Hepp, A. (2026). The imaginative landscape of AI: Locating Silicon Valley’s “quiet futuring”. Media, Culture & Society. https://doi.org/10.1177/01634437261454515
Summary
Hepp’s article intervenes in critical AI studies by arguing that prevailing accounts of Silicon Valley’s sociotechnical imaginaries — focused on the inflammatory pronouncements of CEOs and investors like Altman, Musk, Thiel, and Andreessen — overlook an equally consequential layer of imaginative work. He calls this layer “quiet futuring”: the locally anchored, community-based visioning that takes place in Bay Area labs, start-up houses, Effective Altruist and Rationalist gatherings, and adjacent pioneer communities. Based on multisited ethnography in San Francisco between late 2024 and early 2026, the paper maps an “imaginative landscape” organized across three layers — appropriated places, social figurations, and discursive thickenings — and shows how quiet futuring supplies the ideational reservoir that loud futuring amplifies, all while presenting itself as apolitical despite being deeply political.
Key Contributions
- Introduces the conceptual pair loud futuring / quiet futuring to refine analyses of AI’s power geometries.
- Develops the imaginative landscape as a three-layered analytical framework integrating material places, social figurations, and discursive thickenings.
- Proposes an extended situational analysis as a methodological elaboration of Clarke et al.’s approach, scaled from situations to landscapes.
- Offers ethnographic grounding that complicates personality-driven critiques of Silicon Valley (e.g., focused on Musk, Thiel, Andreessen).
- Foregrounds the spatial and community-based grounding of imaginative agency in debates on AI power geometries.
Methods
Multisited media ethnography in the San Francisco Bay Area across an exploratory phase (Dec 2024) and a core phase of three field trips (June 2025–Feb 2026). The corpus comprises 53 interviews, 8 site observations (OpenAI/Anthropic HQs, AGI House, Frontier Tower, MOX, Constellation, Lighthaven, The Commons, Stanford HAI, Berkeley BAIR), and 7 event observations (e.g., EAGlobal, The Curve, StopAI protests), complemented by netnography of LessWrong, Substack, and Discord. Sampling moved from purposive/snowball to theoretical (Grounded Theory); analysis combined open coding with an extended situational analysis iteratively mapping places, figurations, and discourses.
Findings
- AI development has shifted geographically from lower Silicon Valley into urban San Francisco, anchored by OpenAI, Anthropic, and technologists’ lifestyle preferences.
- Specific places — labs, universities, start-up houses, and community hubs — materially anchor distinct ideological orientations within the landscape.
- Effective Altruism and Rationalism function as influential pioneer communities, supplying long-termist temporal arguments and AGI-as-existential-risk framings.
- At EAGlobal 2026, ~1200 attendees included 38 Anthropic and 16 OpenAI affiliates, evidencing persistent EA–lab ties despite public distancing.
- Three discursive thickenings dominate: (1) AGI arrival (with debate centering on when, not whether), (2) AI-driven job loss and UBI (e.g., Altman’s $14M to OpenResearch), and (3) AI geopolitics, especially the US–China race.
- Quiet-futuring actors consistently frame themselves as apolitical (illustrated by the Reboot “silly tech thing” episode), even as their visions are constitutive of AI’s political order.
- The landscape is fragmented and ideologically contradictory, encompassing cyberlibertarian, technofascist, EA/Rationalist, and counter-currents like “public AI” and StopAI.
Connections
This paper sits somewhat apart from the empirical, platform-and-elections-focused work clustered around it, but it speaks to broader questions about how infrastructural power and discourse co-produce platform governance regimes — concerns that surface in studies of platform infrastructures such as Helmond2026-ll and Rieder2026-pp. Its attention to AI imaginaries and their political consequences also resonates with work on AI in information ecosystems and electoral contexts, such as Mahl2026-hc. Beyond these, the connections to the topic clusters are thematic rather than substantive — most of the listed papers address empirical dynamics on social platforms rather than the imaginative work of AI elites.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify · Apple Podcasts