🇬🇧 Seven exponential policies for a new government
There is an election in the UK. Here is what a new government should do.
So there is a General Election1 in the UK on the 4th of July.
As our American readers sit down to barbecue with friends and family on Independence Day, voters in the UK will be finishing casting out votes for MPs from one of several parties. They include the incumbent Conservative party, in power since 2010; the opposition, Labour party; the Liberal Democrats, the long-standing third party which has no prospect of forming a majority; the right-wing upstart, Reform; or a regional or fringe group. The UK is also famous for comedic candidates, such as Count Binface.
The pollsters put Labour wildly ahead of the Conservatives. But elections are unpredictable and six weeks is a long time in politics.
The government we choose will navigate the critical years of the exponential transition: that short period until 2029.
They’re crucial five years. We are already well into the exponential transition. There is momentum behind the development and deployment of AI and the energy transition.
2029 will look very different to 2019.
If the new government doesn’t act proactively, the exponential gap between the potential of technology and the institutions that help us make sense of the world will grow.
What should be the focus of this government? After 10 years of analysing the development of exponential technologies and their social implications, I want to share my view in this post.
Even if you’re not in the UK, you are likely impacted by an important election this year (and we’re all in one way or another impacted by the US elections later in the year). This assessment is relevant for you, too, no matter where in the world you live.
So here are seven things the next government has to get right to succeed in the exponential age:
One: Plan for an economy of learning
Technology is supplanting resources as the driver of our economies. That is technology, in its fundamental sense: compounded knowledge, iterative learning, improved price-performance.
In 1971, 17% of the S&P500 stock market value was in intangibles. By 2022, it was 90%. Resources are not useless, but they are a lot less valuable than they used to be. And the UK economy has few, if any, big winners in this new economy. Data from UBS shows how a century ago the US and UK markets were anchored by the dominant technology of the time, rail. Scroll forward to today, America has leant into the 21st century, Britain is stuck somewhere else.

An economy of learning is based on a basic understanding of learning rates: technologies start expensive and get cheaper. But they suffer from higher start-up costs and a different risk profile to that of an economy that is simply extracting copper, cotton or sugar.
Specific interventions would be:
In the energy, manufacturing and technology sectors, different stages of technological maturity need capital. But each stage has a different absolute requirement, as well as risk profile and pay-back period. Policies need to open finance for both the expensive first-of-a-kind and decreasingly costly nth-of-a-kind assets that take innovations affordably into every part of the economy. The exponential transition involves many of these leaps as we build new assets for new industrial processes (like carbon nanotubes, batteries, fusion reactors, geothermal). There is a financing gap that policy could help fill — and ultimately tease private capital into this growing market.
Many technologies, such as chips, solar, heat pumps, bioreactors can benefit from learning curves. As we produce more, we become smarter at making them and prices fall. The learning curve is a thing of beauty. Market structures and regulations can influence learning rates. Government should signal (and stick to) stable legal and regulatory conditions that foster rapid learning rates in new technologies. This approach makes them affordable for the general public, delivering widespread welfare benefits. In addition, it will create new opportunities for entrepreneurs. I explain this in Financial Times here.
Further encouraging private sector, especially pension funds, to invest in private assets like venture capital. It is important to protect and strengthen the existing incentives for investment in innovation.
Enhancing support for entrepreneurship and the specific dynamics of younger firms in terms of raising capital, employment and talent. For example, make it easier for innovative companies to recruit from abroad.
Two: Integrate AI
Artificial intelligence is a tool. And over the course of the next Parliament, it can be deployed to boost workers’ skills, human capital and productivity. Study after study has shown that the augmentation of workers by AI improves their performance (although more work has to be done to understand this).

The famous BCG/HBS study showed that the use of AI in knowledge work lifted the below-average worker into the second quartile. Across the economy, even a fraction of this would amount to significant human capital added.
So, what would be the steps for the next government to take towards integrating AI?
First, change the narrative around AI. Start the conversation about the real potential the technology has to improve people’s economic and life outcomes: by helping us earn more, by eliminating friction and volatility in the economy, by shortening waiting lists for healthcare, by deepening the provision of education, by proactively monitoring crucial infrastructure like water and sewage systems.
Second, integrate technology thinking (not just the technology!) across government to improve its effectiveness, service delivery and efficiency.
Third, changing school curricula to include core skills (I call them ur-skills) that are required to make the most of these tools. These skills include
Problem-solving: analysis and logical thinking, critical thinking and textual analysis, quant and stats skills, creative thinking,
Collaboration and judgement: task delegation, machine understanding, judgement and evaluation.
Pupils today may pick these skills up as ancillary outcomes of their school education but these should become core, tested and evaluated outcomes as soon as possible.
Fourth, secure sovereign access to computing resources: ensure that government and academia have access to more computing power than they think they need to research, develop and realise the promise of AI.
Three: Energy innovation
Much of the energy transition is a financing problem rather than a technology problem. The upfront costs of heat pumps, solar or batteries are high, even if their lifetime costs are lower than fossil alternatives. The energy market is broken because it takes on the characteristics of fossil-based electricity. And skills shortages hamper the deployment of no-brainer technologies like heat pumps.
We need three additional shifts in the next five years:
A move to a regime of zonal/nodal pricing for electricity to better reflect regional differences in generation cost. This would create incentives for investment in generation and storage locally where prices of high, further driving down costs. A recent government review ditched zonal pricing. [Note: 25/5/2024 - I made an error here. The review ditched nodal pricing as too complex but zonal pricing is still on the table. My main argument holds.]
This was a mistake. The power market will benefit from a more sensible design that fits the architecture of the future energy system - which is much more decentralised, fragmented and bidirectional than today’s. Long-term thinking is needed.Encouraging the private sector to develop financing packages that enable businesses and homeowners to make sensible, long-term energy choices. Consistent, stable, long-term government signal helps.
Tackling the skills shortages (for example, in heat pump installations) by incentivising up-skilling and ensuring that homeowners understand the benefits. This will create a market for these skills.
Putting the hydrogen boondoggle to rest. The laws of physics make hydrogen unsuitable for domestic heating and most forms of transport. However, it is useful (especially solid-state modalities) for certain classes of long-term storage.
Four: Prepare people
Technological change is nerve-wracking for people at the best of times, worse when people are subjected to systematic scaremongering. After years of this, Edelman reports that more than 54% of Brits claim they would reject the growing use of AI, while only 16% embrace it.
Yet every discussion about AI risk can be flipped on its head. AI as cyber threat? How about AI as an integral part of cyber defense? AI as job replacement, how about AI as job augmentation and worker power tool? The benefits to individuals outweigh the risks, and yet the media and political narrative has largely emphasised the most unlikely bad outcomes.
Politicians need to lead from the front and change this narrative. This requires new stories (fewer front pages about existential risks) and more engagement. Specifically:
Launch a nationwide process to engage, deliberate and consult on key issues of technology change, from AI to energy transition. Such citizens’ assemblies can help identify local fears, needs and priorities, while building broader public trust in the system.
Promote concrete, tangible benefits from the application of new technologies. Whether it is key aspects of health service delivery that have been improved, or people who are better able to manage their household finances because of heat pumps or solar panels.
Encourage more discussion between bosses and workers about the potential for artificial intelligence to enhance jobs and business performance (through convening power or policy).
The basic ideas of shared facts and narratives needs to be addressed. Promoting epistemic integrity is important. I write about this in more detail here.
Five: Foster a scientific and cultural renewal
After a difficult start, the 2020s could still be remembered as another ‘roaring twenties’. Periods of technological change have been accompanied by scientific and cultural renewal. A century ago, in the shadow of a devastating war and pandemic, we saw the technological transition of the car, radio and electricity accompanied by cultural and artistic renewal. Think of the BBC, the Bloomsbury Group, importing jazz from America, developing modern physics. Intellectual endeavours don’t exist in silos.
Three specific policies will help:
Bring science and scientific decision-making closer to the fore including in the way in which government communicates more widely.
Reinforce the importance of the independence of thought that is the hallmark of science. Access and participation don’t attack the academy.
It is not the role of the state to determine artistic output but there are steps a government can take to support cultural expression and experimentation. This includes focusing on the fundamentals (liberal political, economic and social rights) that enable such activities.
Six: The dollars and sense of government
There is the challenge of how all of this will be paid for. Britain’s finances are parlous. We need to avoid the “treasury brain” that
beautifully describes: “the incredibly short-termist approach to public spending infects every area of policy-making.”Technology is the gift that keeps giving. After all, prices are falling: that’s the innovation dividend.
Much of the funding can be shared with the private sector, spreading the benefits and the risks. And a lot of it, especially when it comes to the transformation of industry and energy, is just a mismatch in terms of duration. The money is very likely to be there if we think about it creatively.
Partnering with the private sector could unlock more than the new government might expect. Going back to Edelman, Brits trust business more than they trust the (current) government and the media, so any new government will have to figure out how to take a constructive but measured approach to working with bosses.
Closing the exponential gap will also require regulatory innovation. Regulation needs to be made, adapted and withdrawn more quickly. Labour’s proposed Regulatory Innovation Office is the sort of thing that might achieve that.
There are other considerations, of course. The world is more febrile and conflictual. This does mean more investment in defence (both innovation and at scale). It also means more honesty about what that environment looks like.
Seven: Tell a great story
There is a great story to be told about the next 5 to 10 years. Well told, it could mobilise and electrify the country — literally and metaphorically.
We know that the economy is changing and where we need to go. We should be building towards an electrified economy, a decarbonised grid, novel industrial processes, heavy use of AI and computation in the economy to support all industries and functions.
We can work backwards from this and figure out what needs to be in place today and what needs to be in place tomorrow to get there. Some of this work is not very sexy, like planning regulations2 and some of it makes the headlines, like flashy technology.
But most of it matters.
The British are not optimistic about technology (who would be after the global financial crisis, austerity and the aftermath of Brexit), so we have to find the stories that can make people see that the future could be better than the past.
One of the key strengths of leaders like Blair, Thatcher and Reagan, whether we like their policies or not, was the stories they told. Just look at Reagan’s highly effective 1984 TV ad that launched his re-election campaign.
And, most importantly…
Whether you’re casting your vote in the Summer in the UK or later this year in the US, and everywhere in between, keep in mind that we are amidst a major transition, one that we can still shape the way we want it to be.
Cheers,
A
Many thanks to Kenneth Cukier for helpful comments on these ideas.
While American readers are familiar with presidential elections, a general election in the UK involves electing Members of Parliament (MPs) who will form the next government. This process determines which political party, or coalition of parties, will lead the country and shape its policies for the next five years.
Labour has already announced some proposed sensible reforms to infrastructure policy.
Hey Azeem, Josh here—we were introduced a while back via Kevin W. Given some of your recommendations, I wanted to put on your radar the movement for public AI, or AI as public infrastructure. Think of a nationally-funded general-purpose model akin to a BBC for AI, or a specialized NHS model for healthcare. Public AI builds on public compute investments, but emphasizes more public use-cases and public accessibility. E.g. look at AuroraGPT, a GPT for science being built at Argonne National Labs in the US; similar efforts in Switzerland and Barcelona. Also, ref. many existing nationally-funded language models (ref. Dutch, Swiss, Spanish, Swedish, Polish, UAE).
https://publicai.network is where some of the organizing is happening, including UK-specific advocacy.
Azeem, let me know if you ever stand as an MP, I may move to your constituency!