Write-up: Signals from the Future

人类对变化的线性思维低估了AI的变革速度

Mathematician Hannah Fry points out that the human brain, due to evolutionary reasons, struggles to assess technological change, and the success of large language models precisely illustrates this underestimation. At the Scottish Mortgage "Change Drives Growth" digital conference, manager Tom Slater explored this theme with Fry. Fry explained that humans did not evolve in an environment of rapid technological progress; we tend to assume that "what is true now will still be true in the future." She used LLMs as an example: "Nobody really expected this stuff to work quite as well as it has." AI researchers once widely believed that "language alone cannot produce intelligence," but LLMs have upended that assumption.

AI将彻底变革药物发现与材料科学的搜索范式

Fry argues that AI's impact on drug discovery is "Everything, frankly," with protein structure determination serving as the core breakthrough. She describes how protein function in the human body depends on shape: "Your body is playing LEGO." Determining a protein structure once took years; AI can reduce that to days. Tom Slater directly links this to Scottish Mortgage's holdings: Moderna uses mRNA technology to design drugs, while Enveda applies AI to screening potential drugs from the natural world. In materials science, humans have historically relied on experimentation and luck to discover useful materials, but the number of possible atomic combinations is "mind-bogglingly large," and AI is well suited to exploring that space. Fry comments on current lithium-ion batteries: "They’re the best we have so far, but they are rubbish……about a tenth of the energy density of butter." Even relatively modest improvements could extend vehicle range, help robots operate longer, and make renewable energy easier to store. Slater connects this trend to holding CATL (the world's largest manufacturer of EV batteries by usage in 2025)—advances in battery technology directly expand its addressable market. He emphasizes: "Batteries don’t just change cars, they can change what machines can do."

技术落地需以人为本，独行者才定义变革的回报

Fry uses autonomous driving as an example, arguing that prioritizing technology while ignoring human weaknesses leads to failure; design should compensate for, rather than fill, human shortcomings. When a machine encounters an anomaly and hands control to a human, "Human attention wanders, situational awareness fades and performance can deteriorate under pressure." She contrasts this with an alternative approach: first ask where humans struggle, then have the system automatically fill the gap (e.g., automatic emergency braking). Tom Slater summarizes Scottish Mortgage's investment logic: "It’s the small number of exceptional companies that have driven a large part of our [shareholder] returns." Many investments will not pan out as hoped, but "a fundamental part of the way we think about the world in Scottish Mortgage" is that a few big winners can have a disproportionate impact.

投资含义

The article points to a clear investment direction: focus on scientific fields where AI can materially accelerate breakthroughs (pharma, materials, energy), and prioritize bets on companies addressing key technology bottlenecks (Moderna, Enveda, CATL). Institutional perspective bias: Scottish Mortgage itself holds the above positions, so the article systematically uses Fry's arguments to endorse their holdings. Readers should note this is a holder's perspective, and the narrative implicitly embraces a high-volatility strategy where "a few big winners are enough to cover many failures."

Position Moves