Speed as a promise

Season 1, episode 1. Does AI save us time – and if so, where does it go? A sociologist of time and an AI lead argue about twenty emails instead of ten, the down escalator, the infinite workday, the shrinking present and who pays for saved time.

Chapters
00:00 Cold open
00:17 Welcome
02:04 Twenty emails instead of ten
06:50 The down escalator
10:59 The infinite workday
15:30 The shrinking present
19:40 Who pays for the time AI saves?

Elisabeth talks with Krzysztof, a sociologist of time in Warsaw, and Francesca, who leads the AI team at a tech company in Amsterdam. Why did Keynes' fifteen-hour week never arrive? What does a civil servant do with 26 saved minutes a day? Is the AI race a down escalator or a learning curve? Should AI agents have a right to disconnect too? And why do the young pay most when the present shrinks? The guests try answers to two of the essays' open questions – and leave others open.

Season 1 – Time & AI. The essays behind this season and its conversations start from the work of Hartmut Rosa, professor of sociology at the University of Jena and director of the Max Weber Centre in Erfurt. His theory of social acceleration describes how technology, social change and the pace of life speed each other up; his later books on resonance and on the uncontrollability of the world ask what a good relationship to the world looks like. Each episode tests one part of his work against current research on generative AI.

Essays in this episode (Time & AI → Acceleration & time pressure)

• Twenty emails instead of ten
• The down escalator
• The infinite workday
• The shrinking present
• Who pays for the time AI saves?

Mentioned in this episode

• Hartmut Rosa in ARTE Offene Ideen (2023): Was verlieren wir, wenn wir Zeit sparen?
• John Maynard Keynes (1930): Economic Possibilities for our Grandchildren – the fifteen-hour week
• UK Government Digital Service (2025): Microsoft 365 Copilot trial with 20,000 civil servants, 26 minutes a day (self-reported)
• Barron's on the UK trial: 22 % saved less than five minutes
• Ruth Schwartz Cowan (1983): More Work for Mother – via Joel Mokyr (2000)
• MIT NANDA (2025): The GenAI Divide – 95 % of organisations without measurable return
• William Barnett (2008): The Red Queen among Organizations
• OECD annual hours worked 2023: Poland about 1,800, Germany about 1,340
• Microsoft Work Trend Index (June 2025): Breaking down the infinite workday
• Jonathan Gershuny (2005): Busyness as the badge of honour for the new superordinate working class
• BetterUp Labs & Stanford Social Media Lab, Harvard Business Review (2025): AI-generated workslop
• Volkswagen: no emails to company phones outside working hours (since 2011)
• Right to disconnect: France (2017) and Australia (2024)
• World Economic Forum (2025): Future of Jobs Report – 39 % of core skills expected to change by 2030
• Brynjolfsson, Chandar & Chen: Canaries in the Coal Mine? (Stanford Digital Economy Lab, update August 2026)
• Humlum & Vestergaard (2025): Large Language Models, Small Labor Market Effects – 3–7 % of gains passed to workers
• TIME (2023): Kenyan data labellers behind ChatGPT
• Autonomy (2023): UK four-day week pilot – 56 of 61 companies continued

Speakers: Elisabeth (Host), Krzysztof (sociologist of time), Francesca (AI lead). AI-generated voices, fictional speakers; script researched and written with AI.

Transcript and all sources