The inevitable destination

My short answer: AI absorbs most computer-based knowledge work, humanoid robots take on more and more physical work, and the cost of producing goods and services falls. That supply-side deflation, plus some form of redistribution such as a universal basic income, makes work increasingly optional. The direction is clear. The pace and the politics are not, and the UK has barely started planning for the jobs side.

I have been spending a lot of time recently trying to figure out where all of this actually goes.

Not what AI can do today. Not which tools are hot this quarter. But the destination. The endgame. Because when you zoom out far enough, the trajectory becomes surprisingly clear and, honestly, pretty exciting.

What follows is my attempt to map it out. Not as a prediction, but as a logical sequence of events that are already in motion, backed by data from the people and institutions closest to the action.

01How much knowledge work will AI take over?

This one is already happening, so it barely qualifies as a prediction.

AI is absorbing the majority of computer-based knowledge work. Writing, research, analysis, coding, design, customer support, data processing, financial modelling, legal review. The list keeps growing and the capability gap keeps shrinking.

The World Economic Forum’s Future of Jobs Report 2025 expects structural churn equal to 22% of today’s jobs by 2030, driven by technology and other forces: 170 million roles created and 92 million displaced. McKinsey estimates that generative AI and other technologies could automate work activities that absorb 60 to 70% of employees’ time today. Harvard Business School researchers found that since ChatGPT launched, job postings for structured, repetitive roles dropped 13%, while demand for analytical, technical and creative work grew 20%.

The economic projections are staggering. McKinsey puts the total economic potential of AI at $17.1 trillion to $25.6 trillion a year. Goldman Sachs projects that generative AI could raise global GDP by 7%, almost $7 trillion, over ten years. PwC’s 2017 “Sizing the prize” study puts AI’s contribution at $15.7 trillion by 2030. Nobel laureate Daron Acemoglu at MIT is far more cautious: in The Simple Macroeconomics of AI he estimates that only about 4.6% of all tasks will be affected by AI within ten years, and calls the effects “nontrivial but modest”.

Meanwhile, the people building these systems are not being coy about timelines. Anthropic CEO Dario Amodei has warned that AI could wipe out half of entry-level white-collar jobs within one to five years. Elon Musk has said that work will become optional. And according to Morgan Stanley research reported by Fortune, by the fourth quarter of 2025, 30% of AI-adopting companies were already seeing quantifiable benefits, up from just 16% a year earlier.

“They commodify human expertise, rendering it less valuable and potentially superfluous.”

Acemoglu, Autor & Johnson, on pure automation technologies, Hamilton Project, 2026

Acemoglu, writing with David Autor and Simon Johnson in Building Pro-Worker Artificial Intelligence for the Hamilton Project in February 2026, warned that the current wave of AI development risks following an automation-first path rather than an augmentation-first path, amplifying inequality rather than reducing it.

Whether you think this is five years away or fifteen, the direction is not in question. The only debate is pace.

02Are humanoid robots already here?

While we have been arguing about chatbots, the hardware side has been moving quietly and fast.

Humanoid robots are no longer a concept. They are shipping. Barclays estimates there were roughly 15,000 humanoid robot installations in 2025, more than 85% of them in China, according to CNBC.

Let us do the maths on that. A $20,000 robot that works 24 hours a day, 7 days a week, for 4 years works out to around $0.57 per hour. Compare that to the US federal minimum wage of $7.25, the UK National Living Wage of £12.71, or the UK median full-time hourly pay of £19.67. The economics are not close.

The scale of what is being planned is extraordinary. Morgan Stanley projects more than a billion humanoid robots in use by 2050, in a market worth $5 trillion.

China’s dominance in this space is not accidental. Barclays puts China’s share of 2025 installations above 85% (CNBC), and TechCrunch has set out why China’s humanoid robot industry is winning the early market.

“This is not science fiction. This is supply chain logistics.”

Baz Furby

03Will AI make everything cheaper?

Here is where it gets really interesting.

When AI handles knowledge work and robots handle physical work, both powered by increasingly cheap renewable energy, the cost of producing almost everything falls dramatically.

Sam Altman has been saying this plainly. At a town hall in January 2026, he said AI is going to be “massively deflationary”. Vinod Khosla has said AI should be “hugely deflationary” over twenty-five years.

Writing in the IMF’s Finance & Development magazine, Nobel laureate Michael Spence argues that AI could relax supply-side constraints on the global economy, the kind of constraints that push prices up.

Think about what deflation actually means in practice. Healthcare becomes cheaper when AI handles diagnostics. Construction costs drop when robots do the building. Legal fees fall when AI drafts contracts. Food costs decline when autonomous farming and logistics remove human labour from the chain.

Everything gets cheaper. Not because demand collapses, but because the cost of producing things falls through the floor. This is supply-side deflation, and it is fundamentally different from the demand-side deflation that economists fear.

Software is the clearest precedent. The US Consumer Price Index for computer software and accessories has fallen by more than 70% since 1997.

There are credible counter-arguments here. The transition itself may push some costs up before it pulls them down, as companies invest heavily in AI infrastructure.

04Where does income come from?

So if robots make things and AI runs the systems, and everything is cheaper, where does income come from?

This is where Universal Basic Income stops being a fringe idea and starts being an inevitability.

The logic is straightforward. Companies using AI and robots generate massive profits from dramatically lower production costs. Governments tax those profits and distribute the proceeds. Citizens receive regular payments that, combined with deflated prices, provide a comfortable standard of living.

This is not a radical new idea. In Moore’s Law for Everything, Sam Altman proposed an American Equity Fund, paid for by taxing companies above a certain valuation 2.5% of their market value each year, in shares, plus 2.5% of the value of privately held land. Bill Gates argued for taxing robots that take human jobs back in 2017. The LSE has published research arguing UBI is the logical new social contract for the AI age.

And it is being tested. Mayors for a Guaranteed Income had grown to 160 mayors by 2024, and a bill in the US House, H.R. 5830, proposes $495 million a year for a federal guaranteed income pilot from 2026 to 2030.

The funding question is real. A meaningful UBI in a country the size of the US costs trillions a year. But the maths changes when deflation cuts the cost of living.

The pieces are already moving. Slowly, politically, messily. But they are moving.

05Does work become optional?

This is the part that makes people uncomfortable. Not the economics, but the philosophy.

If AI handles the knowledge work and robots handle the physical work, and a combination of deflation and redistribution means basic needs are covered, what do people actually do?

The Japanese concept of Ikigai (finding purpose through chosen work rather than obligated labour) becomes the dominant framework.

But I would push back on the doom framing here. Humans did not stop creating when we automated farming. We did not stop inventing when we automated manufacturing. Every wave of automation has freed human energy for higher-order work.

The difference this time is speed. Previous automation waves took decades to unfold. This one is compressing into years.

06What question are governments not asking?

Here is my concern. Not about the technology. Not about the economics. About the conversation.

Most governments are not having it. And as someone building AI products in the UK, I feel this acutely.

The UK has an AI Action Plan. It talks about growth, investment, skills pipelines, and positioning Britain as a global AI leader. But it says almost nothing about what happens when the jobs actually go.

Meanwhile, the US has framed AI as a national security priority, explicitly aiming to win the AI race against China.

And the UK? We are still debating copyright law for AI training data. In March 2026 the government said a broad copyright exception with an opt-out is no longer its preferred way forward, after sustained pushback from the creative industries.

The technology is moving on a timeline of years. UK policy, constrained by fiscal pressures and political cycles, operates on a fundamentally different clock.

“We are not ready for this conversation, let alone the answer.”

Baz Furby

07So where does this end?

Honestly? I think it ends somewhere good. Not perfect, not without turbulence, but fundamentally good.

AI and robotics will make the world more abundant. Things that are expensive today will become dramatically cheaper. The question is not whether this happens, but how we manage the transition.

Start with science. AI is already changing drug discovery. In Benchling’s 2026 Biotech AI Report, half of the organisations surveyed said AI had already sped up their time to target, although the report describes progress as broad rather than deep.

Think about what that means for healthcare. Faster drug discovery. Personalised treatments. AI diagnostics that catch diseases earlier. The cost of healthcare drops while outcomes improve.

Now apply the same logic to climate and energy. AI-optimised power grids. Autonomous systems managing renewable infrastructure. Carbon capture at scale.

Infrastructure gets rebuilt. Not in decades, but in years. Autonomous construction, AI-planned logistics, robotic maintenance.

The surplus created by automation will, eventually, find its way to people through some combination of UBI, deflation, and new economic models.

“The destination looks good. But the journey needs a plan.”

Baz Furby

As a founder building AI products, I think about this constantly. We are not just building tools. We are building the infrastructure for a world that works fundamentally differently. If your business is navigating this shift and you need an experienced hand on the wheel, explore working with us.