Oriole Networks: a paradigm shift for AI interconnect

As companies continue to train larger and larger AI models, the amount of data that needs to be moved between GPUs is growing extremely rapidly - due to highly synchronous workloads. This places huge strain on the networking technology that connects thousands or even hundreds of thousands of GPUs together, the ‘interconnect’. As much as 90% of the time training a large AI model is spent moving data around, which is an inefficient use of hugely expensive GPUs as well as a contributor to the substantial energy footprint of AI clusters.To put this in context, the most recent xAI cluster constructed by Elon Musk is reported to have 100,000 H100 GPUs, costing $6b which requires 100MW of power for the cluster - enough energy for 50,000 homes, all generated by burning natural gas and the associated CO2 emissions. The same story is playing out across the AI industry.

In the UK in 1952 a revolutionary new technology was invented - fibre optical cable. Instead of sending information using electrons along metal wire, it became possible to send it using photons along glass fibres. There were huge benefits in doing this: you can move the information at the speed of light, substantially increasing the fidelity of information transmission, increasing the bandwidth and lowering the energy required.

Fast forward to today and you see the echoes of this revolution in our domestic lives - fibre to the home is now commonplace across London and scars on roads across the city show the signs of a huge infrastructure overhaul - we dug up the ground to lay pipes filled with glass to move information at the speed of light.

The photonic revolution has continued beyond telecoms into ever smaller photonic systems and the rise of ‘photonic integrated circuits’, analogous to the rise of electronic integrated circuits.

Professor George Zervas at University College London is one of the world’s leading experts in optical networked systems. Remarkably he has been winner or runner up in the Fabio Neri Award for the best paper in optical switching and networking for 4 of the last 10 years. He has spent his career pushing the limits of how we connect large networks with photonics. His mission has been to find a way to make a wholesale shift from a datacentre networked via electronic packet switching to one that is fully optical. This requires a whole system solution integrating software and hardware to connect many thousands of GPUs to each other in a way that has never been done previously. He and his team members have made the key technical breakthroughs, and proved that they will deliver extraordinary gains in simulation. With Oriole’s network, George believes it will be possible to unlock up to a 10x gain in AI cluster performance - so you could train a model like GPT-4 10 times faster, or with fewer GPUs. As importantly, the proposed optical network consumes a fraction of the power compared to current AI networks.

Simulation is great, but to convince the industry to make such a paradigm shift you need to actually build it, so George formed a spinout and recruited James Regan to join him as CEO and co-founder. While George has spent his career pioneering research into optical networking, James has spent his career industrialising photonics. From Nortel, to Agility Communications to Isca Photonics to EFFECT Photonics, James has lived the photonics revolution and deeply understands what it takes to build these products. He has seen the photonic supply chain mature to the point where a startup like Oriole can be ‘fabless’ and take advantage of foundries that produce photonic integrated circuits. 

I’m deeply motivated by the potential climate impact of Oriole. If society continues to scale these AI systems, we urgently need a way to reduce their carbon footprint. This is a longstanding area of focus for us at Plural with our investments in Proxima Fusion (stellarator fusion), Isometric (carbon removal verification) and Field Energy (grid scale batteries) among others. 

I’m also keen to see a UK start-up play a real leadership role in the future of the AI compute stack. Nvidia recognised the importance of better interconnect with their $6.9b acquisition of Mellanox (creators of Infiniband) in 2019 - roughly 4% of their market cap at the time. At today’s market cap of $3.3 trillion, this would be over $130b - and if you scan GPU list, Infiniband’s dominance is obvious. The most important AI startup to come out of UCL, DeepMind, failed to find the support it deserved from Europe’s venture capital ecosystem and had to go to Silicon Valley and Hong Kong to find ambitious enough early investors, and as I wrote in 2018, DeepMind’s early sale to Google was a hugely significant branching event for the UK.

With Oriole, it’s time for the next generation of UCL founders to get better support. The company has accelerated out of the blocks, raising almost $40m in the first year and a half since founding. They are hiring some of the top experts in optical networking and sprinting to get this network built and to put an alpha product in the hands of customers next year. They’ve also brought on strategic investors like XTX Markets, who operate one of the largest clusters in Europe, and have a longstanding commitment to tackling their climate footprint.

We are excited to support Oriole in their mission to bring a paradigm shift in AI interconnect. Thanks to Jeremy at Fortune for digging into the story more here.

UK stand up tall!

A new European champion for fusion: Proxima spins out from Max Planck

I’m very excited to announce that Plural and UVC Partners are going to be co-leading a €7m seed round for Proxima Fusion, the first ever spin out from the Max Planck Institute of Plasma Physics, creators of the world’s most advanced stellarator. Given the excitement that is building around fusion, I thought it would be interesting to lay out why we are committing to Proxima.

1. Why fusion?

Clean, reliable, safe, cheap, zero carbon, baseload energy generation is increasingly critical as Europe navigates both the climate crisis and works towards true energy security - fusion has the potential to satisfy every single one of these requirements. At Plural we are big believers in wind and solar and are lead investors in Field Energy, a developer of grid-scale battery storage (needed to help manage the volatility of renewables). However renewables are still limited in their ability to provide baseload power to the grid, and require significant land resources. Nuclear fission is an amazing and proven technology but challenges around disposing of nuclear waste, and the risks of catastrophic meltdowns drive both political resistance and increased cost. Fusion has the potential to take advantage of the energy density of nuclear energy but without these constraints.

Beyond the climate crisis and energy security it is important to remember that fusion is the energy source that powers the universe. Harnessing it directly on Earth would be a foundational technology for humanity and one that could take us to a world of true energy abundance. So much human progress has been unlocked by cheaper energy and fusion could take us into a fundamentally new era. Like today’s most advanced submarines, the spacecraft of the future will likely be powered by nuclear reactors. I first fell in love with the vision of fusion playing Sim City as a child - as a game it beautifully captured the path towards more and more ideal sources of energy and by the end of a game my city would always be powered by fusion!

2. Why now?

The old joke about fusion is “it’s 30 years away…and it always will be”. This is not true anymore due to spectacular progress in the field over the past decades - I believe it is now likely we will see the first fusion power plants connected to the grid in the 2030s. This graph shows the progress towards the conditions required for a fusion power plant over time:

Source: [1]

Progress in magnetic confinement (a kind of fusion reactor that uses powerful magnets to confine plasma) has been advancing faster than Moore’s law and we must now have some confidence in extrapolating what might be possible from here and not falling back on tired cliches. Fusion reactors on Earth now routinely reach temperatures above 100 million degrees: we are getting close!

A major enabling technology for magnetic confinement fusion has been incredible improvements in superconducting materials. You can think of the relationship between superconductors and fusion as similar to how progress in semiconductors has underpinned progress in computing.

Over the past decades we now have new superconducting materials that can be made into tape, that superconduct at higher temperatures and produce much higher magnetic fields. This is transformational for magnetic confinement fusion because many properties dramatically improve as you increase the magnetic field. In the case of stellarators, doubling the magnetic field can increase the power output of the machine by as much as 16x (power output scales to the fourth power of the magnetic field). This allows us for example to design much smaller fusion reactors that can achieve similar performance. 

I believe that in the coming decades we will continue to see breakthroughs in higher field superconductors and if this happens, it will be utterly transformational for fusion. However this is all upside - just with the superconducting materials of today we should be capable of putting fusion on the grid.

3. Why stellarators?

Stellarators are maybe the best kept secret in fusion. The stellarator was the original design for a fusion device, created by Lyman Spitzer at Princeton in 1951. It is a fundamentally beautiful design that uses a series of external magnetic coils to entirely confine a plasma across 3 dimensions. If you were trying to come up with the simplest and most elegant way to make magnetic confinement fusion happen you would probably arrive at the stellarator.

The big challenge with stellarators is that in the 1950s we did not yet have the computing resources to fully optimise the design of these complex coils and in 1958 a new kind of fusion device was proposed, the tokamak. This radically simplified things by only confining the plasma on two dimensions using external magnets. The final dimensionality reduction of confinement was achieved by inducing a current in the plasma. As you can see this yields a dramatically simpler machine:

If stellarators are a pretzel, tokamaks are a donut.

This simplification in design complexity led to tokamaks racing ahead of stellarators and absorbing the vast majority of funding for magnetic confinement fusion over the following 50 years.

However there is a serious challenge with tokamaks - what is simpler to design is not simpler to operate. Although the initial design is easier, inducing a current in the plasma can cause what are known as disruptions. Disruptions are almost like a lightning strike inside the tokamak and they can be extremely violent. When I recently toured JET, one of the most advanced tokamaks in the world, it was amazing to see the vast metal frame that is needed to hold the 2800 tonne reactor down in the event of a disruption. Unbelievably a machine the size of 14 blue whales can literally lift off the ground under the explosive force of a disruption! Disruptions remain a major unsolved challenge for tokamaks.

However some true believers continued to keep the dream of stellarators alive. Progress in computing allowed for more complex magnetic fields to be designed. New manufacturing techniques and superconducting technologies allowed for new possibilities. The most remarkable stellarator effort has been led by the Max Planck Institute for Plasma Physics in Germany. The team there continued to build larger and more powerful stellarators, starting in 1988 with the Wendelstein 7-AS which established remarkable new records for stellarators and was then succeeded by the Wendelstein 7-X which took stellarators into a totally different performance regime. As part of making the decision to commit to Proxima Fusion I visited this incredible device and it is one of the most impressive things I’ve ever seen in my life - a masterpiece of physics and engineering. 

Stellarators, whilst harder to design, are radically easier to operate. They do not suffer from disruptions and can be run for much longer than tokamaks allowing them to achieve remarkable new performance records. This graph shows two key attributes of magnetic confinement fusion, the triple product (a good measure of how close a fusion reactor is to net power output) and the maximum confinement time achieved at this triple product (longer confinement is both a desirable engineering achievement and allows for higher output from a power plant):

Figure: triple product of ion density, ion temperature and energy confinement time as a function of the time the triple product was sustained. Adapted from Refs. [2] and [3].

The ideal is to be in the top right hand corner of this graph, and stellarators are taking us closer by the year.

Stellarators and specifically the Wendelstein 7-X are now achieving breakthrough results compared to the field of tokamaks. This is an even more remarkable achievement when you consider that the total funding for stellarators has been tiny compared to funding for tokamaks over the past 50 years. The German approach to nuclear energy has recently come under fire as it has closed down its remaining 3 nuclear fission power plants whilst continuing to keep coal powered power plants in operation. This tweet alone had over 25m views showing that in the face of climate change and Putin’s invasion of Ukraine, European energy policy is a major public issue. The German government who has provided the funding for the Wendelstein series of stellarators deserves enormous credit as does the exceptional team of engineers and scientists who achieved these breakthroughs. I am a visiting professor at Mariana Mazzucato’s institute at UCL. Mariana coined the idea of an Entrepreneurial State - one that has vision to pick critical missions and underwrite the development of breakthrough technology. The German government’s leadership in stellarators is a perfect example of that. Thanks to their incredible leadership and billion plus Euro investment over the last 30 years stellarators have re-emerged as a leading fusion design and we are now much closer to putting fusion on the grid.

4. Why Proxima?

The world’s most advanced stellarator is in Germany and fusion is moving within reach. It is now time to spin up a start-up to take the advances in the field and put fusion on the grid by developing a first of its kind powerplant. This is Proxima Fusion. Founded by an incredible team of engineers and scientists with backgrounds spanning Max Planck, MIT, McLaren and Google they are a remarkable group of high energy founders with deep support from the Max Planck Institute of Plasma Physics. The founding team includes world experts in stellarator optimisation and I’m thrilled to see a true intergenerational effort where Dr. Lutz Wegener and Dr. Felix Schauer who recently retired after leading the development of the Wendelstein 7-X are joining as engineering advisors and Prof. Per Helander and Prof. Elisabeth Wolfrum key leaders at IPP who are scientific advisors to Proxima. It was wonderful to spend time recently in Munich with a team who have been working on fusion for a combined 200 years!

Proxima is taking a simulation-first approach. The design space for stellarators is vast and there is likely a significantly better stellarator design waiting to be discovered prior to building a powerplant. I believe a combination of advanced simulation and the latest superconducting materials will allow Proxima to design the most high performance, grid-ready fusion reactor in the world. 

Europe and the world need fusion as soon as possible. Time to accelerate! This is exactly the kind of mission we want to support at Plural and I am thrilled to be leading the round alongside UVC Partners with participation from Wilbe, HTGF and Torsten Reil.

See also coverage today in the FT, Handelsblatt and Il Sole 24 Ore.

[1] https://www.fusionenergybase.com/article/measuring-progress-in-fusion-energy-the-triple-products

[2] Kikuchi and Azumi, Frontiers in Fusion Research II, Springer, Berlin, 2015

[3] Wolf et al., Physics of Plasmas 26, 082504, 2019.

mRNA pioneer and founder turned investor Carina Namih joins Plural

I’m really happy to announce that Carina Namih has joined Khaled, Sten, Taavet and me at Plural as a peer. She’s been a founder, a CEO, an investor and she has a deep and considered view on how to direct technology and build companies for positive impact.

At Plural we are building a platform for serious founders to put capital, experience and energy behind important missions. Carina is the perfect example of someone who does that.

Carina’s path to becoming a founder started when a member of her family became seriously ill. At 22, she bought a copy of the foundational biology textbook The Cell, read it cover to cover, quit her job, moved from London to San Francisco and committed to founding a company that could make a difference. Along with her co-founders she zeroed in on applying AI to RNA-based medicine and founded HelixNano. 

This was back in 2013, before Covid made the rest of us aware of how important RNA vaccines could be. We believe that non-obvious, hard companies matter, and the BioNTech founders are one of the inspirations we cite on our homepage. It feels so exciting that another RNA pioneer like Carina is joining us. 

What is even more unbelievable about Carina and her co-founder Hannu’s story is that neither of them were insiders. In general starting a biotech company has one dominant path whereby a biotech VC partners with a distinguished academic lab and brings in seasoned managers from the industry. Being a veteran insider is critical to getting off the ground. Carina and Hannu had none of the usual advantages, just vision, intelligence, will and scrappiness. Sometimes, to break new ground in a field, it takes an outsider. 

Fast forward ten years and HelixNano is on the cusp of becoming a truly remarkable company. With their computationally driven approach, they have quietly solved the technical bottlenecks that keep first generation RNA confined to vaccines.Their next-generation platform outperforms Moderna’s by over 10x in trials, in terms of tolerable dose and immune response. Passing these thresholds means that essentially any biologic drug, including cancer drugs, can be turned into RNA and produced in the human body, with the same kind of speed, scalability and cost curves we have seen with COVID vaccines. 

It’s hard to exaggerate how significant this could be - transforming cancer treatment and turning the tide against infectious diseases. HelixNano is now focussed on developing vaccines and cancer immunotherapies that were impossible with previous generation RNA. The company plans to move out of stealth mode with announcements coming later this year. 

This kind of hard tech, science driven, company, is a long, hard slog. To put in context, the first generation RNA companies like BioNTech and Moderna have been 18 years in the making. 

Back in 2012, Carina had to be unbelievably scrappy, hustling money and support from industry partnerships, grants - anything. HelixNano had an ambitious, difficult mission that most people didn't understand or recognise. Eventually they found visionary investors, outside of the traditional biotech circles - Eric Schmidt, 50 Years, Sam Altman, etc - who recognised their potential as founders and the importance of their mission. The company is now backed by a broad base of Silicon Valley’s top investors.

These are the kinds of missions we are here for all day at Plural - our view is that with someone like Carina in their corner, the next visionary, contrarian founder can use shortcuts to tackle important missions.

One thing I really admire about Carina’s journey with HelixNano is that several years in, she had the realness to acknowledge she wasn’t the right CEO for the next phase and handed that role over to her co-founder Hannu - staying involved as a co-founder and board member. All of us at Plural know that leaving the start-up you founded triggers a bit of an identity crisis - you’ve merged your identity so much with a single mission it takes a while to find what’s next.

Carina’s version of decompression was to start working with founders as an investor. She’s gone on to invest in some of the UK’s fastest growing and most exciting companies, including Omnipresent, Phasecraft and Robin. I got to know Carina through the Phasecraft board that I Chair. I was struck by the wisdom and judgment she brought to an extremely challenging mission - making the first useful application of a quantum computer.

What most struck me about Carina was the extreme response she provoked in one of the Phasecraft co-founders. He spent 3 hours with Carina and afterwards basically said he wouldn’t really consider any investment syndicate that didn’t allow for her to be involved - he saw her depth of experience in building a hard tech company and bringing cutting edge science out of the lab, and was utterly convinced she’d be a huge advantage to the company. 

The core premise of Plural is that serious founders and operators have unique scar tissue that they can bring to important missions. Carina’s scar tissue from HelixNano spans so many areas and is not something the typical spreadsheet crunching VC can draw upon. 

Our hope with Plural is to make a significant contribution to the role European start-ups play in shaping our future. We believe Carina is the perfect example of someone who can do that with us and are so excited she’s joining as a peer. 

My favourite books, films and music of 2022

For the 11th year running…here’s what I loved the most in 2022.

Books:

1. Native Speaker

2. Skinship

3. The Starship and the Canoe

4. Starmakers

5. The Man from the Future

6. Crying in H Mart

7. The Power Law

8. With Winning in Mind

9. 8000 Hours

10. The Myth of Freedom

Films:

1. The Lost Daughter

2. Aftersun

3. Drive My Car

4. The Rescue

5. The Alpinist

6. Boiling Point

7. Azor

8. Plus One

9. Riders of Justice

10. Triangle of Sadness

Music (including standout track):

1. Pusha T (Diet Coke)

2. Yaya Bey (Reprise)

3. Daliwonga (Abo Mvelo)

4. Wanitwa Mos (Sofa Silahlane)

5. Felo Le Tee (Dipatje Tsa Felo)

6. DBN Gogo (Bells)

7. 9umba (uMlando)

8. Ye (Easy)

9. Big Thief (Shark Smile)

10. Yung Kayo (YEET)
    

My favourite books, films and music of 2021

For the tenth year running…here’s what I loved the most in 2021.

Books:

1. Night of the Gun (Carr)

2. Who They Was (Krauze)

3. Ministry for the Future (Stanley Robinson)

4. How to Speak Whale (Mustill!)

5. Transitions (Bridges)

6. Where Reasons End (Li)

7. Moms (Ma)

8. The World for Sale: Money, Power and the Traders Who Barter the Earth’s Resources (Farchy)

9. Never Split the Difference (Voss)

10. The Wall (Lancaster)

Films:

1. 7 Prisoners

2. Minari

3. Another Round

4. Nomadland

5. Babyteeth

6. The Painter and the Thief

7. Wrath of Man

8. Les Miserables

9. Sweat

10. No Time to Die

Music (including standout track):

1. Tyler, the Creator - what an album! (LUMBERJACK)

2. Vince Staples (Lil Fade)

3. Cardi B (Up)

4. V9 (Hole in One)

5. J Cole (Pride is the Devil)

6. Backroad Gee (A Yo)

7. Pop Smoke (Manslaughter)

8. Haim (Gasoline)

9. Pete & Bas (Speeding)

10. Adrianne Lenker (Zombie Girl)

The 1.5 Quadrillion Dollar Opportunity

I stayed up last night reading The Ministry for the Future after recommendations by Albert and Yancey. It is a searing read, making the climate tragedy we are living through visceral. The opening story of a heatwave in India chilled me to the bone and I appreciate Kim Stanley Robinson’s attempt to wake more of us up. I certainly woke up this morning feeling a greater sense of urgency to do more to help with the climate crisis.

So here is my small message to my network in tech where I perhaps have some small influence: building technology and start-ups to accelerate the transition away from fossil fuels is one of the most fucking epic business opportunities of all time. At Kim Stanley Robinson lays out crisply:

“Humans are burning about 40 gigatons of fossil carbon a year...Scientists have calculated that we can burn about 500 more gigatons of fossil carbon before we push the average global temperature over 2 degrees C higher than it was when the industrial revolution began...but meanwhile, the fossil fuels industry has already located at least 3000 gigatons of fossil carbon in the ground. All these concentrations of carbon are listed as assets by the corporations that have located them and they are regarded as national resources by the nation states in which they have been found...The notional value of the 2500 gigatons of carbon that should be left in the ground, calculated by using the current price of oil is on the order of 1500 trillion US dollars”

That broadly means that at as we handbrake turn our way out of fossil fuel dependence, 1500 trillion dollars of fossil fuels (1.5 quadrillion dollars!) that would be burned over the coming decades will be replaced by something new and better. You know what’s cooler than a trillion dollars? A quadrillion dollars of climate tech!

It’s a breathtakingly large amount of money and the companies that accelerate and own that transition will likely be some of the largest ever created. There’s a reason some of the most successful venture investors are increasingly focused on climate tech.

To make this more personal, I have been angel investing in climate focused startups since 2017. Some of these startups are focused on food (Ovipost, Hoxton Farms), some on work (Hopin), some on the built environment (Infogrid), some on travel (VirtualTrips), some on transformational new materials (Stealth) and some in green infrastructure development (Stealth). My co-founder from Songkick, Michelle is also now working on a stealth climate project that I’m incredibly excited about. What has been surprising is how well these climate investments have performed. 

A large part of my reason for investing in Hopin in 2019 was that I had stopped flying for work, and hoped that over time more people would see the carbon footprint of international conferences as unsustainable until we decarbonise aviation. Hopin is now probably the fastest growing start-up in the world. Infogrid and VirtualTrips are growing extremely quickly and I think might be two of breakout start-ups in the world this year. The founders of those companies could become incredibly wealthy, while making a dent in the climate crisis.

There are obviously many ways to contribute to tackling the climate crisis beyond technology and business. I have so much respect for friends like Claire Farrell (XR), Bryony Worthington (Quadrature Climate Foundation), Mariana Mazzucato (IIPP), Sam Geall (China Dialogue) and Tom Mustill (Nature Now) who are all helping to pivot us into a more ambitious and aggressive response, but it takes all sorts, and there is a role that start-ups and investors can play too.

So if you are a founder or an investor looking for more of a reason to build or invest in climate focused startups, here’s one - you could get exceptionally rich whilst literally helping to save the world.

My favourite books, films and music of 2020

For the ninth year running…here’s what I loved the most in 2020. Not a great year for media consumption!

Books:

1. The Dark Forest (Remembrance of Earth’s Past, #2) (Liu Cixin)

2. The Three-Body Problem (Remembrance of Earth’s Past, #1) (Liu Cixin)

3. Death’s End (Remembrance of Earth’s Past, #3) (Liu Cixin)

4. Wilding (Isabella Tree)

5. The Secret of Our Success: How Culture Is Driving Human Evolution, Domesticating Our Species, and Making Us Smarter (Joseph Henrich)

6. An Accelerationist Reader (Various)

7. Brave Genius: A Scientist, a Philosopher, and Their Daring Adventures from the French Resistance to the Nobel Prize (Sean Carroll)

8. The Precipice (Toby Ord)

9. Real Life (Brandon Taylor)

10. Range (David Epstein)

Films:

1. Uncut Gems

2. So Long My Son

3. Queen & Slim

4. The Favourite

5. The Assistant

6. Lovers Rock

7. Mogul Mowgli

8. Clemency

9. King of Staten Island

10. The Wild Goose Lake

Music (including standout track):

1. Lil Simz (101 fm)

2. Grimes (Delete Forever)

3. Drakeo the Ruler (Keep it 100)

4. Psychs (Spreadin’ Coronavirus)

5. Haim (Don’t Wanna)

6. Kamikaze (Memories)

7. Waxahatchee (Can’t Do Much)

My favourite books, films, music of 2019

For the eighth year running…here’s what I loved the most in 2019:

Books:

1. Crossing to Safety (Stegner)

2. The Tree of Man (White)

3. A Bend in the River (Naipaul)

4. Crashed: How a Decade of Financial Crises Changed the World (Tooze) (my notes)

5. Technological Revolutions and Financial Capital (Perez)

6. This Could Be Our Future: A Manifesto for a More Generous World (Strickler)

7. The Kingdom (Carrère)

8. Free as in Software (Stallman)

9. Architects of Intelligence (Ford)

10. Human Compatible: AI and the Problem of Control (Russell)

Films (released in the UK in 2019):

1. Marriage Story

2. Burning

3. Dragged Across Concrete

4. A Quiet Place

5. Bait

6. Honeyland

7. Ash is Purest White

8. Parasite

9. Transit

10. Minding the Gap

Music (including standout track):

1. Purple Mountains (All My Happiness is Gone)

2. Pusha T/Lauren Hill/Kanye (Coming Home)

3. Dave (Streatham)

4. Kanye (Follow God)

5. Maxo Kream (Meet Again)

6. Skepta (Bullet from a Gun)

7. Kodak Black (Zeze)

8. Playboi Carti (Shoota)

9. Bonny Doon (A Lotta Things)

10. Stormzy (Vossi Bop)

My notes on Crashed by Adam Tooze

I have just finished Crashed - How a Decade of Financial Crises Changed the World by Adam Tooze and can recommend it.

I came across Tooze on the excellent Talking Politics podcast where David Runciman, Helen Thompson and Tooze discussed the connections between the 2008 financial crisis, the eurozone crisis and more generally, how futile it is to talk about economics and politics as separate topics. I bought his book in the hope that it would level up my understanding of monetary policy and the connections between geopolitics and financial markets. It is a dense, long read (600+pp) but I found it a real page turner.

Some of the things that I found most interesting:

• How liquidity works in globalised financial markets and the connections between central banks, investment banks, money market funds, and other major financial actors like sovereign wealth funds, the IMF and insurance companies.

• With the historical perspective of writing in 2019 about about the 2008 crisis, Tooze captures how the credit crisis of 2008 rumbled on for almost a decade and arguably with the 5th round of QE is still affecting financial markets.

• How, In the lead up to the 2008 crisis most macroeconomists misallocated risk. Most were concerned about interstate economic relations and a potential sovereign debt crisis (primarily China’s accumulation of US treasuries, broadly as a result of China’s mercantilist trade policies and the cost of Bush’s tax cuts ($1.35 trillion over 10 years) and the Iraq and Afghanistan wars ($1-3 trillion)). This obscured the tension building up in the interbank system - in particular the exposure of European banks to both the US housing market and the balance of interbank dollar flows. Connections between banks based in different countries create systemic risks in the same way as international trade imbalances. However, unlike trade imbalances these relationships are harder to scrutinise, and can mutate much more rapidly in the event of a ‘global bank run’. Tooze gives the example of how Larry Summers “slapped down” Raghuram Rajan at Alan Greenspan’s farewell party (!) for flagging the new systemic risks as “luddite” and “misguided”.

• How the financial crisis was fundamentally an Atlantic crisis “America’s securitized mortgage system had been designed from the outset to suck foreign capital into US financial markets….overall, two thirds of the commercial paper issued had European sponsors, including 57% of the dollar-denominated commercial paper...how did European banks end up owning such a large slice of American mortgage debt? The answer is that European banks operated just like their adventurous American counterparts. They borrowed dollars to lend dollars...Indeed, in 2007, roughly twice as much money flowed from the UK to the United States as from China...hundreds of billions of dollars...flowed out of the US from the branches of foreign banks in New York to the head offices of European banks, from which they returned for investment in the US, sometimes by way of an offshore tax haven...European banking claims on the US were the largest link in the system...in the process the European financial system came to function, in the words of Fed analysts, as a ‘global hedge fund’, borrowing short and lending long...the entire structure of international banking in the early twenty-first century was transatlantic...52% of the mortgage-backed securities sold to the Fed under QE were sold by foreign banks, with Europeans far in the lead”.

• How ‘market based insurance’ aka financial derivatives totally failed to stabilise the system and how instead “it turned out that we lived in an age not of limited, but of big government, of massive executive action, of interventionism that had more in common with military operations or emergency medicine than with law-bound governance...the decision made by the American crisis fighters to take those questions off the table and to give absolute priority to saving the financial system shaped everything else that followed. It set the stage for a remarkable and bitterly ironic inversion. Whereas since the 1970s the incessant mantra of the spokespeople of the financial industry had been free markets and light touch regulation, what they were now demanding was the mobilisation of all the resources of the state to save society’s financial infrastructure from a threat of systematic implosion... Martin Wolf, the FT’s esteemed chief economic commentator, dubbed March 14 2008, ‘the day the dream of global free market capitalism died’...A conservative, free-market administration lead by businessmen was proposing unlimited state spending to nationalise a large part of the housing finance system”.

• How despite a narrative of globalisation, the global financial system is fundamentally hierarchical with the dollar at the top, and how the availability of Fed swap lines (with 14 other central banks) during the crisis defined the next level of the hierarchy: “As two US analysts attached to the National Intelligence Council remarked at the end of 2009: ‘Artificial divisions between ‘economic’ and ‘ foreign’ policy present a false dichotomy. To whom one extends swap lines is as much a foreign policy as economic decisions”. The Fed broadly appears to have maintained this authority throughout the crisis through massive global intervention “every major bank in the entire world was taking liquidity assistance on a grand scale from its local central bank, and either directly indirectly by way of the swap lines with the Fed...what happened in the fall of 2008 was not the relativisation of the dollar, but the reverse, a dramatic reassertion of the pivotal role of America’s central bank. Far from withering away, the Fed’s response gave an entirely new dimension to the global dollar”.

• How important differences between central banks can be - he analyses the fundamental differences in mandate and agency between the Fed, the Bank of England and the ECB and the consequences for how the US, the UK and the Eurozone handle the crisis: “What the ECB did not have was a mandate to concern itself with the economic health of the eurozone or its member states in any broader sense...The Fed never took such a narrow view. It had a mandate both to preserve price stability and to maximise employment”.

• How geopolitical the ‘economic’ decisions are - one example is how Paulson described his rationale for bailing out Freddy Mac and Fannie May as a result of them being “too Chinese to fail”. The Ukraine crisis of 2013 emerges partly as a result of the unexpectedly weak support the IMF and EU offered to a Ukraine split between EU and Russian interests (“25% of Ukraine’s exports went to the EU, but 26% went to Russia”) and the resulting economic shock. Putin captures this pithily with his line “geopolitics is geoeconomics”.

• How, in a era where ‘elite’ bashing has become part of populist political rhetoric, Tooze argues that ‘elite closure’ (where government treasury staff, heads of investment banks and central bankers share common backgrounds and are tightly networked) enabled certain countries (the USA, France) to more rapidly take aggressive, coordinated actions and move past a stage of the crisis faster than other where the financial actors found it harder to coordinate.

• How poorly the various commercial rating agencies perform in determining the risk associated with different securities and one possibility as to why: “since the 1980s it was issuers of debt who paid the ratings agencies to make their classifications, not the subscribers to their information services. Payment by the issuer created a conflict of interest”. During the heart of the crisis “effectively the Treasury and the Fed would make themselves the credit-rating agencies in chief - the ‘United States of Moody’s’ - official arbiters of private creditworthiness”.

• How quietly, subtly and profoundly regulation is be rewritten in favour of finance, “in July 2004, as subprime was really hitting its stride, the regulators agreed to provide a permanent exception that effectively allowed assets held in SIVs to be backed by only 10% of the capital that would have been required if the assets were held on the balance sheets of the banks themselves….It was following that regulatory shift that the ABCP market exploded from $650b to in excess of $1t....More than the grand gestures of deregulation, like the 1999 act, it was this kind of apparently small-scale regulatory change that unfettered the growth of shadow banking”.

• How powerfully defaults can act: “automatic stabilisers are the unsung heroes of modern fiscal policy. In the US, no more than one third of federal government spending is discretionary. The rest is made up of mandatory expenditures required by existing ‘entitlements’....these tend to increase in a recession...Between 2007 and 2011, demand in the world economy was stabilised by the largest surge in public debt since WWII.”

• How and why the Eurozone coped so poorly with the crisis - broadly because it was a “monetary union that unified financial markets but provided none of the institutions of governance required for a banking union”.

• How major economic policy decisions are often driven by narratives grounded in sloppy, simplistic analysis. He gives the example of austerity decisions which were based on research from “ultrarespectable...former IMF economists Reinhart and Rogoff…In January 2010 they launched a research paper...this purported to show that as public debts passed the threshold of 90% of GDP, economic growth slowed...to avoid this fate it was necessary to take action sooner rather than later. On closer inspection Reinhart and Rogoff’s analysis turned out to be riddled with errors. Once their Excel spreadsheet was properly edited, there was no sharp discontinuity at the 90% markt and the case for emergency action was far weaker than they made out. But in early 2010 their arguments ruled the roost...Earlier and more sharply than in any other recession in recent history, the fiscal screw was turned. On both sides of the Atlantic the result was to stunt the recovery...The Reinhart and Rogoff meme had reached Europe. Finance minister Schäuble invoked the menacing 90% threshold”. Another stark example is some dodgy economic modeling by the IMF that “systematically underestimated the negative impact of budget cuts. Where they had started the crisis believing that the multiplier was on average around 0.5, they now concluded that from 2010 forward it had been in excess of 1. This meant that cutting government spending by 1 euro, as the austerity programs demanded, would reduced economic activity by more than 1 euro…It was a staggering admission. Bad economics and faulty empirical assumptions had lead the IMF to advocate a policy that destroyed the economic prospects for a generation of young people in Southern Europe.”

I'm joining UCL as visiting professor

I have a bit of news to share - I’ve been appointed Visiting Professor (!) at UCL working with the wonderful Mariana Mazzucato at the department she founded, The Institute of Innovation and Public Purpose (IIPP).

I first came across Mariana back in 2015 when I read her book, The Entrepreneurial State. In it she examines how states have actively shaped technology markets by declaring ambitious missions (for example putting a man on the moon) and making visionary and significant investments in basic scientific research. It made my best books of the year list and the data and arguments she presented fundamentally shifted my worldview.

After I wrote my essay on AI Nationalism last year, I bumped into Mariana at an event and we hit it off. She is a formidable intellectual sparring partner and hugely fun to spend time with - she engages in serious issues without taking herself seriously. I greatly admire the way she has collaborated with politicians across the ideological spectrum (from David Willetts to Alexandria Ocasio-Cortez!) and is influencing policy in many different areas from pharma to the Green New Deal. I believe her UCL Institute for Innovation and Public Purpose (IIPP) will significantly influence the most ambitious policy makers in the years to come.

I’m also excited to spend more time within the UCL world - UCL has played an important role in recent developments in machine learning - two of the co-founders of DeepMind - Demis and Shane met while studying at UCL’s Gatsby Computational Neuroscience Unit.

Finally I’m looking forward to learning from the great group Mariana has assembled at her institute including people like Mike Bracken who lead the digitisation of UK government services at GDS, Carlota Perez who wrote a seminal book on technological revolutions that has greatly influenced investors like Fred Wilson and Josh Ryan-Collins who wrote a phenomenal book on the economics and housing.

I’m going to be using IIPP as my new base to work on the political and economic implications of machine learning. This will be a part-time thing and I’ll continue investing in start-ups and tracking the latest machine learning research with the rest of my time.

My favourite books, films, music of 2018

For the seventh year running….here’s what I loved the most in 2018:

Books:
1. The Making of the Atomic Bomb (Rhodes)
2. Capital (Marx)
3. Darkness at Noon (Koestler)
4. The Prize (Yergin)
5. The Heart is a Lonely Hunter (McCullers)
6. All Things Shining (Dreyfus & Kelly)
7. Jean Monnet, First Statesman of Interdependence (Duchene)
8. Stories of Your Life and Others (Chiang)
9. A Life’s Work (Cusk)
10. What Belongs to You (Greenwell)

Honourable mentions: Just in Time (Hoskyns), Around the Coast in 80 Waves (Bennett), The Value of Everything (Mazzucato), Destined for War (Allison), The Common Thread (Sulston), The Fabric of Reality (Deutsch). Thanks to Demis, Dom and Jeff for the strongest recs this year!

Films (released in UK in 2018):
1. The Square
2. Florida Project
3. Annihilation
4. Leave No Trace
5. Tully
6. I, Tonya
7. Loveless
8. A Prayer Before Dawn
9. Black Panther
10. Revenge

Music (for albums released in 2018 and standout track):
1. Pusha T (If You Know You Know)
2. Chance the Rapper (65th and Ingleside)
3. Anderson .Paak (6 Summers)
4. Jimothy Lacoste (Subway System)
5. Lily Allen (Trigger Bang)
6. Tierra Whack (Pretty Ugly)
7. Kendrick Lamar (X)
8. Cardi B (I Like It)
9. Mitski (Old Friend)
10. Low (Tempest)

AI Nationalism

For the past 9 months I have been presenting versions of this talk to AI researchers, investors, politicians and policy makers. I felt it was time to share these ideas with a wider audience. Thanks to the Ditchley conference on Machine Learning in 2017 for giving me a fantastic platform to get early feedback on my ideas. Thanks also to Nathan Benaich, Jack Clark, Matt Clifford, Jeff Ding, Paul Graham, Michael Page, Nick Srnicek, Yancey Strickler and Michelle You for helpful conversations and feedback on this piece.

Summary

The central prediction I want to make and defend in this post is that continued rapid progress in machine learning will drive the emergence of a new kind of geopolitics; I have been calling it AI Nationalism. Machine learning is an omni-use technology that will come to touch all sectors and parts of society. The transformation of both the economy and the military by machine learning will create instability at the national and international level forcing governments to act. AI policy will become the single most important area of government policy. An accelerated arms race will emerge between key countries and we will see increased protectionist state action to support national champions, block takeovers by foreign firms and attract talent. I use the example of Google, DeepMind and the UK as a specific example of this issue. This arms race will potentially speed up the pace of AI development and shorten the timescale for getting to AGI. Although there will be many common aspects to this techno-nationalist agenda, there will also be important state specific policies. There is a difference between predicting that something will happen and believing this is a good thing. Nationalism is a dangerous path, particular when the international order and international norms will be in flux as a result and in the concluding section I discuss how a period of AI Nationalism might transition to one of global cooperation where AI is treated as a global public good.

Progress in machine learning

The last few years have seen developments in machine learning research and commercialisation that have been pretty astounding. As just a few examples:

• Image recognition starts to achieve human-level accuracy at complex tasks, for example skin cancer classification.

• Big steps forward in applying neural networks to machine translation at Baidu, Google, Microsoft etc. Microsoft’s system achieving human-parity on Mandarin-English translation of news stories (when compared with non-expert translators).

• In March 2016, DeepMind developed AlphaGo--the first computer program to defeat a world champion at Go. This is significant given that machine learning researchers have been trying to develop a system that could defeat a professional player for decades. AlphaGo was trained on 30 million moves played by human experts.

• 18 months later, DeepMind released AlphaZero. Unlike AlphaGo, AlphaZero did not use any moves from human experts to train. Instead, it learned solely by playing against itself. AlphaZero was not only able to defeat its predecessor AlphaGo, but in what is known as ‘transfer learning’ it was also able to defeat best-in-class chess and shogi computers. Leading ML researchers I have spoken with have consistently remarked on the ‘uncanny’ significance of a simpler algorithm that used zero human data ending up being more competent and exhibiting more transferable intelligence. There is a huge gulf between the achievement of AlphaZero and Artificial General Intelligence, but nonetheless there is a sense that this could be another small step in that direction.

Beyond research, there has been incredible progress in applying machine learning to large markets, from search engines (Baidu) to ad targeting (Facebook) to warehouse automation (Amazon) to many new areas like self-driving cars, drug discovery, cybersecurity and robotics. CB Insights provides a good overview of all the markets that start-ups are applying machine learning to today.

This rapid pace of change has caused leading AI practitioners to think seriously about its impact on society. Even at Google, the quintessential applied machine learning company of my lifetime, leadership seems to be shifting away from a techno-utopian stance and is starting to publicly acknowledge the attendant risks in accelerated machine learning research and commercialisation:

“How will they affect employment across different sectors? How can we understand what they are doing under the hood? What about measures of fairness? How might they manipulate people? Are they safe?” - Sergey Brin, April 2018

 

Three forms of instability

So why does this matter to nation states? There are 3 main ways in which accelerating progress in machine learning could create instability in the international order:

1. Commercial applications of machine learning will create vast new businesses and destroy millions of jobs. In the extreme case, the country that invests the most effectively may end up the strongest economically.

2. Machine learning will enable new modes of warfare - both sophisticated cyber offense and defense capabilities but also various forms of autonomous and semiautonomous weaponry for example Lockheed Martin’s Long Range Anti-Ship Missile. In the most extreme case, the country that invests the earliest and most aggressively may end up in a position of military supremacy.

3. Eventually more general purpose AI will enable a fundamental speedup in science and technology research. In my opinion, this might actually be the most profound source of instability. Consider for example the state whose leadership in AI enables them to be the first to develop a viable fusion reactor for power generation. Again, in the extreme case this might enable a country to achieve Wakandan technological supremacy.

Machine learning, to use Jack Clark’s term, is a uniquely omni-use technology that could impact almost every area of national policy. Human intelligence has shaped everything we see around us, so our ability to build machines with greater and greater intelligence could eventually have the same impact. Despite that, we can find some historical parallels to help us think through how things might unfold. Nuclear technology is a dual-use technology with both civilian and military uses (nuclear weapons, radiography, power generation) as is oil (the use of which expanded from lighting to heating, to an incredibly broad range of industrial and military uses). Both of these technologies have had enormous influence on geopolitics and relatively rapidly governments became primary actors and remain so today (consider America’s 6,800 nuclear warheads or the 695 million barrels of oil in the Strategic Petroleum Reserve).

Ambitious governments have already started to see machine learning as the core differentiating technology of the twenty first century and a race has already commenced. This race will come to bear some similarity to nuclear arms race of the last century and the geopolitical tensions and alliances between nation states and multinational companies over oil. Economic, military and technological supremacy have always been extremely powerful motivators for countries.

Industry mix, labour cost, demographics, domestic champions

While the broad threats and rewards of forward-thinking AI policy are common across states, the impact of machine learning is going to vary substantially by country:

Firstly, each country has a different mix of dominant industries and automation is not affecting all industries at the same pace. Compare for example the manufacturing and construction sectors. The construction sector has only recently started to be transformed by digital technologies like Building Information Modeling, whereas manufacturing has seen substantial applications of robotics and automation. That is clear when you look at their comparative productivity gains since 1995:

The impact on wages and jobs will be felt very differently by countries whose core industries are automated sooner. Consider for example Germany, where the automotive industry represents over 10% of GDP; they are going to be more affected by the dynamics around self-driving cars than, for example, the UK where the automotive industry contributes 4% of GDP.

Secondly, every country has a different labour cost that machines will compete against. I have seen this most clearly with a cleaning robotics company called Avidbots (full disclosure: I am an investor). The start-up is headquartered in Waterloo, Canada and produces industrial robots that use computer vision to clean large commercial spaces at a lower price than human cleaning teams in most developed countries. They are seeing orders for their robots from all over the world; however, growth is fastest in Australia due to higher labour costs in the cleaning sector there.

This chart captures well how the economic consequences of automation may vary by country:

If the OECD’s analysis is directionally correct then Slovakia will face a greater challenge in the near term than Norway, with twice as many jobs at risk of automation.

Thirdly, as Kai-Fu Lee articulated very eloquently in his recent New York Times article, only America and China currently are headquarters for the largest AI companies - Google, Apple, Amazon, Facebook, Baidu, Tencent, and Alibaba. National industrial strategy is very different when you are the home of these companies vs. just a customer state. I discuss this in more detail in a later section on the role of national champions.

Finally, in a time when AI is going to materially impact the labour market, different countries have very different attitudes to redistribution, and this will significantly affect how they approach sharing the value created by automation. It is worth noting that while both China and America are home to the leading AI companies, they also both have levels of income inequality at or near their historic peaks.

Blurring line between public & private sectors

This is complicated by the fact that there are incredibly powerful non-state actors who are also competing furiously to develop this technology. All of the 7 most important technology companies in the world--Google, Apple, Amazon, Facebook, Alibaba, Tencent, Baidu--are making huge investments in AI, from low level frameworks and silicon to consumer products.  It goes without saying that their expertise in machine learning leads any state actor at the moment.

As the applications of machine learning grow, the interactions between these companies and different nation states will grow in complexity. Consider for example road transportation, where we are gradually moving towards on demand, autonomous cars. This will increasingly blur the line between publicly funded mass transportation (e.g. a bus) and private transport (a shared Uber). If this leads to a new natural monopoly in road transportation should it be managed by the state (e.g. the call in London for “Khan’s Cars”) or by a British company, or by a multinational company like Uber?

As Mariana Mazzucato outlined in her fantastic book The Entrepreneurial State, states have historically played a crucial role in underwriting long term, high risk research in science and technology by funding either academic research or the military. These technologies are often then commercialised by private companies. With the rise of visionary and wealthy technology companies like Google we are seeing more high risk long term research being funded by the private sector. DeepMind is a prime example of this. This creates tension when the interests of a private company like Google and a state are not aligned. An example of this is the recent interactions between Google and the Pentagon where over 4000 Google employees protested against Google’s participation in “warfare technologies” and as a result Google decided to not renew its contract with the Pentagon. This is a rapidly evolving topic. Only a week earlier Sergey Brin had said that “he understood the controversy and had discussed the matter extensively with Mr. Page and Mr. Pichai. However, he said he thought that it was better for peace if the world’s militaries were intertwined with international organizations like Google rather than working solely with nationalistic defense contractors”.

AI with Chinese Characteristics

In developing a national strategy for AI, China is way out ahead of everyone else. Call it ‘AI with Chinese Characteristics.’ For China over the past couple decades, protectionism has been a winning strategy in developing enduring domestic technology companies, and it has ultimately enabled China to be the only other country in the world with AI companies to rival America’s. Beyond this, China’s technology companies are far more coupled to national policy than in the UK or US, with talk of the Chinese government taking equity ownership in them via 1% ‘special management shares’.

Some notable aspects of China’s early efforts in AI nationalism:

• China has an explicit goal developed at the highest level of government to make itself the global leader in AI by the year 2030. As Jeff Ding notes, China viewed themselves as behind the US in AI policy and this was a major effort to catch up.
• China has committed to a $2 billion AI technology park in Beijing.
• China has developed the ‘Big Fund’ (Credit Suisse estimates total investment at ~$140 billion) to grow the Chinese semiconductor industry. Semiconductor performance is a key driver behind progress in machine learning research and applications
• The state appears to be explicitly focusing their domestic champions on key fields, for example Tencent in computer vision for medical imaging and Baidu for autonomous driving.
• The Chinese state appears to have recognised the importance of data to its AI nationalism efforts. China’s latest cybersecurity law mandates that data being exported out of China have to be reviewed.
• China is implementing specific incentives for key foreign AI talent to relocate to China

The effects of this are starting to be felt. Andrew Moore, Dean of Computer Science at Carnegie Mellon, has estimated that the percentage of papers submitted from China to big AI conferences has increased from 5% a decade ago to 50% today (discussed eight minutes into this interview). This assumes that China is openly publishing all its research. Quantity is obviously not the same as quality and for now researchers based in North America and Europe remain the most influential (for example see Google Scholar ranking by citation). It seems reasonable to assume that this gap will start to close.

Beyond research, Chinese AI startups accounted for an astonishing 48% of global AI funding to startups last year, up from 11% in 2016.

Arguably the weakest link in China’s AI strategy at present is in semiconductors, hence the centrality of that to both the Big Fund and China 2030 and the tension between the US and China in this area, e.g. the US blocking the $117 billion takeover of Qualcomm. China’s annual imports of semiconductor-related products are now $260 billion and have recently risen above spending on oil.

The following graphics illustrate the gaps that China is trying to close in semiconductors and how much smaller the Chinese companies are than the US, Taiwanese or South Korean market leaders. This would also suggest that Taiwan and the Korean peninsula will become an even more geopolitically fraught area for US and Chinese foreign policy.   

(Source)

Key events in the arms race so far

While China has the most developed public position on AI Nationalism, there is a clear and growing competition between major countries to lead the world in AI. When referring to an arms race I am primarily using this term figuratively to describe a competitive dynamic between actors where the value they are creating is partly a function of their relative strength over a competitor. There is also a smaller component of this that is a literal arms race, where states are focused on autonomous and semi-autonomous weapons and machine learning enabled capabilities for cyberattack and defense. Here are the key events so far as I see them.

2014:

• China launches the National Integrated Circuit Industry Investment Fund (aka the ’Big Fund’ with 138 billion yuan ($21.9 billion) to boost fledgling semiconductor industry.

2016:

• Obama White House releases report on future of artificial intelligence. Report is widely read and discussed in China.
• US gov spend of $1.2 billion on unclassified AI-related R&D.
• AlphaGo as a ‘Sputnik Moment’ for China and AI. Sixty million people watch AlphaGo vs. Lee Sedol live. For Westerners who don’t understand the historical significance and popularity of Go in China, consider AlphaGo’s victory as analogous to a scenario where Tencent developed a team of humanoid robots that could play American football and then went on to defeat the New England Patriots at the Super Bowl. Given Go’s deep history as a vehicle for military strategy, the PLA also takes note. Workshops like “A Summary of the Workshop on the Game between AlphaGo and Lee Sedol and the Intelligentization of Military Command and Decision-Making” start to be held.
• Partly in response to AlphaGo, South Korea announces investment of $863 million in AI research over the following 5 years.
• Germany fails to prevent €4.5 billion Chinese takeover of industrial robotics manufacturer Kuka.

2017:

• AlphaGo defeated world No.1 Kie Jie 3-0 in Wuzhen, China. Live video coverage of AlphaGo vs. Ke Jie was blocked in China.
• China announces deeply ambitious plan to become the world leader in AI by 2030.
• Pentagon publicly raises concerns around technology transfer from US to China in various AI related areas.
• Increasing use of CFIUS (Committee on Foreign Investment in the United States) to block acquisitions and investments in US technology companies from Chinese companies or investors. Not limited to US companies--for example, CFIUS also used to block Chinese takeover of Aixtron (German chip equipment maker used in US weapons systems).

2018 so far:

• January: France announces that foreign takeovers of AI companies will be subject to government approval.
• March: France announces its AI plan - plan to invest €1.5 billion over 4 years. Meaningful vision for France’s role laid out by Cédric Villani. Trump uses CFIUS to block Qualcomm takeover.
• April: The UK announces its AI plan to invest £600 million over the coming years (exact annual spend unclear). The EU Commission announces desire to invest €20 billion into AI by 2020. US considers using International Emergency Economic Powers Act to move beyond the blocking of Chinese investment and acquisitions to potentially blocking business partnerships between American and Chinese companies.
• May: South Korea expands 2016 AI plan to $2.2 billion including 6 new AI institutes, a $1 billion fund for AI semiconductors and an overarching goal to reach the “global [AI] top four by 2022”.

AI Nationalism policies

It is helpful to consider the various fundamental actions a state can take in trying to advance its interests in AI. I am listing these roughly in order of how commonly taken these actions have been by governments over the past decade:

• Invest money in research or academic institutions focused on machine learning.
• Help to set standards/regulations so that the technology develops in a way that is most aligned/beneficial to the state’s domestic concerns and companies.
• Indirectly invest money in the sector by subsidising venture capital.
• Directly invest money in key companies.
• Have the state become a key customer for your domestic champions e.g. the relationship between SenseTime and Chinese local and national government.
• Block acquisitions of your domestic AI companies by foreign companies to preserve their independence.
• Block investment into your domestic AI companies by foreign investors.
• Block partnerships between your domestic AI companies and foreign companies.
• Nationalise key domestic AI companies.

My personal belief is that we will see a lot more activity at the bottom of the list over the next few years. In particular, political leaders will start to question whether acquisitions of key AI startups should be blocked or perhaps even reversed. The canonical example for me is Google and DeepMind, which I will discuss more towards the end of this essay.

Domestic champions

Domestic champions are companies that are global commercial leaders in AI but are also headquartered in a specific country, for example Baidu and China or Google and the US. It is worth discussing domestic champions in more detail:

• Commercially, gains from progress in machine learning continue to be dominated by seven companies - Google, Amazon, Apple, Facebook, Alibaba, Tencent, Baidu.
• Only the US and China at present have domestic champions.
• Those companies are contributing to a greater and greater share of their respective equity markets.
• And simultaneously paying less tax: effective reported tax rates paid by top 10 multinational companies has fallen nearly one-third since 2000, from 34% to 24% - this list includes Google, Apple, Amazon, Facebook and Microsoft.
• And importantly they are paying even less tax outside of their home markets.

source

This presents issues for the US and China and even bigger issues for other countries when it comes to redistributing the gains from automation and reducing inequality. If these companies continue to take a larger and larger share of the global economy the delta between tax revenues for China or America and everyone else becomes a bigger and bigger issue for politicians.

Kai-Fu Lee, formerly of Google China and now a leading venture capitalist in Beijing presents a bleak view on how this plays out for countries that are not the US or China,

"[I]f most countries will not be able to tax ultra-profitable A.I. companies to subsidize their workers, what options will they have? I foresee only one: Unless they wish to plunge their people into poverty, they will be forced to negotiate with whichever country supplies most of their A.I. software — China or the United States — to essentially become that country’s economic dependent, taking in welfare subsidies in exchange for letting the “parent” nation’s A.I. companies continue to profit from the dependent country’s users. Such economic arrangements would reshape today’s geopolitical alliances."

This kind of dependency would be tantamount to a new kind of colonialism.

We can see small examples of new geopolitical relationships emerging. In March, Zimbabwe’s government signed a strategic cooperation framework agreement with a Guangzhou-based startup, CloudWalk Technology for a large-scale facial recognition program where Zimbabwe will export a database of their citizens’ faces to China, allowing CloudWalk to improve their underlying algorithms with more data and Zimbabwe to get access to CloudWalk’s computer vision technology. This is part of the much broader Belt and Road initiative of the Chinese Government.

There are historical parallels in all of this with the development of the oil industry. As Daniel Yergin explains in his masterful history of oil:

“two contradictory, even schizophrenic, strands of public policy towards the major oil companies have appeared and reappeared in the United States. On occasion, Washington would champion the companies and their expansion in order to promote America’s political and economic interests, protect its strategic objectives, and enhance the nation’s well-being. At other times, these same companies were subjected to populist assaults against “big oil” for their allegedly greedy, monopolistic ways and indeed for being arrogant and secretive”.

My prediction is that domestic antitrust action against Google and Amazon will not materialise, because for now Washington will care more about strengthening its hand against China. The notes Mark Zuckerberg prepared for his Senate hearing capture this pithily:

“Break up FB? US tech companies key asset for America, break up strengthens Chinese companies.”

 

What can countries that aren’t China or America do?

To answer that question we need to consider the resources that are important to a country in the race to develop a leading position in AI:

• Compute. The compute resources associated with machine learning progress are increasing rapidly. Consider for example this Open AI analysis. While compute costs run into the hundreds of millions for the leading machine learning corporations, this is still small compared to government budgets, so in theory smaller states like Germany, Singapore, the UK or Canada can compete head to head with the US and China.
• Deeply specific talent. At present, progress in machine learning is very sensitive to a talent pool that is microscopically small compared to the world’s population. There are perhaps 700 people in the world who can contribute to the leading edge of AI research, perhaps 70,000 who can understand their work and participate actively in commercialising it and 7 billion people who will be impacted by it. There are parallels with nuclear weapons, where the pool of scientists like Fermi, Szilard, Segre, Hahn, Frisch, Heisenberg capable of designing an atomic bomb was incredibly small compared to the consequences of their work. This suggests that specific talent could be a huge determiner in any AI arms race. China certainly thinks so. In this regard, some smaller countries--notably the UK and Canada--punch massively above their weight.
• General STEM talent. The alternative is that you don’t need a Fermi or an Oppenheimer, you just need a lot of competent engineers, mathematicians and physicists. If so, the balance tips in favour of the largest most-developed countries, with the US and China squarely at the forefront.  
• Adjacent technologies. I have restricted this discussion to machine learning, but it is worth noting that there are various technologies that could contribute to progress in machine learning. For example if quantum computing enables a breakthrough in computing power, this would further accelerate progress in machine learning. A state’s ability to win an AI arms race will be partly enabled by a broader set of technology investments in particular software and semiconductors.
• Political environment - clearly any state action around AI will consume a portion of the leaderships political capital and will trade off against other key issues consuming the country. If a country’s political leadership is absorbed by dealing with another form of instability - for example climate change or Brexit then it will be harder for them to focus attention on AI.

The strange case of the UK

My interest in this topic partly stems from my concern that the UK government is not getting its  AI strategy right.

The UK finds itself in a fortunate position of having DeepMind--arguably the most important AI lab on the planet--headquartered in London. DeepMind has the magical combination of visionary, exceptional leadership in Demis Hassabis, Shane Legg and Mustafa Suleyman as well as the greatest density of AI research talent in the world. If humanity builds Artificial General Intelligence, many of the deepest thinkers on the topic believe that it will happen in Kings Cross. If you were looking for a domestic champion for the UK, you would be hard pressed to find a better candidate.

However, DeepMind is no longer an independent British company. It was acquired by Google in 2014 for £400 million at a critical inflection point: after their success with Atari DQN, but before the big AlphaGo/AlphaZero breakthroughs. It was a brilliant acquisition. In general, it appears that Google has been an excellent parent company for DeepMind, providing substantial resources to increase both the compute spend and the talent base (reported by Quartz as $160 million in 2016) as well as being able to tap into Google’s existing talent in machine learning--for example the Google Brain team. For a pre-revenue startup, remaining independent would have required DeepMind to raise close to half a billion dollars between 2014 and now to execute a similar plan. Today, in the middle of an bull market for AI startups, that seems reasonable, but looking back at 2014--before SoftBank’s Vision Fund and the escalation in huge growth rounds for pre-revenue companies--it would have been a tall order. Ultimately, DeepMind probably chose the highest impact and ambition path available to them in 2014 by selling to Google. I have always had enormous respect for Google and the principled and visionary leadership there is likely a very good fit with the DeepMind culture.

However I find it hard to believe that the UK would not be better off were DeepMind still an independent company. How much would Google sell DeepMind for today? $5 billion? $10 billion? $50 billion? It’s hard to imagine Google selling DeepMind to Amazon, or Tencent or Facebook at almost any price. With hindsight, would it have been better for the UK government to block this acquisition and help keep it independent? Even now, is there a case to be made for the UK to reverse this acquisition and buy DeepMind out of Google and reinstate it as some kind of independent entity?

The two main political parties in the UK both struggle with this kind of question for different reasons. The Conservative MPs I have spoken to about this topic will always cite the troubled history of British Leyland; that spectre of failed market interference still looms large over their thinking. They remain convinced that the only path is laissez-faire economics.

The Labour party has a different challenge. They assert the importance of state action, for example Jeremy Corbyn’s desire to nationalise railways, water and energy companies. But this thinking focuses on those historic battles over privatisation and doesn’t look to the future. Corbyn and McDonnell today are more interested in Great Western Rail than DeepMind.

All of this is further complicated by the fact that the government is hugely distracted by Brexit.

DeepMind is not the only example of an exceptional British company working on cutting edge machine learning. The UK has made many fundamental contributions to the field of machine learning and is home to some of the world’s very best universities for machine learning research including Cambridge, Edinburgh, Imperial, Oxford and UCL. With the growth of the UK’s startup sector over the past decade, there are now many great teams working to combine the UK’s expertise in building great technology companies like Arm, and its academic talent in machine learning.  Prowler is applying reinforcement learning to the general field of decision making. Graphcore is building a new type of processor for machine learning. Ocado is arguably the most sophisticated global player in warehouse automation after Amazon. DarkTrace is one of the leading companies applying machine learning to cybersecurity. Benevolent is doing pioneering work in applying machine learning to drug discovery. All these companies are growing incredibly quickly, doing transformational work in their fields and building deep talent pools. They are all still independent startups. What will the UK government do when Amazon, Google or Tencent make them a multi-billion dollar offer? At present, nothing. This is a good thing if you’re Google, Amazon or Alibaba looking to further cement your position and indirectly a good thing for the US or China. Is it a good thing for the average UK citizen?

Rogue actors

Most of this essay has focused on the national interests of countries. There are other non-state political actors who also have to be considered - for example terrorist cells or rogue states. This is most relevant when it comes to machine-learning-enabled cyberattacks and autonomous weaponry. For those interested to learn more about these risks, they were covered well in this report on malicious uses of AI. The key question for me is the extent to which key labs, corporations or nation states ‘go dark’ in terms of publishing AI research to avoid enabling malicious actors. The risk is well captured by Allan Friedman in Cybersecurity and Cyberwar:

“To make a historic comparison, building Stuxnet the first time may have required an advanced team that was the cyber equivalent to the Manhattan Project. But once it was used, it was like the Americans didn’t just drop this new kind of bomb on Hiroshima, but also kindly dropped leaflets with the design plan so anyone else could also build it, with no nuclear reactor required… the proliferation of cyber weapons happens at Internet speed”

This is also complicated by the fact that cyber attacks may not be as easily identified:

“The problem is that, unlike in the Cold War, there is no simple bipolar arrangement, since, as we saw, the weapons are proliferating far more widely. Even more, there are no cyber equivalents to the clear and obvious tracing mechanism of a missile’s smoky exhaust plume heading your way, since the attacks can be networked, globalized, and of course, hidden. Nuclear explosions also present their own, rather irrefutable evidence that atomic weapons have been used, while a successful covert cyber operation could remain undetected for months or years”

The most likely outcome here is that certain key machine learning research ceases to be shared in the public domain to avoid enabling malicious actors. This thinking is captured most clearly in OpenAI’s recent charter:

“We are committed to providing public goods that help society navigate the path to AGI. Today this includes publishing most of our AI research, but we expect that safety and security concerns will reduce our traditional publishing in the future, while increasing the importance of sharing safety, policy, and standards research.“

If we do see key research labs or countries ‘go dark’ on some of their research output, a Cold War dynamic could emerge that will reward the most established and largest state or corporate actors. Ultimately, this reinforces the AI Nationalism dynamic.

 

The great wall of money

So far the amount invested by states is an order of magnitudes lower than that of Google, Alibaba etc. McKinsey estimates that the largest technology multinationals spent $20-30 billion on AI in 2016.

I believe that the current government spending on AI is tiny compared to the investment we will see as they come to realise what is at stake. What if rather than spending ~£500 million of public money on AI over a number of years the UK spent something closer to its annual defence budget of £45 billion?

Consider again the parallel with nuclear weapons, where the US government went from ignoring key scientists like Leo Szilard to recognising the existential importance of nuclear weapons to initiating the Manhattan Project. The Manhattan Project went from employing zero people in 1941 to within 3 years spending $25 billion (in 2016 dollars), employing over 100,000 people and building industrial capacity as large as the entire US automobile industry. States have tremendous inertia, but once they move they can have incredible momentum.

If this happens, then the amount of investment in AI research and commercialisation could be 10-100X what it is today. It is not always the case that more funding enables more progress but nonetheless I think it is prudent to assume that if states substantially increase their investment in machine learning then progress is likely to speed up further. This only reinforces the importance of investing now in research that helps to mitigate risks and ensure that these developments go well for humanity.

Engineers without borders

It is also worth acknowledging that there are connections that transcend the state and nationalism as Jeff Ding notes in his excellent report “Deciphering China’s AI Dream”:

“It is important to consider the interdependent, positive-sum aspects of various AI drivers….Cross-border AI investments, with respect to the U.S. and China, have significantly increased in the past few years. From 2016 to 2017, China-backed equity deals to U.S. startups rose from 19 to 31 and U.S.-backed equity deals to Chinese startups quadrupled from 5 to 20. Moreover, what is often forgotten is the fact that both Tencent and Alibaba are multinational, public companies that are owned in significant portions by international stakeholders (Naspers has a 33.3% stake in Tencent and Yahoo has a 15 percent stake in Alibaba).”

It is also true that economies and fundamental science and technology progress do not neatly track state borders. Talent and capital are global: DeepMind’s initial investors were from Silicon Valley and Hong Kong, their team is extremely international and they now have offices in Canada and France. There is a weakness to viewing things too narrowly through a state-centric lense. However, I believe that overall the economic and military consequences of machine learning will be such a dramatic cause of instability that nation states will be forced to put their citizens ahead of broader goals around internationalism.

Up until now I have just tried to outline what I think will happen. Machine learning becomes a huge differentiator between states--economically, militarily and technologically--and triggers an arms race, which causes progress in AI to speed up faster.

However there is a difference between predicting that something will happen and believing this is a good thing. Nationalism is a dangerous path, particular when the international order and international norms will be in flux as a result. George Orwell writing on nationalism in 1945 captures the tension between a patriotism that is primarily defensive, and a nationalism that seeks to dominate:

“Nationalism is not to be confused with patriotism. Both words are normally used in so vague a way that any definition is liable to be challenged, but one must draw a distinction between them, since two different and even opposing ideas are involved. By ‘patriotism’ I mean devotion to a particular place and a particular way of life, which one believes to be the best in the world but has no wish to force on other people. Patriotism is of its nature defensive, both militarily and culturally. Nationalism, on the other hand, is inseparable from the desire for power. The abiding purpose of every nationalist is to secure more power and more prestige, not for himself but for the nation or other unit in which he has chosen to sink his own individuality.”

Personally, I believe that AI should become a global public good--like GPS, HTTP, TCP/IP, or the English language--and the best long term structure for bringing this to fruition is a non-profit, global organisation with governance mechanics that reflect the interests of all countries and people. The best shorthand I have for this is some kind of cross between Wikipedia, and The UN. One organisation that has made a step in this direction is OpenAI, which operates as a non-profit entity focused on AI research. This doesn’t solve many of the economic issues around machine learning that I have discussed in this essay, but it is a great improvement on machine learning research being primarily the economic domain of large technology companies and the military domain of nation states.

While the idea of AI as a public good provides me personally with a true north, I think it is naive to hope we can make a giant leap there today, given the vested interests and misaligned incentives of nation states, for-profit technology companies and the weakness of international institutions. I believe that we are likely to go through a period of AI Nationalism before we get to a place where AI is treated like a public good, and that, to use Orwell’s distinction, a kind of AI Patriotism is likely to be a good thing for smaller countries in the short term.

Taking the example of the UK again, I am in favour of a more expansive national AI strategy to protect the UK’s economic, military and technological interests and to give the UK a credible seat at the table when global issues around AI are being worked out. That will help ensure that the UK’s economic interests and values are considered. I believe that the stronger the position of smaller countries like the UK, Canada, Singapore or South Korea in the short term, the more likely we are to move in the longer term to AI as a global public good. For that reason I believe it is necessary for the UK government to take steps towards investing in and protecting its homegrown AI companies and institutions to allow them to play a larger role on the world stage independent of America and China. I have lived in both America and China, and during that time developed enormous respect and affection for both of those countries. That does not prevent me from believing the UK should protect the economic interests of its citizens and I would like to see the UK play a material role in shaping the future of AI. Once again I come back to DeepMind - I believe that the UK and the world would be in a better place were DeepMind to be an independent entity. Ideally, in the longer term as a non-profit, international organisation focused on AI as a global public good.

During the coming phase of AI Nationalism that this essay predicts, I believe we need a simultaneous investment in organisations and technologies that can counterbalance this trend and drive an international rather than national agenda. Something analogous to The Baruch Plan led by organisations like DeepMind and OpenAI. I plan to write more about that soon.

My favourite books, films, music of 2017

For the 6th year running I give you some arbitrary rankings of some media products from the last year.

Books
1. The KLF: Chaos, Magic and the Band who Burned a Million Pounds
2. Reinventing Organisations
3. Dharma Bums
4. Let My People Go Surfing
5. From Third World to First: The Singapore Story 1965-2000
6. The Piano Teacher
7. All Out War
8.  Dark Money
9. The Institutional Revolution
10. Life 3.0 

Films (released in UK in 2017)
1. Mountains May Depart
2. American Honey
3. Toni Erdmann
4. Lady Bird
5. Moonlight
6. Get Out
7. 20th Century Women
8. I, Daniel Blake
9. I Am Not Your Negro
10. Elle

Music (for albums released in 2017 and standout track)
1. Kendrick (Pride)
2. Travis Scott (Butterfly Effect) 
4. E-40 (Choices)
4. Cardi B (Bodak Yellow)
5. Stormzy (Mr Skeng)
6. Vince Staples (Big Fish)
7. Future (Mask Off)
8. YFN Lucci (Everyday we Lit)
9. Jeremih (Oui)
10. 21 Savage (Bank Account)

My favourite books, films, music and meals of 2016

For the 5th year running I give you some arbitrary rankings of some media products from 2016.

New non-media categories added this year!

Books
1. JR
2. Gilead
3. All the Presidents Men
4. 1984
5. Under the Volcano
6. The Adventures of Augie Marsh
7. The Recognitions
8. Submission
9. When Breath Becomes Air
10. Jerry Moffatt Revelations

Films (released in UK in 2016)
1. 45 years
2. Dheepan
3. A Bigger Splash
4. Kajacki
5. Wiener
6. Magic Mike XXL
7. Heart of a Dog
8. Leviathan
9. Carol
10. Eye in the Sky

Music (for albums released in 2016 and standout track)
1. Anohni (Watch Me)
2. Skepta (Konnichiwa)
3. Kanye (Real Friends)
4. Ariana Grande (Be Alright)
5. dvsn (With Me)
6. Beyoncé (All Night)
7. Rich Chigga (Dat $tick)
8. YG (Still Brazy)
9. ATCQ (We the People)
10. Lil Yachty & Dram (Broccoli)

Meals
1. Mugaritz in San Sebastián w stag ladz
2. Delhi food walk in Old Delhi w M
3. Bar Nestor in San Sebastián w stag ladz
4. channa masala at Amma’s Ashram w Laura & M
5. Roscioli in Roma w Jeff & Alex
6. mirchi bada at Shahi Samosa in Jodhpur w M
7. golgappa at Kashi Chat Bhandar in Varanasi w M
8. spicy fish bhaji & mixed samosa at roadside coffee shop in Oman w M
9. slow cooked pork carnitas w orange Chez Hog
10. Paolo’s Italian BBQ veg in Somerset w P&K crew

Top 10 active outdoor lyfe:
1. sending Lucky Luka in Kalymnos
2. climbing at Hadash
3. night surfing in Ekas
4. hiking round Telendos
5. walking over the Jean Claude & Christo golden piers installation at Lake Iseo
6. sending Marie Rose in Fontainebleau
7. climbing in Yangshuo
8. running London 10k
9. deer hunting in N Florida
10. Napali coast hike

New Year, new roles

As we start the New Year, I want to share some personal and professional news. I will be moving to a Chairman role and Matt Jones will be CEO. We’ve always known the co-CEO structure was temporary and best suited to the the first stage of our merger - but why make this change now? There are a number of reasons.

The co-CEO approach has been a helpful structure for Matt and me to combine our visions, strategies, products, organisations and cultures - everything that we have had to marry to make this merger a success. We’re now 7 months into the merger, and have created some great momentum, so can revert to a more orthodox leadership structure.

I have also been working towards this transition for very personal reasons. I’ve always tried to use this blog as a place to be real about some of the challenges of building a start-up and should share the other reason why this is the right time. A few years ago my younger sister was diagnosed with a brain tumour. The prognosis was not great, but there was room to hope. At the start of 2015 my sister’s prognosis abruptly shifted to terminal and she died weeks later. I’m a very private person when it comes to my family, so this isn’t easy to share in public, but reading essays by Paul Bucheit on the death of his brother really helped me this year, so I’m trying to be more open. This transition is also about me taking time to properly grieve after spending the last year head down and focused on making a success of our merger.

During 2015 we achieved some great things. We raised two rounds of financing, hired some great new people, launched new products, and saw growth we are all excited by. I’m also really proud of the progress we’ve made against one of the less tangible goals of the merger - to combine the DNA of CrowdSurge - a deep understanding of the needs of artists surrounding ticketing - with the DNA of Songkick - building scalable consumer products for fans. We didn’t expect to see the results of that within the first year, but when Adele’s team approached us to help them counteract what they expected to be unprecedented levels of scalping around their upcoming tour, we launched a new product that drew on all the strengths of the merged company. In the words of one industry commentator “Songkick has done more in one campaign to address the issue of touting than has been achieved to date by any other party in any other sector…the prospects are tantalising and, for once, both the artist and the fan seem to have won”. I’m really looking forward to us scaling this product up with more artists in 2016.

We also, like all start-ups, have challenges ahead and felt that the next phase of execution would benefit from a more orthodox and battle-hardened leadership structure.

Without a doubt, Matt Jones is the best CEO for the next chapter. Matt’s been a long-time friend and collaborator of mine, and our shared vision and mutual respect were big factors in wanting to merge our companies. He’s one of the most impressive people I’ve ever met, with a combination of insane levels of energy, infectious ambition, and downright relentlessness. If you spend a few minutes in his company you’ll see that he cares at the deepest possible level about the future of the concert industry, and has a single-minded determination to use technology to make it better.

His leadership is a huge asset to Songkick, and it’s been a big factor in our ability to continue to hire exceptional team members, raise funding from world class investors and earn the trust of artists like Adele, Metallica and Mumford & Sons. Matt’s vision has always been that an artist should be able to sell tickets wherever their fans are engaged, and that vision of distributed commerce is central to our plans for 2016. But as well as being a visionary, Matt’s also deeply pragmatic - a rare combination in our industry - and a leader dedicated to getting shit done.

I’m grateful to our board and executive team in supporting Matt and me with this transition, and in particular to my co-founder Michelle and my new co-founders Adam and Matt for their support throughout this year.

2016’s a huge year for Songkick, and now the merger’s fully complete we have the right team, technology and structure to make it a success. I’m more convinced than ever that we can have a transformative impact on artists, fans and the wider industry, and I’m excited to continue building towards this in the year ahead as Chairman.

My favourite books, films and music of 2015

Same format as last year, and the years before.

Books (most published before 2015):

1. The Grapes of Wrath (Steinbeck)
2. Moby Dick (Melville)
3. A History of Western Philosophy (Russell)
4. East of Eden (Steinbeck)
5. Any Human Heart (Boyd)
6. Between the World and Me (Coates)
7. H is for Hawk (Macdonald)
8. Stoner (Williams)
9. Barbarian Days: a Surfing Life (Finnegan)
10. The Entrepreneurial State (Mazzucato)

Other good things I read: Strangers Drowning (MacFarquhar), Speak, Memory (Nabakov), USA Trilogy (Dos Passos), Purity (Franzen), Antifragile (Taleb), Post Capitalism (Mason)

Films (released in the UK in 2015)

1. Starred Up
2. Whiplash
3. Still Alice
4. Birdman
5. Margin Call
6. Mommy
7. The Duke of Burgundy
8. Citizenfour
9. Girlhood
10. Creed

Artists (for music released in 2015 and standout track)

1. Kendrick Lamar (Alright)
2. Skepta (Shutdown)
3. Kanye (Only One)
4. Vince Staples (Jump off the Roof)
5. Hot Chip (Huarache Nights)
6. Dej Loaf (Back Up)
7. Nicki Minaj (Feeling Myself)
8. Rae Sremmurd (Throw Sum Mo)
9. D’Angelo (The Charade)
10. Grimes (California)

Full Stack Music: 1 Trillion Streams, 200 Million Tickets

(repost of guest blog post I wrote for TechCrunch)

At present, there are three distinct music industries: radio, on-demand music, and concert ticketing. However, we are starting to enter a new phase, where these industries will converge and produce one integrated experience for artists and fans. I’ve taken to calling this full stack music, because at heart it speaks to a holistic experience that integrates these industries through data.

The integration of these three, previously distinct industries will produce a richer experience for artists and fans, unlock a ton of additional subscription, ticketing and advertising revenue for artists and create a better experience for fans. It will resolve the central tension between fans, artists and technology companies that so much ink has been spilled about.

Three Distinct Music Industries

There are three main businesses of music:

Radio 

Radio is where music discovery happens, and where the majority of casual music fans engage with music. Ninety-two percent of the U.S. population listens to radio at least once per week; on average, they listen for 15 hours. It is critical to artists because a radio station decides which track a fan listens to next, and so radio has an incredible ability to drive new artist discovery. Radio is primarily monetized via advertising, generating $45 billion/year. It also is the primary channel for marketing concerts.

On-Demand Music 

This is when the listener decides exactly which song comes next (unlike radio). It started with vinyl, migrated to CDs, migrated to iTunes and finally has migrated to on-demand streaming services like Apple Music, SoundCloud, Spotify and YouTube. Monetization used to be in the form of direct spend (buying a CD); it is now a mix of advertising and subscription.

Concert Ticketing

Paying to see your favorite artist live. This used to be a side business for the music industry. However, over the past 10 years this has expanded to become the main event, growing from $10 billion in 1999 to $30 billion in 2015 in gross ticket sales. It is where artists make the majority of their income — typically 70-80 percent. Most of the growth has come from increasing ticket prices — 50 percent of concert tickets go unsold and attendance has not increased anything like as fast as prices.

These industries have been loosely coupled in the past. Going back to 1999, the record company would use radio as a way to get fans to discover a new act, then monetize that investment, primarily via selling “on-demand” access in the form of CDs and, finally, drive additional discovery by subsidizing touring (known as “tour support;” a label would underwrite some of the cost of touring to help build an audience to whom to sell CDs). Touring represented a small percentage of artist income.

The industries were also coupled at a corporate level at one point, with ClearChannel. Over the course of many years, a massive roll-up of local U.S. radio stations resulted in ClearChannel. That rolled-up business exists today as iHeartRadio, with 850 local stations and 245 million monthly unique listeners. In parallel, a roll-up of local concert promoters produced a new touring behemoth, SFX Entertainment, and the two businesses — radio and concert promotion — were merged in 2000 to form a new conglomerate. The goal was to combine the No. 1 channel for concert discovery (radio) with the No. 1 promoter of concerts (SFX).

Eventually, these businesses were separated into Clear Channel Communications (iHeartRadio) and Clear Channel Entertainment (LiveNation). Subsequent to that, LiveNation embarked on a huge project of vertical and horizontal integration and, at this point, is the world’s largest artist manager (Maverick/ArtistNation), the world’s largest primary ticketing company (TicketMaster), the world’s second-largest secondary ticketing company (TicketMaster+) and the world’s largest festival owner, venue owner and concert promoter.

Internet Music: Radio And On-Demand Converge

We have seen a massive transformation of the recorded music landscape — with the growth of iTunes/Apple Music, Deezer, Pandora, SoundCloud, Spotify and YouTube — to the point where more than 1 TRILLION tracks are now streamed online across these services each year. The line between radio and on-demand is rapidly blurring across each of these services:

• SoundCloud and YouTube both autoselect another track to play when the one you searched on-demand finishes — that is, a radio experience.
• Pandora, historically a pure radio service, has started to enable whole album streams on-demand.
• Spotify has shifted from a pure on-demand model to offer a radio experience very similar to Pandora, where you can select an artist and listen to a radio mix based on that cue. Every Monday now, Spotify will provide you with a personalized radio stream of songs you might enjoy, based on your listening history.
• Finally, Apple Music has taken this integrated approach the furthest, with a live on-air radio experience seamlessly integrated with a library based on demand experience.

At the same time, discovery of local concerts has started to transform — rather than generic emails about all the tickets on sale in Los Angeles, new services like Bandsintown and Songkick (which I co-founded) will send you personalized alerts whenever the artists you listen to on these music streaming services announce a local show. These concert discovery apps now reach more than 20 million fans each month, and are more personalized and convenient way to find out about concerts.

The Next Phase: Streaming And Ticketing Converge

Leading artists have started to articulate the extent to which streaming music and ticketing are becoming joined at the hip. For example, Ed Sheeran:

“I’m playing three Wembley Stadium (shows) on album two. I’m playing sold-out arena gigs in South America, Korea, south-east Asia and Australia. I don’t think I’d be able to do that without Spotify or if people hadn’t streamed my music. My music has been streamed 860million times, which means that it’s getting out to people. I get a percentage of my record sales, but it’s not a large percentage, (whereas) I get all my ticket sales, so I’d much rather tour. That’s why I got into the business — I love playing gigs. Recording albums, to me, is a means to an end. I put out records so I can tour. For me, Spotify is not even a necessary evil. It helps me do what I want to do.”

Over the next few years we will see this connection between streaming and ticket sales become completely explicit. Streaming services will increasingly make it seamless for fans using their services to see when the artist has a local show; Songkick’s existing API partnerships with Deezer, SoundCloud, Spotify and YouTube are hints at what this could look like. It’s not impossible to imagine a time when you could possibly buy tickets directly from your favorite artist right inside your streaming service.

When that happens, artists will finally be able to see a connected picture of how their music is distributed and monetized. An act who gets 100 million streams will see that 10 million of those were monetized via paying subscribers, 90 million by ads and another 5 million fans via ticket purchases. The outcome will be a more seamless experience that results in casualmusic fans attending more concerts.

This is a big deal — only 20 percent of Americans attended a concert in the last year, and the biggest reason for not going is that they didn’t know when shows for their favorite act were happening. This will finally create firm alignment between artists and music streaming services — to the point where all acts will see the explicit and causal relationship between an ad-supported online radio stream and a paid ticket purchase, as Ed Sheeran does.

This is just the start, though. Along with joined up analytics for artists, fans will be offered new propositions that tie together live and recorded music experiences. For example, imagine if all streaming music subscribers were offered lower booking fees on ticket purchases — creating another reason for fans to subscribe rather than use an ad-supported service, and driving faster growth in subscription income for artists.

After the show, the set list will immediately be available on your streaming service of choice, further helping to reinforce the connection you have built to that artist and increasing the likelihood of buying merchandise from the gig. Finally, tour routing will be impacted by the data from streaming services similar to Spotify’s recent experiment with Hunter Hayes:

“Hayes turned to Spotify to help him route the tour. The online music streamer crunched its numbers and determined the college markets where the country star is the strongest. Hayes’ biggest fans in the target markets will receive pre-sale access. His top 21 fans in each market will receive such prizes as early entry, meet-and-greets, signed memorabilia and other goodies. The fan who streams Hayes’ music the most in each market will be awarded a one-year sub to Spotify Premium.”

The key point across all of this is that the central, most valuable asset of streaming musicservices will be the listener data they generate. As we shift from offline radio to online streaming, artists will know how those 1 trillion tracks of music were streamed — which fan listened to them, where they were based, which concert tickets they purchased in the past — and be able to tailor personalized and richer experiences to their fans.

That is an incredible shift compared to the data-poor ecosystem of 1999. The trend will only continue, as more and more offline listening (in particular, terrestrial radio) migrates to online streaming. Once that is fully complete, we will hit 5-10 trillion streams, and these shifts will be even more critical. Zoe Keating, one of the most visionary artists I have the pleasure of knowing, has been saying this for a few years now:

“I want my data and in 2012 I see absolutely no reason why I shouldn’t own it. It seems like everyone has it, and exploits it…everyone but the creators providing the content that services are built on. I wish I could make this demand: stream my music, but in exchange give me my listener data. But the law doesn’t give me that power. The law only demands I be paid in money, which at this point in my career is not as valuable as information. I’d rather be paid in data….The new model says that in the future I’m not supposed to sell music: I’m supposed to sell concert tickets and t-shirts. Ok fine, so put me in touch with the people who will buy concert tickets and t-shirts.”

There are signs that this integration is coming — Pandora appointed the former CEO of AEG Live, the world’s second-largest promoter, to their board, and have started to experiment with concert marketing — for example, campaigns to promote tours for the Rolling Stones and Odesza. Global Radio, the largest terrestrial radio company in the U.K., has expanded into artist management and concert promotion, again hiring key execs from AEG Live. AppleMusic is broadcasting on Beats 1 all the shows from their upcoming festival, and is encouraging the artists they book to share information about their performances on Connect.

We are in the early stages. Eventually we will know not just how many streams are generated per artist, but how many ticket sales resulted. If this deeper integration of streaming and ticketing results in one ticket sold per 5,000 streams, then we’d know that 1 trillion streamsgenerated 200 million tickets — at an average face value of $50, this would be $10 billion in ticket gross — equivalent to the revenue from 100 million subscribers paying $8/month. It would also have the consequence of making Apple, SoundCloud, Spotify, Pandora or YouTube new power players in ticketing.

Artists will start to focus on promoting the service, which in aggregate generates the most ad, subscription and ticket revenue, which in turn will drive further growth in online listening. We’re about to watch the next big shift in online music play out as we move from three separate music industries to Full Stack Music.

My favourite books, films and artists of 2014

Same format as last year.

Books (most published before 2014):
1. Anna Karenina (Tolstoy)
2. My Struggle books 1 & 2 (Knausgard)
3. Herzog (Bellow)
4. Island (Huxley)
5. Ham on Rye (Bukowski)
6. Post Office (Bukowski)
7. Superintelligence (Bostrom)
8. The Diversity of Life (Wilson)
9. White Girls (Als)
10. Steve Jobs (Isaacson)

Musical artists (new music released in 2014) & standout track:
1. Beyonce (7/11)
2. YG (Bicken Back Bein Bool)
3. Lil Wayne (Rich as Fuck)
4. Makonnen (Tuesday)
5. Future Islands (Seasons)
6. Vince Staples (Limos)
7. Young Thug (Lifestyle)
8. Ariana Grande (Bang Bang)
9. Sleaford Mods (Tiswas)
10. Bobby Schmurda (Hot N***a)

Films (UK release in 2014):
1. All is Lost (Chandor)
2. Le Weekend (Michell)
3. Boyhood (Linklater)
4. Bullhead (Roskam)
5. Under the Skin (Glazer)
6. A Touch of Sin (Zhang Ke)
7. Nightcrawler (Gilroy)
8. Interstellar (Nolan)
9. Blue Ruin (Saulnier)
10. Fury (Ayer)

(haven’t yet seen Ida, ‘71, Leviathan, We are the Best or American Sniper - suspect they will shake the list up)

Pitchfork and Songkick

We just launched a new partnership with Pitchfork to integrate listings from Songkick  into Pitchfork’s core review pages. I’m out in California working round the clock as usual but even if only for myself I wanted to take a minute out of the waves of meetings and emails to remind myself what this means to me.

I have been reading Pitchfork since I was a teenager. I have found more music that I have fallen in love with through Pitchfork than through any other online service. They have made my life as a fan an order of magnitude richer. Ever since we started Songkick in 2007 it has been a dream to work with them. We have some amazing API partners - Bandcamp, HypeMachine, SoundCloud, Spotify, YouTube - they are all examples of companies I find inspiring, but still there’s just such a rush from finally getting this partnership live.

After years of being a founder Pitchfork now stands for something different than it did when I was a teenager or when we were getting started. It stands for building something that endures and gets better every year. Pitchfork has been getting better EVERY YEAR SINCE 1995. Outside of VICE it’s hard to think of a brand and team that has shown as much commitment to online media. How many great online music services have died in just the years while Songkick has been active? iLike, Imeem, MySpace, Lala - this shit is hard. Let alone when you consider everything that has happened since Pitchfork was founded. And yet they endure, improve and prosper.

Michelle, the whole team and I are really proud they chose us as a partner and can’t wait to do more together.

My notes on Superintelligence by Bostrom

On the plane to the US I finished reading Nick Bostrom’s Superintelligence. I jotted down notes as I went and thought a few friends might be interested so posting here.

Bostom’s background spans philosophy (he is a professor at Oxford), computational neuroscience and physics - his breadth of knowledge makes this a broad reaching read. It’s particularly interesting if you have a basic understanding of machine learning and want to understand some of the philosophical and ethical questions raised by superintelligent machines.

A few things that stood out for me:

- various surveys of AI experts (who are plausibly at the optimistic end of the spectrum :) ) peg the likelihood that we will see machines with human level intelligence by 2040 at 50%, and 90% by 2075

- Bostrom convincingly argues that once human level machine intelligence emerges we may rapidly see an ‘intelligence explosion’ where the intelligent machines self-enhance their own software/intelligence at high speed. This leads to machines that are superintelligent. Since software can be copied the population of superintelligent machines can grow rapidly.

- He then argues that given the kinetics of such an explosion one entity may end up rapidly accelerating past other machine intelligence projects and forming a dominant position. This echoes the writing of Lanier and others on the increasing centralisation of power within the technology industry. He makes a particularly interesting point that digital agents may tend to greater centralisation of control due to reduced inter-agent transaction costs. For example the idea that firms or nations of machines could massively increase in size.

- the majority of the book focuses on what happens after a superintelligence emerges. He draws an interesting distinction between having more intelligence and more wisdom - and the risks of one developing without the other. He gives a hilarious example, worthy of Foster-Wallace where a superintelligent machine is tasked with producing 1000 paperclips. The machine, being superintelligent and supercapable rapidly produces 1000 paperclips. However, being a perfect Bayesian agent it is also aware that observational error may mean that it has actually produced fewer paperclips than this - there is a tiny but real chance it has only produced 999. So to remedy this it commandeers all the resources in the known universe to more accurately count whether it has actually produced 1000 paperclips or not. He lays out various types of superintelligence and various ways that things could go badly wrong for humanity from goal functions that on first glance seem to be bounded, but per the paperclip example are not. At lot of this seems to be the difference between programatic logic and 'common sense’ and the complexity in creating a bridge from one to the other.

- he draws an interesting parallel between the fate of humans in a world with superintelligent machines, and the fate of horses in a human world. The horse population grew massively through the 1900s as a complement to carriages and ploughs, but then declined with the arrival of automobiles & tractors. The population of horses was 26m in the US in 1915 but declined to 2m by the early 1950s. The flipside of this is that the horse population subsequently returned to 10m driven by economic growth that have allowed more humans to indulge in leisure activities involving horses. 

- He explores how superintelligent machines might acquire their values. This section on value loading techniques is very interesting and summarises some of the most interesting mathematical and philosophical challenges facing the AI space. For example in one unfinished solution to the value loading problem we have a subset of intelligent machines that are known to have values that are safe for humans. These machines are allowed to develop a incrementally more intelligent machine - where the step in intelligence between the first group of machines and the mutation is small enough that the earlier machines can still test the new, slightly smarter machine to see if its values remain compatible with humanity’s safety. He makes the terrifying point that if there is an arms race going on for one company or nation to develop superintelligent machines first, this kind of caution is unlikely to be on the path of the 'winning’ project - 'move fast and break things’ seems like a bad motto when you are playing with something this powerful.

- Having framed the challenges of loading a superintelligence with values, he then moves to what values we want this superintelligent to have. Bostrom argues that humanity may have made relatively little progress on answering key moral questions and is likely still labouring under some grave moral misconceptions. Given that are we in a position to specify a moral framework for a superintelligent machine? He introduces the concepts of Indirect Normativity and coherent extrapolated volition in response to this - a hedge against our own limited moral framework and a bet that the machine can do better:

“Our coherent extrapolated volition is our wish if we knew more, thought faster, were more the people we wished we were, had grown up farther together; where the extrapolation converges rather than diverges, where our wishes cohere rather than interfere; extrapolated as we wish that extrapolated, interpreted as we wish that interpreted” - Yudakowski

- finally he asks how to ensure that the immense economic windfall resulting from superintelligence should be distributed to benefit all of humanity, not just a narrow set of people (or machines).

Overall I found it very stimulating and would recommend.