Scottish Mortgage is Baillie Gifford's flagship investment trust (founded 1909, LSE ticker SMT), known for its maximalist growth style — long-term stakes in Tesla, Amazon and ASML plus bold allocations to private companies like SpaceX and ByteDance. It is the UK retail investor's flagship vehicle for global disruptive growth.
This interview says AI could be as transformative as the Industrial Revolution, but most people underestimate its speed because they think linearly. Fund manager Tom Slater is bullish on AI for scientific discovery, especially in drugs and materials. Key holdings: Moderna (uses AI to design mRNA drugs), Enveda (AI searches nature for new drugs), and CATL (world's largest EV battery maker; better batteries would expand its market).
At a Glance The Scottish Mortgage research article explores why technological change is often misjudged and how to identify key signals that may reshape industries over the next decade. The core argument is that the human mind is accustomed to linear change and struggles to adapt to exponential tech
Mathematician and podcast host Professor Hannah Fry discusses with Scottish Mortgage fund manager Tom Slater why technological change is consistently misjudged. Core thesis: AI is driving the value of human intelligence toward near zero, potentially marking the most fundamental transformation since the Industrial Revolution—yet most people still perceive exponential change through a linear lens, thus severely underestimating the impending impact.
Hannah Fry argues that human evolutionary history has determined our inability to intuitively grasp the pace of technological change. Over hundreds of thousands of years of civilization, life remained largely unchanged—seasons cycled, day and night alternated, and a decade later looked much the same as today. Our thinking is naturally inclined toward "linear extrapolation": assuming that what is true now will remain true in the future.
This cognitive bias leads to two typical errors: first, a widespread underestimation of truly disruptive technologies (such as language models); second, a systematic misjudgment of technology maturation timelines. Fry cites her own experience as an example—nine years ago, when she was producing a program, nearly all AI researchers believed that "true intelligence cannot be achieved through language alone," because machines lacked embodied experience in the physical world. Yet, in reality, the performance of language models has far exceeded everyone's expectations.
Tom Slater adds: This is precisely the core of Scottish Mortgage's investment framework—seeking out opportunities for exponential change that are underestimated by linear thinking.
Fry believes that the breakthrough of AI in protein folding is the most exciting scientific advancement at present. All functions of the human body—digestion, repair, disease—ultimately boil down to a game of protein "Lego blocks": shape matching determines function. In the past, resolving the folded structure of a single protein required years of effort from scientists. Now, AI has resolved approximately 250 million protein structures and made them freely available worldwide.
This means that for diseases such as Alzheimer's, Parkinson's, and specific cancers, AI-assisted drug design suddenly possesses unprecedented capabilities for microscopic manipulation. Some researchers claim that "within a decade, all diseases will have curable drugs."
Slater notes that Scottish Mortgage has already positioned itself in this field: It has invested in Moderna (mRNA drug design) and Enveda (AI-driven search for potential new drugs in nature).
Fry argues that AI's potential in materials science is also severely underestimated. Virtually all new materials in human history (such as vulcanized rubber) were discovered by accident—"someone knocked over a vase onto something else." However, the number of possible atomic combinations is astronomical, and AI excels at searching for optimal solutions within such vast spaces.
Take batteries as an example: The energy density of current lithium batteries is only one-tenth that of butter. If battery efficiency could be improved by 20%, 50%, or even doubled, it would fundamentally transform energy storage and transportation—the core bottleneck of the clean energy transition. Better batteries would make seawater desalination feasible and clean freshwater abundant.
Slater adds: Scottish Mortgage holds shares in CATL (Contemporary Amperex Technology Co., Limited), the world's largest manufacturer of electric vehicle batteries. Any advancement in battery technology will directly expand its addressable market.
Fry proposes two technology deployment paths and warns of the dangers of a "technology-first" strategy. Early autonomous vehicles adopted a "technology-first" approach: letting the machine drive, with humans ready to take over at any moment. However, what humans are least adept at is precisely: maintaining sustained vigilance, performing well under pressure, and staying focused. The result is that—at the very moment the system most needs human intervention (a sudden incident)—humans are often distracted or overwhelmed by stress.
The superior path is "human-first": First identify what humans cannot do (maintain attention, perform under pressure), then design technology to fill those gaps. This is precisely why "guardian systems" such as automatic emergency braking and collision avoidance systems succeed—they do not replace humans but compensate for human weaknesses. When these algorithms become good enough, under certain cities and conditions, humans can already be completely removed from driving.
Slater believes that this distinction will determine the adoption speed of all automation technologies.
Fry offers three reasons why the current transformation fundamentally differs from the steam engine era:
1. Self-Improvement Capability: AI models are approaching the tipping point where they can improve themselves. Once achieved, this will trigger a "runaway effect"—models will become better and faster at an accelerating pace. Steam engines cannot improve themselves.
2. Intelligence Value Collapse: AI is driving the value of human reasoning and problem-solving abilities toward zero. This is more akin to the invention of electricity—permeating every corner of society—rather than a singular innovation like the steam engine.
3. Innovation Accelerator Effect: AI is not only advancing itself but also enabling breakthroughs in other fields—drug discovery, materials science, quantum computing, and nuclear fusion. These areas, long "stuck" in the scientific community, are beginning to see progress thanks to AI.
On how to determine whether a technology (e.g., quantum computing) is worth taking seriously, Fry proposes two balancing factors:
Slater adds: Quantum computers are almost impossible to develop outside of commercial organizations—the costs are too high. Yet their scientific potential (simulating natural processes, atomic behavior) is equally immense.
Fry predicts that ten years from now, what we are most likely to regret not seeing is that everything needs to be fundamentally rethought. DeepMind co-founder Shane Legg believes that the education system, tax system, retirement system, leisure time, weekends—the entire social structure—requires radical reconstruction.
But Fry is optimistic about this: Over the past century, the Industrial Revolution led us to mistakenly regard "intelligence" as humanity's most valuable quality. When AI reduces the value of intelligence to zero, we instead have the opportunity to rethink: what truly matters? What is the core of being human? — It is the capacity for nurturing, courage, strength, and care, not just intellect.
Slater concludes: We always underestimate the extent to which "everything will change."
| Position | Analyst View | Key Data |
|---|---|---|
| Moderna | Bullish (portfolio holding) | Uses AI to design mRNA drugs |
| Enveda | Bullish (portfolio holding) | AI searches for potential new drugs in nature |
| CATL (Contemporary Amperex Technology) | Bullish (portfolio holding) | World's largest EV battery manufacturer |
| Lithium-ion batteries (industry) | Risk warning (technological bottleneck) | Energy density is only 1/10 of butter |
1. Fry: The human mind is a "linear extrapolator" incapable of understanding exponential change. Hundreds of thousands of years of evolution have conditioned us to believe that "what is real now will remain unchanged in the future"—this is the root of all technological misjudgments.
2. Fry: Nine years ago, most AI researchers believed that "intelligence cannot be achieved through language alone"—yet language models have far exceeded everyone's expectations. This is a classic case of "expert consensus being wrong."
3. Fry: AI has already analyzed approximately 250 million protein structures and made them freely available, whereas in the past, analyzing a single structure took years. Drug discovery is transitioning from a "craft workshop" to an "industrial era."
4. Fry: The energy density of lithium batteries is only one-tenth that of butter—"they are the best, but they are terrible." A 20-50% improvement in battery efficiency could reshape the energy storage landscape, and AI-driven materials discovery may deliver a breakthrough.
5. Fry: The key to technology adoption is not how good the technology is, but whether it is "human-first" or "technology-first." Early autonomous driving failed because it forced humans to take over under pressure—something humans are least suited for.
6. Fry: This time, there are three fundamental differences from the steam engine era—AI can improve itself, the value of intelligence drops to zero, and AI accelerates innovation across all other fields. The steam engine did not improve itself; electricity permeated everything—AI is more like electricity.
7. Fry: To judge whether a technology is worth taking seriously, look at two factors: the "scientific upper bound" and the "commercial incentive." Quantum computing should be physically possible, but it requires massive investment from commercial organizations to be realized.
8. Fry: When AI reduces the value of "being smart" to zero, we have an opportunity to redefine what truly matters as human qualities. Over the past century, we have over-rewarded "intelligence" while neglecting equally important qualities such as nurturing, courage, and care.