Nick Bostrom "superintelligence" book review


Thu 24-Oct-2018

There are few things less fashionable than reading a book that was all the rage 2 years prior. One might as well not bother – the time for all the casual watercooler/dinner party mentions is gone, and the world has moved on. However, despite tape delay caused by “life” and with all social credit devalued, I decided to make an effort and reach for it nonetheless.

In terms of content, there’s a lot in it – and I mean *a lot*. Regardless of discipline, in non-fiction a lot can be a great thing, but also challenging (one can only process, let alone absorb, so much). However, a lot in what is essentially a techno-existential divagation is like… really a lot.

For starters, Bostrom deserves credit for defining the titular superintelligence as “a system that is at least as fast as a human mind and vastly qualitatively smarter”, and for consistently using the correct terminology. The term “AI” – as is increasingly called out these days (“Ultimately, the term artificial intelligence may be a misnomer. To be sure, these machines can solve complex, seemingly abstract problems that had previously yielded only to human cognition. But what they do uniquely is not thinking as heretofore conceived and experienced.”) – is routinely misused and applied to machine-learning and / or statistical systems which are all about pattern recognition, statistical discrimination, and accurate predictions; none of which are close to AI / AGI (artificial general intelligence) proper (ironically, as I’m writing this Microsoft’s AI commercial ft. Common – which conflates AI, VR, AR, and IoT and throws them all into one supermixed bag – is playing in the background; and btw, Common? Whatever made Satya Nadella choose a vaguely recognizable actor when he could choose and afford a great actor? Or a scientist for that matter?).

Anyway, back to “superintelligence”. One of the most repetitive and profound themes in Bostrom’s book is that we cannot predict nor comprehend how an agent orders of magnitude smarter than the smartest of humans will reason; what goals it might have; how it might go about reaching them. It’s kind of a tautology (“we cannot comprehend the incomprehensible; we cannot predict the unpredictable”), but still, makes one think.

Separately – and it gave me pause many times as I was reading it – the word “privilege” is used a lot these days: male privilege, white privilege, Western privilege, straight privilege etc. etc. These privileges are bad, but I think most of us humans – inequalities notwithstanding – is used to and quite fond of the homo sapiens privilege of being the smartest (to our knowledge…) species on Earth. “Smart” may not be the most important attribute for everyone (some people bank on their looks, others on humour, others still on personality, others on love), but I think few people don’t like to think of themselves as broadly smart. Some – myself included – chose to bank most of themselves, our lives, our sense of self, our self-esteem on being smart / clever / educated. What happens when the best, smartest, sharpest, wittiest possible version of me becomes available as an EPROM chip or a download? If ever / whenever the time comes when my intellect becomes vastly inferior to an artificial one, how will I be able to live? What will drive me? I don’t want to be kept as some superintelligence’s ******* pet…

The amount of ideas, technologies, and considerations listed in Bostrom’s book is quite staggering. He’s also not shy to think big, really big (computronium, colonizing the Hubble volume, lastly – what if we’re all living in a simulation in the first place?) – and I love it (the Asimov-loving kid inside me loves it too). Separately though… Bostrom seems to be quite confident that the first superintelligence would end up colonizing and transforming the observable Universe (there would be no 2ndsuperintelligence… and even if there were, there is only one Universe we know of for sure). However – as far as our still rather basic civilisation can observe – the universe is neither colonized nor transformed (unless we are all living in a simulation, in which case it can be). Has the path not been taken before…? To be the first (or only) civilisation in the history of the universe capable of developing AI sounds like being really, really lucky… *too* lucky almost. Then again, it may be the case of trying to comprehend the incomprehensible and predict the unpredictable

It may not be the best written book ever, but the guy did his homework and knows his stuff. Separately, the author deserves credit for not looking at AI in a technological silo, but broadly: from neuroscience, through politics, all the way to philosophy and ethics. For someone who’s a big believer in the future being more interdisciplinary (i.e. myself), that’s confirmation that having wide and diverse interests is worthwhile.

Reaching for hit books is a little bit like reaching for hit albums (regardless of the genre) – sometimes great ones deservedly become hits, and sometimes substance-less **** ones undeservedly become hits all the same. However, despite attaining recognition similar to “4 hour week” or “blink!”, “superintelligence” does actually have substance – and plenty of it. Along with substance comes a certain challenge in reading and following, which makes me wonder how many people who bought the book bought it to read it, and how many merely bought it to be seen reading it in public? (much like this). The substance of “superintelligence” can actually be really overwhelming – not just in terms of mind-boggliness of its content (although that too), but purely on volume. There is a lot in it – and I mean *a lot*. In his efforts to cram as much substance as possible, Bostrom forgot that in order for a book to be great it needs to be – has to be – well written. “Superintelligence” is a lot of things, but well-written it, unfortunately, is not. The first hundred pages are particularly tough to get through – the author could have easily trimmed them to 30 – 40, making the material more concise and comprehensible to a regular reader (it is, after all, meant to be a “popular science” book, not Principia Mathematica – Joe Average should be able to follow it). Yuval Noah Harari’s “homo deus” is a great example of a book that’s got substance, but reads really, really well (on an unrelated note: YNH is a much better auteur than he is a public speaker). Nassim Taleb’s “black swan” is another (though Nassim is all about Nassim more than Kanye is all about Kanye – it does get old real quick).

On top of that, AI occupies a unique place in the zeitgeist. On one hand, FT (“Artificial intelligence: winter is coming”) rightly points out that “We have not moved a byte forward in understanding human intelligence. We have much faster computers, thanks to Moore’s law, but the underlying algorithms are mostly identical to those that powered machines 40 years ago. Instead, we have creatively rebranded those algorithms. Good old-fashioned “data” has suddenly become “big”. And 1970s-vintage neural networks have started to provide the mysterious phenomenon of “deep learning”.”; on the other Alan Turing’s Institute notes that, “artificial intelligence manages to sit at the peak of ‘inflated expectations’ on Gartner’s technology hype curve whilst simultaneously being underestimated in other assessments”.

Consequently, in the end, I was struck by a peculiar dissonance: on one hand, reading the book (which is measured and balanced, it’s not unabashed evangelising) one might get the impression that the titular superintelligence really is inevitable – that it’s a matter of “when” rather than “if” (with the entire focus being on “how”), and that the likelihood of this becoming an existential threat to humanity is substantial. Then, around page 230, Bostrom gives the readers a bit of a cold shower by making them realise that it’s essentially impossible to express human values (such as “happiness”) in code. And then I’m left agitated (the good way) and confused (also the good way): inevitable or impossible? Which one is it?


PS. Not as an alternative, but as a condensed and very well written compendium, I cannot recommend this waitbutwhy classic enough.


Royal Institution: How to Change Your Mind

Mon 11-Jun-2018

The innocent and perhaps slightly dull title hid one of the more exciting RI events in recent weeks: an acclaimed author, “immersive” journalist (“when I was researching the food industry, I bought a cow”), and a UC Berkeley professor Michael Pollan was talking about his most recent book (carrying an innocent and perhaps slightly dull title “how to change your mind”…) about psychedelic substances, in particular psilocybin and LSD.

Psychedelics are currently enjoying a true (academic) renaissance. After half a century of complete obscurity (not to mention criminalisation), they are resurfacing in respected universities’ research facilities (subject to unbelievable admission rigours and control protocols, not to mention procurement), and receiving strong support from foremost academics as potential game-changers in psychiatry as well as cognitive performance enhancers. Ongoing legalisation of cannabis and shifting public perception of drugs’ harmfulness may be the catalysts.

Among the world-class academics leading the debate and the research are Prof. David Nutt of Imperial College London, who published a now-famous (re)classification of drugs based on their harmfulness in 2009, and was sacked by the Labour (sic!) government shortly thereafter, and Prof. Barbara Sahakian of Cambridge University, who – among many other things – is researching performance enhancement and “smart drugs” (you can buy her book on the topic here and read a fascinating interview here).

With mental illness spiralling, a prospect of single (supervised) application of LSD or psilocybin generating results comparable to *years* of traditional therapies was probably what interested the scientific community (you can watch Prof. Nutt’s presentation given at UCL’s Society for the Application of Psychedelics event in Nov-2017 here, it’s fascinating – you can hear me asking his views on legalisation towards the end). Silicon Valley – in its true “quantified self”, entrepreneurial, and somewhat quirky style – responded to “performance enhancement” (the concept of performance-enhancing “microdosing” was brought into public’s attention in a famous Sep-2016 Wired UK article you can read here; another well-known feature is the Aug-2017 article in the FT you can read here.)

Against this backdrop, any presentation on LSD and / or psilocybin is pretty much a guaranteed full house*, and this one was no exception. There were unusually many first-time-ever attendees (RI did a quick show of hands), many of which were friendly, Big Lebowski-style dudes with very long hair and an aversion to footwear. The event was MC’d by Dr. Robin Carhart-Harris, who is one of the rising stars of the younger generation of psychedelic research (you can watch him give his own presentation at UCL’s Society for the Application of Psychedelics event in Feb-2018 (a sequel to Prof. Nutt’s event) here).

Michael Pollan proved to be an experienced, engaging, and charismatic speaker, which definitely contributed to the quality of the event. In his presentation he talked about his (informal, and, by his own admission, at times not entirely legal) research and experiences of psychedelic substances, as well as experiences and views of other people who – for very different reasons – tried psychedelics.

The focus of his talk were mental health and related experiences, not on performance enhancement, which made it a little more profound (particularly the part about terminal patients confronting their fear of imminent death) than it would be otherwise. He stated, matter-of-factly, that the single guided LSD experience he had under the care of an experienced therapist (who takes huge personal risks by running such clandestine sessions) did for him what would take conventional therapy years to accomplish.

Another interesting point raised by Michael was the deep sense of closeness and togetherness with nature being the result of psychedelic experiences. I don’t recall this argument appearing in the public discourse, usually it’s just depression and performance enhancement, which made it all the more interesting.

The event went well into overtime with extended Q&A (you can hear me asking his views on legalisation towards the end) and will hopefully be uploaded to Royal Institution’s YouTube channel.

You can see upcoming RI events here, and make a small contribution here.

Royal Institution Patron’s Night: ExpeRience: A space odyssey


June Patron’s Night event was dedicated to the legacy and themes raised in the cinematic masterpiece 2001: A Space Odyssey.

The movie, due to its many exceptional qualities (stunning in-camera visuals, aesthetics, striking technological predictions (see astronauts’ using tablet computers in a movie from 1968), the rogue AI, lastly its philosophical layer) has never been entirely out of the popular discourse, but as of late it’s enjoying additional resurgence due to its rogue AI, HAL9000, appearing in each and every single conversation about AI in general, AI ethics, and AI threats.

The event was focused on some of the other themes of the movie, such as space exploration in general, space robotics, hibernation and cryonics, and the psychological aspects of space travel. RI always has great speakers, and this time was no exception.

The panel included:

  • Marek Kukula, Public Astronomer at the Royal Observatory Greenwich, as the discussion host and moderator
  • Calum Hervieu, space systems engineer with a focus on lunar exploration mission architectures and crew engineer for a Mars mission simulation in Hawaii.
  • Professor Yang Gao, Associate Dean for Faculty of Engineering and Physical Sciences (FEPS) and the Professor of Space Autonomous Systems at Surrey Space Centre (SSC) and the Director of EPSRC/UKSA National Hub on Future AI & Robotics for Space (FAIR-SPACE).
  • João Pedro de Magalhães, Institute of Ageing and Chronic Disease, University of Liverpool and coordinator of the UK Cryonics and Cryopreservation Research Network (, which aims to advance research in cryopreservation and its applications.
  • Iya Whiteley, Director of the Centre for Space Medicine at Mullard Space Science Laboratory, UCL. She is a Space Psychologist, who worked on developing Tim Peake’s astronaut selection training programmes.

If the academic cred of the guests weren’t enough, Iya’s hobby is competitive skydiving, João is an amateur stand-up comedian, and Calum recreationally runs half-marathons. Try to compete with that.

The event paid tribute to Stanley Kubrick’s masterpiece and used it as a starting point for speakers’ individual short presentations. Rogue AI wasn’t one of them, and that’s a good thing, because there were enough interesting topics to discuss without it. Calum compared the scope and difficulty of lunar missions with the potential Mars mission, making it very clear how much more challenging the latter will be (if it ever happens…), and showed some concepts for human habitat on Mars (kudos to Foster+Partners for thinking about extraterrestrial markets so early in the game). Iya talked about the unique and numerous psychological challenges and considerations of a long-term (here defined as going to Mars and back, a couple of years altogether, including a minimum of 220 days in the space craft each way) space mission, some of which are difficult to relate to for ordinary people – e.g. the lack of sensation of wind or slightly varying temperature on one’s skin. Yang talked about UK’s ambition of becoming a leading player in space robotics, and the drive to enthuse the youngest generations to space exploration and robotics. João talked about hibernation and cryonics as they occur in some species of mammals and amphibians, and about ongoing research into both hibernation and cryonics of humans (mentioning Alcor, the cryonics company which first appeared on my radar in unforgettable “Future Fantastic” series hosted by Gillian Anderson many moons ago).

Presentations were followed by the usual Q&A, which was then followed by hands-on experiences, which aren’t usually the part of RI events. Attendees could see the impact of micrometeorites on space crafts, experience firsthand (literally) different rocket fuels, see what would happen to an unprotected human in the vacuum of space, or try the zero-G simulation chair. My favourite one were the rocket fuels, presented by charming and beautiful robot lady

Loved it!

Nesta FutureFest: Will my job exist in 2030?

Fri 06-Jul-2018

Hosted by Nesta’s Eliza Easton and Jed Cinnamon, the “will my job exist in 2030?” session took place at high noon of what was probably the hottest day of the year, in a modest size auditorium, under a glass roof.

To say that it was hot would be a misunderstatement: it was boiling, it was a sauna, it was Dubai. And still, the auditorium was *packed*. I know why I braved it, and I suspect everyone else in there braved it for the same reason: as white collar professionals we are mortified about the prospect of being rendered obsolete by an algo, and we want to find out:

  • How likely it is exactly
  • What (if anything) we can do about it

While the presentation echoed Nesta’s report bearing the same title, it was entirely self-contained and, as far as I could tell, told from a slightly different angle than the report.

Part of the presentation was delivered from the perspective of educating and training today’s middle- and high-school students (so only tangentially relevant to grown ups), though the skills and competencies listed as enhancing the employability of modern day teenagers’ were definitely good to know for everyone (along the lines of: do I have this? Could I plausibly argue that I have this? What can I still do in order to have this? How can I rephrase my resume in order to say I have this?). The list included:

  • Judgement
  • Decision making
  • Complex problem solving
  • Fluidity of ideas
  • Collaborative problem solving
  • Creative problem solving
  • Resilience
  • Critical thinking

I think that list alone made it worth attending the event, because the skills and competencies listed above are indeed in the Venn diagram sweet spot where analytical, creative, interpersonal and imaginative overlap; in short, those (at least in 2018) appear to be abilities relatively most difficult to automate.

Separate consideration was given to the interdisciplinary. The hosts mentioned how important it is to hone creative and artistic skills alongside modern STEM curriculum, as well as remove barriers between different, previously silo’ed disciplines (where Finland’s success was used as an example). This point echoed with me quite strongly because only a few years ago my own interdisciplinary approach towards career development (defined as pursuing a relatively wide variety of roles based on how interesting they appeared rather than how closely they aligned to my direct experience) earned rather limited support and understanding of my London Business School colleagues, most of whom spent their professional careers within one specialism (equity sales; credit derivs etc.) and were all about climbing the ranks of higher and better paid positions within their respective specialisms.

There was a somewhat fresh (and refreshingly sane) angle on the nemesis-du-jour topic of automation. While the authors echoed the general ennui of many jobs currently done by humans being automated away, they thought their conclusions over more thoroughly than the prevailing “unskilled jobs will all go, and many skilled jobs as well” and pointed out that some of the lower-skilled jobs are not only relatively safe, but also likely to grow (examples listed were agriculture and construction; I would also think that all types of non-specialised carer jobs were in the same category; basically jobs requiring body mobility and dexterity and / or emotional connection).

Another original observation was that while there is a continuity of certain occupations and job titles, the day-to-day work itself – and the skillset it requires – may have very little in common over the years (the example given was typesetter: modern day InDesign user vs. a heavy machinery operator from couple of decades ago; finance automatically comes to mind where many roles have much more to do with technology than any “pure” finance).

The final point of relevance to me was raised during the Q&A (by myself…): given my personal experiences with some employers being more open towards the concept of lifetime education than others (with few being actively hostile), I wanted to know whether businesses begin to recognise the value of ongoing education, learning and training of their staff as opposed to the entrenched view of seeing any upskilling as a distraction and / or a threat (basically a variation on: “they don’t need this for their current job… they will want more money and leave”). I was quite happy to hear that there is indeed a certain pivot happening (albeit not very fast) as we speak, and businesses begin to see the (commercial) added value of their employees gaining more skills.

Nesta FutureFest 2018

Fri 06 – Sat 07-Jul-2018

First weekend of July saw the 4th edition of Nesta FutureFest. For those of you who don’t know, Nesta is UK government’s agency for innovation (one of very few government bodies I feel isn’t wasting my taxpayer contributions), and FutureFest is its sorta-annual (taking place about once every 18 months) two-day festival.

The event consists of talks on a number of stages + smaller accompanying events and presentations. The closest reference point to FutureFest I can think of is New Scientist Live, but NS Live is mostly “pure” science: whether it’s quantum physics, CERN, DNA memory storage, or nuclear fusion, it’s mostly just talks on some exciting discoveries, ideas, and inventions. It is thought-provoking but largely uncontroversial and apolitical.

FutureFest is a little different. While the focus is firmly on the future (as the name would suggest), it is broader than science and technology and also considers social, urban, artistic, and even philosophical and religious angles. This holistic scope makes the event really unique.

The 2018 event was held in the Tobacco Dock in London over one of the hottest weekends of the year, which, given that most of the building is covered with glass roof, made attending some of the jam-packed talks (and vast majority of them were jam-packed…) something of a challenge, but trust me, well worth it. Plus the venue itself is really nice too (I believe some of the Wired magazine’s events are being held there as well).

The Nesta team did a brilliant job booking diverse headline speakers to deliver talks across a very broad spectrum of subject I’d describe as not just “future”, but more so “us humans and our future”.

Just to give you a little taste of the diversity of the 2018 talks, here are titles of the ones I attended:

  • The geopolitics of AI
  • Will my job exist in 2030?
  • 2027: When the post-work era begins
  • Digital workers of the world, unite?
  • How blockchain can, literally, save the world
  • The tangle of mind and matter
  • Future humans: augmented selves
  • Let there be bytes
    (I’ll expand on a couple of these in greater detail in separate posts).

In parallel, there was a “meet the author” stage, where I resisted buying even more books I may never get a chance to read (I regretted afterwards), and where I finally managed to talk to my idol Dr. Julia Shaw, and caught up with the inimitable and very, very candid Ruby Wax.

What struck me as surprising was that of the 2 days, Friday was definitely busier and more packed with attendees. I mean… I had to take a day off work to attend, was everyone else there on business? Or was it the World Cup quarterfinal match (England vs. Sweden – we won) on Saturday? In any case, I showed up on both days and had an absolute blast. It was intense and towards Saturday evening my brain was definitely overflowing, but it was absolutely worth it. I can’t wait for the next one.

Mariana Mazzucato: the value of everything

British Library, 09-Jul-2018

In 2018 Mariana Mazzucato is a brand name. I first came across her in the short article Wired’s Ideas Bank, which first made me aware that everything I thought I knew about entrepreneurship and innovation was a fallacy, sold to me by corporate sector’s PR and general “hijacking of the narrative”.

The article may have been short, but the Mazzucato’s point was huge. Then there was the lunch with FT, and then the British Library event. Mazzucato’s “the value of everything” lecture was part of the series organized by the British Library and UCL, and was linked to the release of her new book (also titled “the value of everything”) which analyses how (and why) modern economies reward value extraction and rent seeking rather than genuine value creation.

And let me tell you one thing: did she deliver. With passion, confidence, and charisma Mazzucato was one of a couple of speakers whose presentations I attended in the recent weeks who proved that not only *what* you say, but also *how* you say it really, really counts (others were Ruby Wax, Izabella Kaminska and Eugenia Cheng).

The quote the presentation revolved around comes from Big Bill Haywood, who in 1929 set up first trade union, which reads “The barbarous gold barons – they did not find the gold, the did not mine the gold, they did not mill the gold, but by some weird alchemy all the gold belonged to them”.

What I found very interesting was that rather than focus on purely economic arguments (which could lead to the discussion turning academic, niche, and otherwise boring), Mazzucato, originally, provocatively, even somewhat eccentrically, stated how important storytelling and narratives are in this discussion, and how important it is to contest what is told to the public as the official version of events (Mazzucato quoted Plato’s “storytellers rule the world”; by contrast I feel tempted to quote Reagan’s “government is no the solution to our problem; government *is* the problem”; to be clear, I’m siding with Plato).

Financial services sector took the brunt of Mazzucato’s criticism, with modern politics coming a close second. The criticism of financial services ran along the standard lines of productivity and adding value, while modern day liberal politicians and public sectoras a whole were criticised for not countering the neoliberal, neoclassical corporate narrative. Pharma came third for rigging prices of medicines.
In terms of innovative thoughts and ideas, I appreciated Mazzucato joining the ranks of more and more public figures who fight to recognise user-generated data as something that has value in the economic sense, and therefore something its creators should be compensated for (e.g. in the form of UBI).
On the more iconoclastic side, I was in equal measure surprised and happy to hear Mazzucato predicting a boom (and bubble) in cleantech. I’m all for it. I could not be happier to hear it. That’s one bubble I support in full.
One thing I missed throughout the lecture was Mazzucato clearly defining “value”. She was pointing out may dysfunctional aspects of modern economy (casino banking, rigging medicine prices) and constantly referring to what does and doesn’t contribute to value, but never once had she actually defined it. The question came up in the Q&A and Mazzucato defined it fairly loosely as public purpose- / mission-driven actions, such as, for example, cleaning up the oceans. She also said that “value is created collectively”, which stands in contrast to prevailing individualist approach to entrepreneurship and life in general, so that’s definitely food for thought.

You can watch the entire lecture (87 minutes) on YouTube.

BBC Machine Learning Fireside Chat: The ML Arms Race: China vs. USA vs. Europe

BBC Machine Learning Fireside chat: The ML Arms Race: China vs. USA vs. Europe

BBC ML Fireside chats are a fairly new initiative launched by a group of machine learning specialists at the BBC. They are being held as free panel sessions open to the general public via, where BBC ML Fireside chats have their own burgeoning group.

The 06-Jun-2018 session was the 6th or 7th held by the group and focused on the unsettling similarity between modern day artificial intelligence technological race and the not-so-distant arms and space race between the East and the West. Granted, some things *did* change between the Cold War and the present day, namely that instead of 2 key players (the US and the Soviet Union) we now have 3: the US, Russia, and China, with UK and the EU being additional important players.

The panel featured as its speakers:

Lucy Beard: a dual UK/US national, Lucy is a former actuary who then joined Intuit as a data scientist (“before we all started using the term”), and a couple of years ago had her entrepreneurial epiphany and launched Feetz, the world’s first company producing custom-fitted 3D-printed shoes. Lucy joined via Skype from California.

Charlotte Stix: Research Associate and Policy Officer for AI Policy at the Leverhulme Centre for the Future of Intelligence, University of Cambridge, and formerly a member of European Commission’s Robotics and Artificial Intelligence Unit.
Jeffrey Ding: China lead for the Governance of AI Program at the Future of Humanity Institute, University of Oxford, and formerly an advisor to the U.S. Department of State and the Hong Kong Legislative Council.

The panel was perfectly selected to discuss the complex, politically sensitive topic: Lucy, as a dual US and UK national was in the position to highlight the importance of cultural differences in development of new technologies between the US and UK / Europe; Charlotte, as an EU national living in London was the right person to discuss the impact of Brexit on UK’s science and technology scene; while Jeffrey, as American born Chinese was an ideal person to discuss both the scale of ML efforts by the 2 major players, but also to delineate some ideological considerations, and to highlight scenarios more nuanced than “the winner takes it all”.

The event was attended by approx. 100 guests from different backgrounds (diversity and inclusion side note: overwhelming majority of them were men) and the discussion flowed smoothly and energetically from the get-go. One valid point clarified at the outset was that currently “AI” and “machine learning” generally refer to different forms of intelligent automation rather than to sentient, self-conscious agents the like of HAL9000 or Ex Machina’s Ava.

It’d require a post 10 times the length of this one to transcribe the entire conversation, but the more salient points were:

We may not be in the machine learning “arms race” situation just yet, but if we keep framing the ongoing conversation like that, we might end up with one.
There is a frenzied race for ML / data science talent taking place right now, with Big Tech firms sparing no effort and no expense to compete the little talent that’s out there, leading to a very real “salary race”. (If I may add my own perspective here, this is very similar to what we’ve seen in financial services in early 2000’s, where there was a very similar race for junior investment bankers or credit derivatives specialists, or the race for iOS developers circa 2007/2008. In case of financial services, the demand was just as intense as it was short-lived, and many junior bankers and credit derivatives structuring specialists experienced on their own skins just how unsentimental banks are; in case of iOS developers the supply quickly met the demand, leading to an adjustment in wages – ML, as purely technological, may experience the latter fate; as for data science, the jury’s out, as it’s more interdisciplinary, and not purely technological).
The importance of (often marginalised) ethical considerations.
In light of the increasingly globalised and interconnected nature of computer technology, win-win scenarios are possible (Jeffrey gave an example of Microsoft’s China research hub, where the best and brightest scientists from China work for the benefit of a US company, but on the other hand, gain a lot of expertise themselves and leave to set up some of China’s leading ML companies).
EU does have quality academic and entrepreneurial hubs, developing quality research and talent, but it’s not doing a great job of promoting it and holding on to it.
UK is trying to pre-empt the impact of Brexit and maintain (if not enhance) its position in the ML / AI world by launching numerous agencies and initiatives (author’s side note: one of the examples of recently launched agencies is Fair-Space, whose head, Professor Yang Gao, was a guest in the recent Royal Institution event; one of the examples of recent initiatives are recent report by House of Lords Select Committee on AI or PM’s bold declaration to use AI in cancer detection) while at the same time, it’s facing a prospect of losing hundreds of millions of EUR per year in EU funding.
Some of the big concerns around Big Data and machine learning are in fact very 1st world types of problems – as Charlotte rightly pointed out, a typical citizen of the Global South wouldn’t care how her or his data is being used, as long as they would receive something that would improve their lives (e.g. healthcare) in return.

The session ended with a Q&A that lasted about 20 minutes, followed by networking. Had it not been for the fact that WeWork had to close the building at 21:00, the networking part could have easily lasted much longer.

It is worth mentioning that the events are organised on voluntary basis by a group of ML enthusiasts at the BBC, who put their own time and effort into making them happen. WeWork (which recently acquired in order to enhance the brand) kindly provided the space (and catering! You don’t usually have any at paid events, let alone the free ones), and the result was absolutely great. Big “thank you” to everyone who made the event happen, especially Gabriel and Ahmed at the BBC.

In short: brilliant event. You can join the BBC Machine Learning Fireside chat group on here.

1 reason why I really, really, REALLY hate Bitcoin

1 reason why I really, *really*, REALLY hate Bitcoin

It’s 2018 and Bitcoin is all the rage. At USD 7,000 it is perhaps less of a rage than it was at USD 20,000 in Dec-2018 (although people who bought it then… well, they must be literally “all (the) rage”), but still, this price is nothing short of astonishing.

I am an enthusiast of blockchain technology, but at the same time, I know that BitCoin and blockchain are not the same thing. The latter is, in its core, a concept, a new approach of maintaining, storing, and updating ledgers and books of records. The concept requires tangible technological solutions in order to become a real-world solution, and that is happening as we speak. A number of providers (e.g. IBM, whose solution my former employer Northern Trust used in their innovative blockchain-based solution for Private Equity) offer blockchain products. BitCoin, whilst running on a blockchain network (the first, or one of the first blockchain networks in the world, and by far and large the most successful and best-known one), has no intrinsic value at all (same applies to all other “on-chain” assets, BitCoin is in no way exceptional here).

Ok, so BitCoin is a purely digital, largely unregulated*, (over)hyped, somewhat mysterious (the coin’s mysterious founder Satoshi Nakamoto certainly couldn’t predict that his decision to remain anonymous will do wonders for BitCoin’s PR and status), and first cryptocurrency of its kind. It is a speculator’s dream. Of course, for this dream to come true, BitCoin needed to catch on in the mainstream, which it did, on a scale no one could have predicted. Discussing the merits of BitCoin as an investment is like discussing the merits of a lottery ticket as an investment. Attributing to early BitCoin miners or buyers some sort of prodigious investment foresight is, frankly, ridiculous. Most of them were tech enthusiast and geeks from the depths of IT community, who were enthusiastic about the technology and were mining earliest BitCoins for fun. The second wave were speculators, who saw a developing new market and speculatively invested without caring much about the technology’s potential. There were exceptions, who seemed to be genuine BitCoin believers (most famous among them being the Winklevoss twins**), but basically, it was a speculators’ market from Day 2 onwards. And then, with some serendipity and luck, this happened:

Chart source: 

NB. The peak is actually slightly understated here, BitCoin got very close to USD 20,000 on some exchanges in Dec-2018.

Fortunes were made, “get rich quick” seminars abounded, evangelists of “it can only go up” appeared and quickly became truly fanatical. All the conditions for an asset bubble / mania were met, and then exceeded. I’m modestly proud to be able to prove that I already held these views when BitCoin was reaching new heights on a daily, a while before its Dec-2018 peak, as can be evidenced by this LinkedIn post from late Nov-2017:


Anyone can do with their money what they see fit, and if it’s speculative BitCoin investment, so be it.****

What is decidedly *not* fine is the exponentially growing environmental cost of BitCoin mining as a result of mining operations’ energy use. The first big news (which took a lot of people, author included) by surprise was that BitCoin mining used more energy per annum than all of Ecuador. Then Ireland. Then Denmark. Then Ireland. Then more than 159 countries in the world.*****

That’s a hell of a lot of energy.

And still, this wouldn’t cause the author so much grief (no grief in fact, perhaps even some jubilation) if that energy was put to good use. Currently the algorithms miners need to crunch in order to demonstrate their proof of work and earn BitCoin transaction validation rights (which is how BitCoins are generated) don’t mean anything. They’re just complicated equations consuming enormous amounts of electricity and computing power. Which, in author’s view, shows that the closest to what can be a tangible, real-world classification of BitCoin is that of pollution derivative (which also makes BitCoin a self-referential instrument, because it’s pollution generated in the process of BitCoin mining). BitCoin is pollution, it’s as simple as that. If only, instead of solving explicitly pointless equations, the credits were given for participation in science- or healthcare-related distributed computing projects such as:

  • LHC@Home: Aiding CERN’s particle research
  • Einstein@Home: Detection of gravitational waves (actually, they were detected in 2015 and won a Nobel prize in 2017, but the project is still running, and my laptop is contributing to it because I saw a popular science programme about LIGO in Louisiana, I thought it was awesome, and that was that; proudly contributing my CPU power since 2016).
  • SETI@Home: The original @Home project aiding search for extraterrestrial intelligence.
  • Or, if they’d prefer something more human-centric, there’s always Folding@Home, whose research helps in a fight with cancer, Zika, Alzheimer’s, Parkinson’s, and several other serious, often terminal conditions.

If the computing and electric power currently used to mine BitCoin were repurposed for the purpose of science, one can only imagine how many advances we’d have achieved by now. A thorough understanding of Alzheimer’s and Parkinson’s? Advanced climate modelling? Better understanding of the human brain? A detailed 3D model of the Milky Way? Advanced mathematics? Particle physics? Nuclear fusion?

It truly boggles the mind how much good could be accomplished out of something that today goes to complete, asset-bubble-fuelled, waste. We’d get so much closer to Stage I on the Kardashev scale on this alone.

And yet this is no happening, and it doesn’t look like it will. One can blame the BitCoin foundation for this as the people who set the rules for the BitCoin network (which is a mess), but even they can’t be blamed in full. The real parties to blame are the usual suspects: greed, short-sightedness, bandwagon effect, etc. 

And this is why I hate BitCoin.

PS. This article is not moaning of a sore loser. I neither lost money on BitCoin nor missed on it out altogether. In the brief period when I was allowed to trade it (i.e. approx. 2 months between jobs), I made approx. 13% by betting on BitCoin’s downfall and volatility. 13% is a negligible profit by BitCoin standards, but by real world standards it’s not too shabby. So I haven’t missed out, and if I still could hold my positions, I’d have made much more than 13% by holding on to my all-short position. Still, on risk-return basis, this wasn’t a very good return.

* That’s perhaps not entirely true: South Korea flirted with the idea of banning cryptocurrency trading via anonymised accounts (mostly concerned about money laundering and concerns over its citizens gambling away their savings) around Jan-2018 before deciding not to proceed; China banned ICO’s and cryptocurrency trading in full in Jan-2018; India introduced a similar ban in Feb-2018


  1. This is an incredibly fast-moving market, and any government / regulatory restrictions may be lifted just as quickly as they were applied
  2. For purely digital assets (which they are) the restrictions can be easily bypassed: VPN access to a network of a country that doesn’t prohibit crypto trading, a TOR browser, an exchange permitting anonymised accounts. Obviously, these activities will constitute a breach of the law, but at the same time, they will be fairly difficult to detect.

** The author doesn’t usually feel sorry for people who got USD 65mln as a court payout, but he couldn’t help but feel that they walked from the Facebook affair somewhat shortchanged. Therefore their early investment in BitCoin (some would say lucky, but they appeared to be solid believers back in 2013, so they deserve some kudos) and its subsequent exorbitant rise in value (100,000 coins worth, at the Dec-2018 peak, close to USD 2bn) was welcome. As a reminder, their idea for a BitCoin ETF (which isn’t that much different to existing BitCoin futures and their offshoots on financial spread betting platforms) was met with reactions best captured in this ZeroHedge article***.

*** The author doesn’t like or read ZH since approx. 2014, when it was becoming more political and less financial, but in its heyday it had one hell of news and commentary.

**** The author believes there will be a lot of tears, buyers’ remorse, and fortunes and savings lost sooner rather than later.

***** The Digiconomist blog has an ongoing analysis of BitCoin’s energy consumption. Another good analysis can be found on PowerCompare.