What Ronaldo can teach us about Technology and Geopolitics...

...or what Manchester City has in common with a Large Language Model

Some tech and geopolitics, with a side order of football (soccer)

On the face of it Al-Nassr and Falcon 40B have little in common other than the fact that most people have never heard of them. The former is a Saudi football team; the latter a large language model (LLM) of the type that is proliferating as the artificial intelligence revolution gathers pace. But both are based in the Gulf and form part of the strategies of states in the region to diversify their economies away from hydrocarbons. And both are in danger of being underestimated.

After the 2022 World Cup Finals, Al-Nassr secured the services of Cristiano Ronaldo, one of the finest players of his generation. Falcon 40B, which has been built and trained by the Technology Innovation Institute (TII) of Abu Dhabi, is currently ranked as the world’s highest-rated open-source LLM. In sport and in tech, they pose a challenge to the old order.

The Premier League has been the dominant competition in European, and therefore World football, for much of the last decade. This trend has been underpinned by money - first from TV rights and then from foreign owners.

Pep Guardiola's Manchester City are a case in point.  The club's fortunes were (quite literally) transformed when it was taken over by Sheikh Mansour bin Zayed al-Nahyan, a member of the United Arab Emirates (UAE) royal family, in 2008. Since then, Manchester City have risen from Premier League also-ran to become arguably the world's most successful club side.

This success didn’t go unnoticed by the UAE's peers in the Gulf region, most notably Qatar and Saudi Arabia. The tiny Gulf emirate of Qatar bid for and sensationally won the right to host the 2022 FIFA World Cup; Saudi Arabia’s Public Investment Fund (the nation’s sovereign wealth fund) bought Newcastle United.

But now this ambition has morphed: rather than simply buying existing clubs and hosting existing competitions, Gulf states want to create their own. In business terms, they are augmenting an acquisition strategy with a move into venture capital.

To date these efforts have largely engendered derision. Admittedly, Ronaldo is in the twilight of his career. His signing has been dismissed as sports washing by Saudi Arabia, a poor investment by Al-Nassr, and a sell-out by Ronaldo.

But as Liverpool's Roberto Firmiño, Real Madrid’s Karim Benzema, and most recently Wolves' sought-after talent Ruben Neves accept transfers to the Kingdom - and an array of talented international players like Jordan Henderson and Fabinho consider playing there - we should perhaps ask whether we are witnessing the start of something.

For all the legitimate criticism levelled at Saudi Arabia and Qatar, and to a lesser extent the UAE, each country is very clearly using sport to rebrand their image by (and as a means of) attracting talent from abroad as they seek to reconfigure their economies.

And this ambition is not limited to sport. We are starting to see the “Ronaldo Effect” in technology, and its impact on geopolitics.

Since late 2022, when OpenAI released ChatGPT and the GPT3 model that underpinned it, much tech commentary has been dominated by the seemingly astonishing progress made in artificial intelligence and its potential to transform societies.

In reality, the GPT “breakthrough” was not as sudden as it may seem, and can be traced back roughly a decade. As, Alex Jaimes, the chief scientist and AI officer at Dataminr, recently explained to me, that was the point at which algorithmic innovations, allied to radical increases in the scale of data and the power of available compute set us on a path to GPT3 and its peers/successors in the world of large language models.

There’s a perception that the AI revolution is being driven almost entirely by US-based researchers like Jaimes, along with US innovators and startups, hyperscale US tech companies (Google, Microsoft, Meta and Amazon), underpinned by precious and scarce specialist hardware in the form of Nvidia graphics processing units. Sam Altman and OpenAI have certainly attracted a near monopoly of AI-related media coverage in the past nine months.

But, as Jaimes explained to me, there are now over 100,000 large language models in existence. The vast majority of them are open source and accessible to pretty much anyone with a laptop and an internet connection. Amid this panoply, the model currently regarded as the most powerful is Falcon 40B, an LLM built and trained by the Technology Innovation Institute (TII).

Ever heard of the TII? Didn't think so. That's probably because it is part of the UAE's Advanced Technology Research Council, under a team led by Dr Ebtesam Al-Mazrouei, based in Abu Dhabi. Dr Ebtesam and her colleagues have, like Pep Guardiola at Manchester City, assembled a formidable array of AI and machine learning talent and applied it to the challenge of building a world-leading foundation model, backed by investment-at-scale in data and compute to turn their ambition into reality.

How many people had heard of Al-Nassr until Ronaldo signed? Or Al-Hillal, Ruben Neves' new Club? Very few, I’d venture. But as in sport, Gulf-based research bodies are investing serious money into tech, science talent and facilities. The evidence of Falcon 40B suggests this strategy is already paying dividends.

Jared Cohen, ex-White House and State Department technology advisor, and the former head of Google Ideas, said recently: "Generative AI has given the US a very significant lifeline in terms of China-US tech competition.” This framing casts the AI arms race in starkly bipolar terms.

But Cohen also identifies the emergence of “geopolitical swing states” as an important new phenomenon. By training the world's most highly-rated open source LLM, the UAE has clearly staked its claim to become one such player. [Cohen’s recent essay on geopolitical swings states is well worth reading: https://www.goldmansachs.com/intelligence/pages/the-rise-of-geopolitical-swing-states.html]

I have been as guilty as anyone in overlooking the technology progress made in the UAE and other Gulf States, dismissing too readily their initiatives as PR, much as the investment by Qatar and Saudi Arabia into sport is dismissed as “sports washing”. Without wishing to downplay or diminish the very real and important criticisms of human rights abuses in these countries, perhaps we should take more seriously the strategic intent behind the “Ronaldo Effect” in sports, and its parallel in the domain of tech and AI.

Sports fans will recall the unsuccessful, but nonetheless serious, attempt to create a super league of European football clubs by a consortium supported by Gulf money. Then there’s the psychodrama in golf as LIV Golf, an upstart Saudi-backed league, shakes up the complacent established order, forcing the US-based PGA Tour into an alliance whose consequences continue to play out, not least in recent Senate hearings.

From a British perspective, the lessons from sport are clear. Complacency can be fatal, and the same is true of the technologies driving shifts in geopolitical power. Our historical role as the birthplace of both football and of computing entitles us to nothing. Nations are already engaged in an AI arms race, and in the UK a serious national-level effort is required to retain our position as a key swing state, with some measure of control over our destiny.