Lauren Crichton on AI's Contradictions and Choosing a Future

ZUSAMMENFASSUNG

In 30 Sekunden

Sana vice president Lauren Crichton opens the 2026 Sana AI Summit by comparing current questions about intelligent machines with Arthur C. Clarke's 1964 predictions. She describes rapid agent progress alongside conflicting signals about jobs, token costs, infrastructure investment and politics. Calling the moment ‘artificial general absurdity,’ Crichton argues that people can respond with humor and action. She explains why the summit brings together disciplines beyond technology to consider the futures they want to build.

1An old set of questionsCrichton recalls Arthur C. Clarke's 1964 program and asks how people should respond now that machine intelligence is no longer a distant prediction.
2Agents and software workShe says agents began completing longer tasks with tools and cites claims about coding productivity and AI-generated code.
3The economics of AIShe contrasts rising token usage and capital spending with falling inference costs and uncertainty about future supply and pricing.
4Politics and absurdityCrichton points to disagreements among AI companies and unexpected shifts by other businesses as examples of a contradictory moment.
5A wider conversationShe says Sana's summit brings together technology, science, economics, philosophy and the arts to help imagine and build better futures.

Das Wichtigste

Crichton sees today's questions about intelligent machines as echoes of those raised in the 1960s, but with more immediate consequences.
She describes agents as crossing a threshold into longer, tool-using tasks, especially in software engineering.
Her account places productivity claims alongside fears of job loss and organizational disruption.
She sees a tension between rising token use and investment on one hand and falling inference costs on the other.
Crichton calls for people across disciplines to help shape AI's future.

ThemenAI Adoption Future of Work AI cost & performance Geopolitics

VOLLTEXT

Den bearbeiteten Text nach Kapiteln lesen

Transkript zeigen

Nur das Transkript

From prediction to response

Crichton recalls Arthur C. Clarke's questions about computers. She argues that similar questions now concern a future being built in the present.

Lauren Crichton opens the 2026 Sana AI Summit by recalling Arthur C. Clarke's 1964 BBC program The Knowledge Explosion, filmed at the New York World's Fair. Amid the space race, nuclear tension and early computing, visitors were trying to understand what computers would mean for everyday life.

Crichton says the questions remain familiar: what will people do as computers become smarter, what will happen to the economy, and could people create their own destroyers? The difference, she argues, is that these are no longer only predictions about a distant future but choices about one being built now.

Agents, jobs and the economics of AI

Crichton describes a perceived leap in agent capability and conflicting signals about productivity and employment. She then discusses token demand, inference costs and investment.

She describes late 2025 as a point when agents crossed a threshold of coherence: with systems such as Claude Code and Opus 4.5, they appeared able to use tools and complete tasks lasting hours. She identifies software engineering as the first area to feel a major effect.

Crichton cites claims that AI writes a large share of new code at Google and that some startups report sharply higher productivity with coding tools. She also describes ‘tokenmaxxing’ as a status game, with employees comparing how many tokens they use. These figures are the claims presented in her talk.

Alongside that enthusiasm, she points to fears that junior technical roles may be replaced, layoffs at technology companies and reported employee resistance to AI rollouts. She calls the anxiety FOBO: fear of becoming obsolete. Her examples present a contrast between productivity gains and insecurity about work.

The economics appear contradictory to Crichton as well. She cites a rise in OpenAI API token processing from six billion per minute in October to 15 billion in April, alongside a claimed rapid fall in per-token inference costs. She says future capacity and pricing remain uncertain.

She contrasts the large sums frontier labs are spending on computing infrastructure with the possibility that model development is becoming cheaper, including through work by Chinese labs. She presents competing interpretations: computing as a scarce resource, or an apparent supply monopoly presented as a demand-driven revolution.

Absurdity and collective action

Crichton turns to AI politics and unexpected business pivots. She closes by calling for humor, action and participation across disciplines.

Crichton turns to politics, contrasting Anthropic's and OpenAI's approaches to the US government and noting disputes among AI founders. She asks whether public debate has lost sight of reality amid conflicting claims and corporate rivalries.

To illustrate what she calls absurdity, she cites a toilet manufacturer increasing production of semiconductor components and Allbirds moving from shoes toward AI infrastructure. She jokes about replacing the shoes with AI-powered footwear and labels the moment ‘artificial general absurdity.’

Faced with an absurd present, Crichton says people can laugh, despair or act; she favors laughter and action. She argues that people can bring desired futures into existence and asks what kind of future they actually want.

Crichton says Sana's first AI Summit in Stockholm in 2023 began from the belief that the technology industry should not shape the future alone. The event brings together robotics, economics, machine learning, physics, biology, philosophy and literature in the hope that different disciplines can challenge and inspire one another.