Most Valuable Painting - Art exhibition on Feral File

Nora O Murchú on Jonas Lund’s MVP (2022, 2026) for Feral File

The tricky thing with belief in a system is how quickly it can break. I have been online for most of my adult life – probably like most of you reading this – and in this time I have witnessed the rise and fall of platforms, technologies that promised transformation, and hype cycles that boomed, burst, and repeated.

Some years ago, I used to work as an ATM switch engineer in a company in Fremont in the East Bay region of the Bay Area. It was my first technical job, my first boom, and my very first economic crash. Every morning I got on the BART and commuted for an hour to the company headquarters – an air-conditioned, tinted-glass building on a tree-lined boulevard. It was inside this building I encountered an early version of the tech culture that was soon to proliferate globally and unsettle the economic and social logic of everyday life: I would move from my cubicle to a lab testing cloud infrastructure, to standup meetings, to unenthusiastically participating in pizza Fridays against the backdrop of an industry convinced it was building the future. Here I learnt how the backend of the internet operated, running tests between software and hardware, and checking the results against the physical switches in the lab – a suite of machines that formed part of the infrastructure supporting what was then being sold to us as the World Wide Web. I worked there for nearly a year but returned home to finish my degree, not thinking so much about the place I had left behind. It wasn’t until a few months later on an island some 8,000km away that I thought back to the smell and the hum of those machines as I watched the news and the fallout of the dot-com bubble crash.

Triggered by the rise and fall of technology stocks, the early days of the internet were flashy and fiscally irresponsible. Venture capital funding for startups was readily available and investors poured heavily into internet start ups with the hope that one day they would become profitable and for fear of missing out on cashing in on the growing use of the internet. With money being thrown at the sector, start-ups raced to scale quickly, spending heavily on marketing to establish their brands. Growth over profits became the prevailing logic, with companies offering products and services for free or at a discount on the assumption that an audience could be turned into profit later. The more people invested, the more others followed, with decisions increasingly shaped by what everyone else was doing.

This speculation fuelled exponential growth in the market and the Nasdaq index skyrocketed, experiencing a five fold increase, peaking in March 2000 before plummeting by nearly 78% by October 2002. As investment funds dried up, many companies with high valuations and little to no profits, riding the wave and hype of the new technology, folded or filed for bankruptcy after burning through massive amounts of capital. Of course some companies survived and, despite the massive loss in fortunes and the lessons supposedly learnt, the crash left behind an infrastructural legacy of dark fibre and hardware that inadvertently laid the physical foundation for the internet we have today, and a culture in which speculation and risk could still yield value.

I survived that crash, and once you survive one crash, you can survive them all. Right?

Speculation and hype about technology is nothing new. You must have experienced at least one? We are currently in the middle of an AI boom – an old technology whose spectre has returned from its own summer and winter cycles – repackaged and rebranded to automate us into our futures. While technically still early days for AI adoption, AI-related valuations have never been higher, despite the overhyped promises of what AI can deliver. These cycles of technological hype are familiar: think back four years and you might remember NFTs, or think a little longer and you might remember blockchain, or even an earlier version of artificial intelligence. Cycles come and go, leaving behind winners and losers, their infrastructures and systems, and the possibility that maybe, next time, it will be different.

But I’m not here to discuss the rise and fall of markets – there are others far better equipped to deal with Capital than me. Instead, what interests me is what we do when we find ourselves inside one, and what happens to our understanding of what we value along the way. When does something become valuable? And is being valuable the same as being financially worth something? And when did betting on what might happen next become such an ordinary way of relating to the future?

The thing with art made in relation to technology is that it often gets caught up in cycles of attention, with particular technologies demanding our focus at particular moments. As curators and artists, we inevitably get pulled into these cycles too. As a curator working in this field for some time, I’m often asked what I think about the latest technology, or whether my programme will engage with it, and often I find myself seeking out practices that engage with and critique the current technological moment. I am not cynical about these questions, but I do recognise them as demands that are never just about the technology itself, but about the aspirations, fears and fantasies that surround it. They are shaped by markets and flows of capital, by decisions about what matters and deserves our attention, and ultimately by what becomes valuable. I remember when NFTs first appeared and how quickly the conversation moved between the possibilities of a new market for digital art, the money being made, and questions about the environmental cost of the infrastructures supporting it. Four years later, many of those same questions have attached themselves to the vast extractive infrastructures supporting artificial intelligence. But being caught in these cycles doesn’t always mean accepting their terms. Whether you agree or not, technologies caught up in moments of hype can be material for experimentation: ways of working through the systems that produce attention, speculation and value from within. Sometimes this means adopting the very logics you are trying to examine, allowing them to operate inside the work itself.

When I think about how value and speculation operate through art, I often return to the work of Jonas Lund, whose practice I have followed since 2012. His practice has, for as long as I can remember, laid bare the logics of markets and capital, using their systems and mechanisms as part of the work itself, while asking questions about the nature of participation as it is translated into data: a click, a like or a glance becomes a measurable signal through which attention is quantified, preferences are produced and decisions about value are made.

In 2022, at the peak of NFT hype, these questions converged in MVP (Most Valuable Painting), a work commissioned by Aorist that set out to produce 512 generations of digital paintings through a process of optimisation and collective decision-making. The premise was relatively simple: could a painting optimise itself towards becoming the most valuable? And, if so, who or what gets to decide what valuable means? So Lund set out on an experiment — by now a core characteristic of his work — in which 512 generations of paintings would evolve through an algorithm that responded to likes, clicks and other forms of engagement. This idea had its roots in earlier works such as Viewer Improved Painting (2014), a self-optimising painting that used eye-tracking to measure where viewers looked, continually adjusting its composition towards what Lund described as the most viewed and attention-grabbing image. With MVP, looking became only one part of a much larger system of measurement where each painting was assigned a fitness score based on its ability to attract attention and engagement. Taste became data, and individual choices accumulated into something like a collective preference. And so, Lund’s rules for MVP formed a system that measured, responded and determined what happened next. The better a painting performed, the greater its influence on what came next.

MVP launched with Aorist in 2022, beginning its journey through 512 generations towards the final Most Valuable Painting. But optimisation was only one measure of value within the work. As the generations increased, so too did the price, bringing questions of taste and collective preference together with what someone was prepared to pay — and what they imagined that value might become. By Generation 449, the price had reached $2,048. But before MVP could go any further, Aorist stopped operating, leaving behind an unfinished work for Lund. For all the belief in alternative markets at the time, MVP held these different ideas of value in tension: a collective decision-making system built around something notoriously difficult to quantify — taste — while at the same time participating in the speculation and hype of the NFT market itself.

For the past four years, MVP has remained at Generation 449. For some, it may be a cautionary tale about the promises and failures of the NFT boom. And yes, in lots of ways MVP tells the story of a boom-and-bust cycle. But in our current technological moment, I think it also tells us something about what happens to value when the conditions through which it was produced have changed. Paused in 2022, MVP in 2026 enters a very different technological landscape, where the systems of measurement and optimisation it was built around have become increasingly bound up with questions of governance, prediction and speculation. The logic of optimisation that is inherent to MVP has certainly not gone away. We still live through systems that measure, rank and respond to our behaviour, despite the logic of these algorithms becoming more opaque. But the culture of optimisation that has defined so much of the last decade of technological progress has given rise to a culture of technical governance, increasingly operating as a way of deciding what gets seen, what gets prioritised, where resources go, and what actions should be taken. At the same time, speculation has moved further into everyday life in what Ann Pettifor describes as a “global casino”, where remote financial markets speculate on everything from food and energy to housing and pensions, determining the costs through which everyday life is organised. For Pettifor, this casino is deliberately constructed through the management of speculation, while the risks it produces are passed on to “distant others” who may have had no part in making the bet. Perhaps nowhere is this more literal than in the rise of prediction markets like Kalshi or Polymarket, where almost anything that might happen next can become something to bet on. Earlier this year, in the hours before the US and Israel began strikes on Iran, a number of newly created Polymarket accounts began betting that the US would attack by February 28. Six accounts identified by blockchain analysts made around $1.2 million from those bets; most had been funded within the previous 24 hours and had no history of betting on anything else. To this day, no one has established who was behind those accounts or what data they had. But the timing raised an obvious question: what happens to the idea of collective prediction when some of the people participating might know something that everyone else doesn’t?

It is into these changed conditions that MVP now returns. Four years after it stopped at Generation 449, Lund is completing the work with Feral File, but the conditions through which its final generations will be produced are very different from those in which it began. This time, MVP begins again with a Dutch auction. Where the price rose as the generations advanced in 2022, this time it moves in the opposite direction, starting at $8,192 and falling to $2,048. So the speculation remains, but without the momentum of the NFT market that surrounded the work in 2022. At the same time the question of value is left more openly with those who choose to participate. What you are buying is a position in the sequence: the earlier you buy, the closer your generation is to MVP 1 – the final painting, Generation 512 – and the later you will choose from the paintings the system has evolved. How long are you prepared to wait? What are you prepared to pay? And how much does it matter to you where you end up?

And yet there is one final contradiction built into MVP. After 512 generations of measuring, selecting, buying and optimising towards the Most Valuable Painting, MVP 1 can never be bought or sold. The work arrives at its final measure of value at precisely the point where price can no longer determine it.

I don't think MVP can tell us what is valuable. Nor do I think the work tells us what the Most Valuable Painting should be. Instead, four years later, I think of it as a question. A question about what happens when taste becomes measurable, when individual preferences become collective decisions, and when those decisions are fed back into systems designed to optimise what comes next. It asks who gets to decide what counts, what kinds of collective decision-making we want to participate in, and what happens when the values of a system are different from our own. And perhaps, after all the optimisation, speculation, prediction and hype, the fact that the Most Valuable Painting is the one painting that cannot be bought leaves us with a much simpler question: what do we value?