpinned by maximal data share?" Good point, let me go through a few more things first.

2. Forctis: this is a blockchain company based in Switzerland where I am board member and equity holder. Its founder asked me for help long time ago and we ended up becoming friends. The point about this task is that, through them, I am being a privileged witness of the evolution of the businesses around this tech ecosystem - I started crafting my academic view at UCL several years ago but this is almost orthogonal information. Too much to be said about the pros and the cons of blockchains but let me highlight only one thing here: the public ones and the business around them are, pretty much, B2C thus it all boils down to the size of your pockets for marketing. This is, don't you think the ones that are killing it are the best tech out there. In fact, most of them lack a tone of much needed features. From those related to efficiency and crypto value stability (whether natural or manipulated volatility) up to those that would allow precisely the opposite move to the digitalization of real assets, i.e. the realization of digital assets, which has been the target of such company since the beginning - and the angle why I started embracing blockchain at the beginning. And that leads to the relevance of the NFTs, by the way. Unique objects of value transacted peer to peer (well, there's more often than not a middlemen though, typically, a broker). The beauty of having organized micromarkets per object (where object can be anything... including, as proposed in the earlys 2010s, your own avatar) is to me quite the core disruption here. But transactions in the real world as a consequence of digital ones is still very far from the current state blockchain ecosystem, so... what do I see in it nowadays? What is the key pro in blockchain which goes far beyond the technology itself? There is a reason why blockchain unlocks innovation that otherwise would never happen...

3. Quantum Mads: this is a quantum computing company where I am board as well. It has been founded by brilliant youngsters surprisingly influential across their peers (PhD candidates in quantum computing) - from Europe all the way to China! I love the way they cut the usual bs from part of the quantum industry and seek only realistic approaches upon this innovative technology. They are focused on finance and, you may think, as such they could be considered a competitor of SciTheWorld (very strong in finance and energy). But they are not. Why? You guessed it: because of the spirit behind Federated Algorithmics. I am sure that I will be able to onboard them in our platform when a client needs to achieve a faster optimization at a project. This is, they can become a turbo boost of our smart tech. Note, by the way, that quantum computing is not smarter than other technologies. It is faster. It can leverage brute force in the calibration of algorithms at an unbeatable pace. Basically, academia tries to find the quantum twin of current models (say, PCA or neural nets) to run it crazily fast - and sometimes, that's crucial. But you need first to find the right model to solve your problem and that requires a completely different set of skills evolved at the non-quantum world.

[From my pro-bono]

1. GPAI: it is a global project around AI hosted by the OECD. There are different working groups fairly diversified internationally so that we get a general overview of how things can be improved beyond each one's frontiers. I have been involved in different projects, mostly: IP protection, AI regulation, AI in SMEs and AI in agriculture. As you can imagine, highly aligned with what we do at SciTheWorld hence, it doesn't take much extra time for me to be able to add value - actually, that's quite the trick behind most things I do, really. Among other things, I see it as my opportunity to try to erode the usual bias against methodologies like, for instance, unsupervised learning - very much at risk of being regulatory forbidden by default even for cases where it can be a good reinforcer of pattern detection upon supervised techniques (or even the best solution when there is no other valid tool available). What am I concluding, overall, across all this projects? Interestingly enough, I am arriving to the conclusion that all I thought it made sense a few years ago, remains making sense: it is difficult for SMEs to leverage AI and to face its regulation. But they are the real deal going forward, they can't be left behind. The way to minimize their burden implies structuring the info for them as much as possible and helping them minimize their responsibility in terms of data, cyber security, etc. Hence, the aforementioned equilibrium between bringing the data in vs pushing the algo out becomes crucial again - i.e. as said, Federated Algorithmics... quite the new asset.

2. Alastria: it is the Spanish blockchain consortium. I am somehow leading the Algorithmics Horizontal (A.I.) in the attempt to show cases where blockchain technology unlocks projects that otherwise would have not been done. Dependent on the complexity of the project overall it will require us to use more or less of a Federated Algorithmics platform and, within it, more public or more private blockchains. There are a tone of projects to evolve - the issue is convincing key players to collaborate with the required level of effort. How can different agents trust a project that requires each of them doing things right upon their own, protected assets? (where protection includes both, IP and regulation). Not easy, as it really needs sophisticated and ad-hoc tools, but doable as we'll briefly see below. And, if achieved, it can unlock a whole new level of impact - e.g. imagine a project of cyber security or AML collaboration across companies in the seek for hacking patterns across them instead of within them (my first proposal).

3. Madrid blockchain cluster: it was formally announced very recently. So humbled to have been selected for this. So far, my conclusion is that many regions in the world are taking this field very seriously - just as the traders' say goes: they are putting their money where their mouth is. There is a fierce competition not to be left behind - nowadays you need to compete aggressively to survive... quite a new motto at many verticals. Why? Whether your sparks of innovation are right or wrong they can surely trigger a solid tech industry around them. I've been pitching this benefits already for a while in Spain, along with many others, and I love to see significant results, finally. That is the real value to me from the different regional efforts around the world - the true plan is not the product but its byproducts, its positive economies. Another interesting thing is that Madrid will create some more clusters which will be largely connected to this one. In especial the metaverse cluster (or metaverses). Why? The metaverse is fairly similar to online video games that leverage blockchain-like tech for the aforementioned objects of value (which can range from features to help from gamers) to be transacted easily. Again it is the easy bit (that spans through exchanges well beyond the video games themselves) that triggers the possibility for new industries, brokers free, within web or video gaming to be unlocked. Imagine YouTube without YouTube... meaning, no change of fees policies nor algorithms whenever the company decides to (there was quite a polemic some years ago around that). Or imagine your avatar working 24-7 for you in marketing campaigns. Whether the experience is immersive or not is just a matter of UX and channel.

4. Several more pro-bono projects: given the current circumstances we are helping fight phishing, along with quite a bunch of friends, through AlertaHack (we are just launching it for free; it is fairly unique and it was designed for insurance companies... stay tuned); also trying to educate people around tech advances and risks by collaborating with dynamic journalists from prime time Spanish TV news; and a number of other projects that keep adding judgement to where we are and where we are going. Where are we now? In a world that seems utterly complex but it is not - it is just slightly different. Where are we going: we are going to exploit synergies, at last.

[From academia]

I like testing myself at deep level discussions on hot topics at tier one universities (UCL, Oxford, IE...). Specially, dealing with PhD students and top end researchers (obviously, mostly from UCL). It helps me audit how robust I am at my thinking - interestingly enough I often include in this bucket national (govs and central banks) and supranational entities as the discussions can become very similar. And, along these years, I've found out how rare it is to have bottom-up built, zoom-out views. And how valuable - C-suites are willing to pay utterly well for this type of consultancy. And how demanding as well - sometimes, the level of zoom-in required to discuss with different agents is the part that allows me gaining their respect (it is easy though, just saying you don't know when you don't know and discussing the rest, saves you a lot of trouble). And that's the beauty: when you can convince of the zoom-out upon ad-hoc zoom-ins. Note that storytellers typically lack zoom-ins and that leads to not insightful zoom-outs - and also note that I am not against storytellers as they have a very healthy role unlocking brainstorming. And this important. But, what can you do to avoid confusing brainstorming with insightful information when you are not a master of the zoom-ins yourself? Easy, I'll finish with that.


OUTPUT

And now the output. Very quickly, because I want this all to be just a bold driver for you to keep thinking, to keep allocating concepts orderly. As you can imagine, being this, as said, eclectic, there is a tone of fine details that could well take a couple of handbooks to polish but we'll have to live with that as I am seeking eloquence here.

By the way, this is the type of messages that we convey to our clients when we organize events at our headquarters - which, actually, are as cool as it can get as we mix C-suite and board members from different blue chips with coders and mid-managers... - hence, I may be able to write a few more of these zoom-outs around topics like robots advisors (not roboadvisors), (real) exponential innovation, algorithms black-box auditing, green algorithms, innovation entrepreneurship... I don't know, this type of things.

Anyhow, there we go:

1. Innovation within departments: digitalization

2. Innovation across departments: algorithmization

3. Innovation across companies: federated algorithmics

4. Innovation across countries: blockchain

5. Innovation across cultures: metaverse(s)

(they kind of click like the dolls of the picture...)

Where:

A. Digitalization is the process by which a company starts accumulating data and trying isolated and rigid use cases within departments. Silos are perpetuated and there is focus on the L in ML. Burdens, in the shape of legacies, follow.

B. Algorithmization is the process by which a company starts transforming its activity through new protocols upon a (deep) tech platform. The whole business becomes an algorithm across departments. There is maximal synergies exploitation and a focus on the M in ML. The L can subsequently reach further upon the well thought M (again, ours took 6 years to start with). And quantum computing can give a major boost to the calibration of part of its multiple pieces - which otherwise can be utterly slow to optimize even in parallelized cloud structures.

C. Federated algorithmics is the process by which several companies collaborate to create a super-algorithm which leverages their main IP assets without having access to them in a raw manner. There is no one-size-fits-all solution, it has to be done in a very bespoke manner upon the right platform (algorithmic trading seems a good seed for this second-to-none challenge). When there is no trust between the stakeholders around whether each one has played its role in the overall calculation of the algorithm, a blockchain technology (typically, private) can be used to prove (value or stake) that a stakeholder's (previously audited) routine has been correctly triggered. Talking about IPs immediately implies black-box auditing... we are, for instance pioneers on black-box auditing for trading algorithms so we do know that there should be space for more innovation here and for many more agents to join. Especially, since more A.I. regulation is round the corner (the business can only grow).

D. Blockchain is a tech that solves for poor trust in a P2P transfer of value (or execution of routines). It is often described as efficient and cheap technology. But the truth is that it probably couldn't be less efficient nor more expensive (and even contaminating) as it basically is massive redundant technology. And yet it is true that it can generate efficient and cheap industries. How? Because it does not need to ask for people to gather across countries and convince them all in order to improve the efficiency of an entire industry. I love that and I think it is its real value. What do I mean? Well, for example, the Spanish payment ecosystem is utterly efficient/digital and yet many other countries in the EU (the same region!) are not that efficient? Why? Easy, have you ever tried to convince someone to do something that takes extra efforts to her duties? In Spain we were lucky that there was an overall pioneer digital appetite in payments. Everyone agreed at the table because it was part of their duties. It seems banks in other countries were not on the same page. Now, would you imagine doing that across countries so that an entire industry becomes more efficient? Here is the behind-the-scene magic of blockchain. Blockchain, with its major momentum, basically has ended up saying: "hey, you in or out? ah, and by out I mean out of business". And, all of a sudden, all those change resistance guardians are all in. Globally. Surely, there would be better solutions for most of the disruptions that public blockchains will bring - brute force, a corner solution in mathematics, is rarely the best solution. But the tech still accounts for two key dimensions to sneak in efficiently. First, the private blockchains which I am sure will gain massive relevance in the next 5 years as they count on the flexibility and efficiency of ad-hoc SLAs (Service Level Agreements) around its nodes ecosystem. And second, the modularity of its business ecosystem: brokers, exchanges, custodians, identity... It is true that those companies are not visionaries (they are not truly inventing anything at all) but they are trying to make sure they are all seamlessly connectable between them. Making sure they are Lego pieces. And that, openness by design, is a major beauty. We will see more innovative pieces being clicked going forward - e.g. don't we need central bank-like stability control for many cryptos? And yet, it is true that we have not seen as many use cases as we wished so far? Why? Speculation is running the show in the low hanging fruits space. But there is much more to be done once the speculation gets eroded/regulated. How? Leveraging the positive side of the speculators as the main asset of blockchain, no doubt, are the crypto investors themselves. Those who sustain the momentum no matter the success cases so far - they are willing to take massive risks backing up projects that otherwise would never reach any impact.

E. Metaverses are a wrap up of the former upon engaging UX, with a major idiosyncratic feature: everything is digital native there. Note that it is again a low hanging fruit when compared to the realization of digital assets. Thus, not my ideal usage of blockchain yet. But very interesting, nevertheless. Its beauty is that, taking into account that web3 allows for easy transfer of value, P2P, a lot of new business opportunities will arise across the metaverses. And those opportunities will be for avatars. Not people. It kind of resembles pure meritocracy from many angles as therein your real origin, accent, image, etc won't count much. You have to be able to add value and that's it. I love that - it boosts the upside behind the Globalization that we have witnessed since the 2000s. But there are also scary things. To start with, being it fully digital, it is going to be easy to get all your data around your digital life to track you like never before. Upon all your digital footprint tones of increasingly accurate digital twins of your can be created - if you think of it for a while it can get very creepy. I can't stop thinking that we better take care of all that from inception...

... And it is such an attitude that brings me back to more investigation and more tasks across ventures, academia and pro-bono in a rather endless loop that I have been chasing for almost 15 years now. Time flies.


So there you go for now. These are our bold thoughts about how all those different innovation approaches are connected (from business departments to entire cultures) and why we think we are walking towards massive efficiencies in the future. When zooming-out you realize that everything has its own space - most of the advances complement each other and radical supporters or haters are just ignorant of the pros and the cons of each tech

The sweet spot, the magic, is the area where them all are smoothly combined; where they all click - which we will see in the rest of the pills.

So, there you go, long but not that eclectic article after all! Thanks for reading (this long...) and no worries - next one is much lighter.

Built with Mobirise free builder