A Better Way to Think About Tech

Cal Newport says technology has become difficult to discuss because commentators increasingly split into two extreme camps. Techno-pessimists emphasize harms and the values new tools may destroy; he uses the early-1800s Luddites, who opposed automated manufacturing such as the steam loom and said they would rather die than see hand weavers’ families starve. Cal argues their immediate concerns were real, but the Industrial Revolution’s long-term gains created much of modern wealth. Techno-optimists make the opposite mistake by dismissing present harms for a promised technological utopia, illustrated by Elon Musk calling humans “biological bootloaders” for digital superintelligence.

Cal proposes “technoselectionism” as a third position for people who want technological progress without treating every innovation as inevitable. Using a hand-drawn map, he places the current moment in one corner and a better future in another. Pessimists circle near the starting point to avoid snakes, fire, and other dangers; optimists draw a straight line through every obstacle. The technoselectionist instead steers around as many hazards as possible. In his 2023 New Yorker definition, Cal says innovation can improve life while some inventions cause more harm than good, so society should encourage experiments but curate adoption intentionally.

The first two of Cal’s four “Cs” are choosing and configuring. Choosing means popularity does not obligate anyone to use a technology: Cal cites Brad Stulberg deciding that LLM efficiencies were not useful for his writing and coaching, and Cal’s own decision in the 2010s not to use social media after weighing benefits against harms. Configuring means setting personal rules rather than treating tools as all-or-nothing. Examples include using Facebook only for necessary groups while hiding the feed, checking Instagram only on Fridays for 20 minutes to follow visual artists, or having a GoPro and employee handle Instagram uploads.

Cal’s third C is changing: starting to use a technology does not create a permanent obligation to keep using it. He says social platforms changed after TikTok became popular, shifting from showing posts by chosen friends and accounts toward endless algorithmic recommendations designed to maximize engagement. People who joined Facebook or Instagram for a selected social network may now find themselves trapped in addictive, unwanted feeds and are entitled to leave. He applies the same logic to AI users who move from useful automation to cognitive surrender, losing the ability to write simple emails or experiencing blurred boundaries between fantasy and reality. His dating analogy: attraction can end when the Pokémon-card binder appears.

The fourth C is complaining: individuals should tell schools, workplaces, and communities which technologies they oppose and why. Cal argues institutional change usually begins with people saying, “I don’t like this for me” or “I don’t like this for my kids.” He attributes 2025’s spread of bell-to-bell school smartphone bans to parental pressure rather than a sudden discovery by school administrators. He also says that in 2026, lawmakers and parents in 16 states proposed limits on students’ use of non-phone educational screens. Personal choices therefore become collective pressure that can redirect institutions and culture.

Cal says discussion of AI increasingly resembles a “techno discourse prison yard,” with people pressured to join either the enthusiastic disruption camp or the camp rejecting AI altogether. He cites exaggerated reporting about a PHP cybersecurity crisis: vulnerabilities were found using familiar tools, but the involvement of an AI model made the event seem like proof that AI was breaking the internet. He also warns that individuals gave new AI agents access to banking systems, creating a real but self-inflicted security risk. His proposed alternative is to say which uses work, which cause harm, and to reject both hype and blanket dismissal.

In conversation with Jesse, Cal explains that he uses chatbots as imperfect technical manuals. They help troubleshoot poorly documented open-source software for his immersive Halloween display, which uses a Raspberry Pi, an isolated Ethernet network, a separate wireless network, and lighting controls. The models often hallucinate nonexistent menu items and sent him down wrong networking paths, while a more aggressive search tool gave the correct fix: unplugging and reconnecting Ethernet to reset the configuration. Cal compares the chatbot to a knowledgeable but not especially smart friend. He also says his scripts are AI-generated and voiced with ElevenLabs, though he bought sound effects when AI-generated ones were inferior.

Answering Robert, Cal agrees that LLMs can be configured for learning rather than cognitive offloading: users can ask for quizzes, explanations, or practice questions instead of answers. He recommends grounding a chatbot in a source page because it confidently produces errors. Responding to medical professor Augustino, he accepts that avoiding difficult work may weaken not only cognition but also resilience and mental health; convenience is not automatically beneficial. Jay’s story supplies a positive social-media example: after quitting platforms, he read newspapers to stay informed and then developed a rewarding reading habit to fill the remaining time. Cal calls this a deliberate redesign of one’s technological environment.

James sends an article about New Yorker writer Lawrence Wright, whose research system for The Looming Tower consisted of many handwritten notes in labeled manila folders, including sections on Osama bin Laden, al-Qaeda, health, Syria, and jihad. Cal contrasts this with the current pressure to build an AI agent, cloud database, and retrieval system that automatically organizes research. Wright’s acclaimed book, he argues, depended less on research-management efficiency than on touching ideas, moving papers, and finding the right note at the right moment. Cal makes the same point about index cards used by writer Ryan Holiday: slowness can help ideas settle in the mind.

Cal reports finishing five October books: Simplify by Jim VandeHei, Mike Allen, and Roy Schwartz; Telephone by John Brooks; Roll the Calls by Ari Emanuel; Cancel Me If You Can by Dave Portnoy; and The Treeline by Ben Rawlence. He praises Emanuel’s ghostwriter but notes that the memoir focuses on Emanuel’s life rather than detailed stories about clients such as Mark Wahlberg, unlike Mike Ovitz’s Hollywood-centered memoir. Portnoy’s book recounts events such as the Roger Goodell protest but offers less business strategy than Cal wanted. The Treeline is a deeply researched account of forests’ northern limits and their movement amid warming. Cal also previews a lighter studio backdrop and his completed Halloween display.