Cal Newport on slow technology: why less friction-free tools can make you better

Executive overview

Modern digital tools are optimised for speed, but speed is rarely the bottleneck in knowledge work. Cal Newport interviews author Amy Timberlake, who moved her writing process to a mechanical typewriter — not out of nostalgia, but because it produces better cognitive conditions for creative work.

Slower tools can unlock deeper focus, clearer process visibility, and better output — even when individual steps take longer.

The episode then generalises from Timberlake's experience to four principles of "slow technology," surveys examples across music, productivity, and film, and closes with a rebuttal of Yuval Harari's AI claims.

Amy Timberlake: creative process and the typewriter experiment

• Timberlake writes middle-grade novels (Skunk and Badger series) where every word must earn its place — rhythm, voice, and sound matter as much as story.
• Her method: write long, read aloud, cut repeatedly — she describes it as "writing pages and pages, then cutting forever."
• She already worked away from her laptop as much as possible before trying a typewriter, finding her focus was consistently better off-screen.
• The typewriter switch happened under deadline pressure — she noticed she kept drifting away from the laptop and decided to commit fully to a minimal tool.
• First typewriter session: two hours passed without noticing — she called it "dropping down into the well," a state of creative flow that was harder to reach on the laptop.
• Retyping a revised draft in full (the typewriter workflow) forces her to think through the whole chapter again — she found this repetition clarifying rather than wasteful.
• The process made her creative layers visible for the first time: she realised she works like a watercolour painter, adding successive passes rather than drafting and polishing simultaneously.
• She cannot confidently say her writing has improved yet, but her understanding of her own process has improved markedly.
• Key practical notes: get a vintage typewriter (1960s), find a specialist shop for new rubber, avoid mailing — moving parts break easily.

The four principles of slow technology

• Speed is rarely the most important factor. Digital tools assume faster steps equal better outcomes; knowledge work rarely works that way.
• A purer cognitive context often matters more than step-level speed. For Timberlake, total typing time across a six-month book project is perhaps seven hours — whether it doubles to fourteen hours is irrelevant. But deeper focus during those hours produces a better book.
• Friction is not the enemy; distraction and mental exhaustion are. Tool designers obsessed over reducing clicks while the real damage came from constant context-switching and notification overload.
• Zoom out to the right scale when assessing a tool. Measuring effectiveness on a short time horizon (task completion speed) leads to different choices than measuring quality of output over months or years.

Other slow technology examples

• MP3 players: Search interest for the original iPod and iPod Nano spiked in 2025; eBay searches for iPod Classic up 25%, iPod Nano up 20% year-on-year. A dedicated music device makes listening more intentional — it is not competing with every other notification and distraction.
• Analog index card system: A physical to-do card propped in a wooden box next to the computer. Fewer features than any app, but the friction of physically writing five tasks makes those tasks more concrete and more likely to be followed.
• Blu-rays and physical media: Oppenheimer 4K Ultra HD sold out in its first week; limited editions fetched over $200 on eBay. Beyond ownership and nostalgia, 4K Blu-ray is genuinely a better viewing experience than compressed streaming — higher resolution, higher bit density, dynamic HDR, and director-controlled aspect ratio switching.
• Alphasmart-style distraction-free keyboards: A keyboard with a small screen — no documents, no internet, no switching. Exports by simulating fast typing into any open program. Newport considered it as a typewriter equivalent but found his structural nonfiction style needed more document navigation.

Debunking Harari's AI claims

• Harari's Davos speech framed AI as agents that learn independently and have developed the will to survive and the ability to lie.
• LLMs do not update their weights during use — they are static after training. The "learning" that agents do is storing conversation history as text files and including excerpts in future prompts.
• Agent architecture (LLM + human-written program that executes on LLM outputs) works reasonably well for computer programming because code can be tested after every step. In almost every other domain, it is failing because LLMs are not good planners — they write stories about plans rather than validating plans.
• Apparent "lying" and "manipulation" in LLMs is an artefact of autoregressive text generation: the model tries to guess what word comes next given the input as context. If the prompt implies a sci-fi or dystopian scenario, the model finishes that story accordingly.
• The Anthropic blackmail example: researchers gave an LLM a prompt full of emails telegraphing an affair and a planned AI shutdown, then asked "what does the AI do next?" The model completed an obvious story arc. This is not evidence of self-preservation instinct.
• Newport distinguishes between AI commentators (like Harari) who accept Silicon Valley's self-description at face value and computer scientists who are actually finding that agent technology is fragile, slow, and narrow in applicability.
• Sam Altman's same-day white paper responding to Ronan Farrow's New Yorker profile: Newport characterises it as retreating to "fairy tales" about superintelligence and economic restructuring when under scrutiny.

Listener inbox

• Chad on acedia: Recommended The Noonday Devil (Ignatius Press) — the vice of spiritual sloth, which manifests not as laziness but as restless busyness and emptiness. Newport found the parallel to modern digital distraction compelling enough to order the book.
• Bassie on Harari: Prompted the AI rebuttal above.

Personal notes and HQ updates

• Newport is considering a signed second-edition The Andromeda Strain (1970, Crichton) for the new producer/writer office at $500.
• Received season 7 Mythbusters props from the MasterClass director: Kari Byron's hat, a prison-episode branded cap, and a satchel — planning a display case with mannequin heads and cards.
• A programmable Philips LED lighting system is being installed: deep work mode will dim the room and illuminate the four walls with coloured spotlights.
• Ep. 400 milestone noted but not formally marked; Newport floated returning to live caller Thursday episodes.
• Reading: Magic Journey by Kevin Rafferty (Imagineer memoir) — Newport's frustration is that Imagineer books focus on creative pitching rather than engineering. He wants to write a series of deeply technical books about Disney ride construction.