Written by

, ,

Vibe Coding Workshop at the University of Warwick

On May 14, 2026, the Convivial.Code Workshop at the University of Warwick was hosted by the Centre for Digital Inquiry and Centre for Interdisciplinary Methodologies, where David M. Berry offered a keynote with a sharp and inspiring perspective on how programming is changing in the age of vector computation and vibe coding. The event was organised by Michael Dieter and featured Michael Castell and me as respondents. The discussion was mediated by the dear Carolina Bandnelli. Unfortunately, due to illness, Bernhard Rieder couldn’t join the discussion, and we were all looking forward to hearing from him.

My response was organised into three parts:

1. ๐•๐ข๐›๐ž ๐‚๐จ๐๐ข๐ง๐  ๐‘๐ž๐š๐œ๐ญ๐ข๐ฏ๐š๐ญ๐ž๐ฌ ๐š๐ง ๐Ž๐ฅ๐ ๐ƒ๐ข๐ ๐ข๐ญ๐š๐ฅ ๐Œ๐ž๐ญ๐ก๐จ๐๐ฌ ๐‚๐ก๐š๐ฅ๐ฅ๐ž๐ง๐ ๐ž
The ease of tools can make us forget what we don’t understand. The methodological question isn’t how to remove friction entirely, but how to navigate it with discernmentโ€”recognising when something is broken, partial, or misleading, and when to intervene.

2. ๐‡๐จ๐ฐ ๐ฐ๐ž ๐›๐š๐ฅ๐š๐ง๐œ๐ž ๐š๐ง๐ ๐š๐œ๐œ๐จ๐ฎ๐ง๐ญ ๐Ÿ๐จ๐ซ ๐ž๐ง๐ ๐š๐ ๐ž๐ฆ๐ž๐ง๐ญ๐ฌ ๐ฐ๐ข๐ญ๐ก ๐€๐ˆ
Rather than identifying which AI mode you occupy (delegation, augmentation, overhead), the skill lies in moving between them. I propose thinking of AI as:
A Thinking Catalyst (challenging & expanding thinking)
A Learning Collaborator (planning & structuring)
A Precision Companion (instrumental tasks)
This keeps you as the author while remaining discerning about engagement.

3. ๐„๐ง๐ ๐š๐ ๐ข๐ง๐  ๐ฐ๐ข๐ญ๐ก ๐€๐ˆโ€™๐ฌ โ€œ๐ฆ๐ž๐œ๐ก๐š๐ง๐ข๐œ๐ฌ ๐จ๐Ÿ ๐ซ๐ž๐š๐ฌ๐ฌ๐ž๐ฆ๐›๐ฅ๐ฒโ€ ( or those moments when AI appear to recover, reconstruct, or reassemble concepts, patterns, and individual/collective intellectual trajectories from ๐ฐ๐ž๐› ๐ญ๐ซ๐š๐œ๐ž๐ฌ)
Agentic AI can reassemble traces of your own (or others) intellectual work in ways that interrupt immediate tasks and provoke unexpected insights. This requires a different posture: temporarily surrendering to what the machine reassembles, rather than immediately taking over, correct, measure or instrumentalising it. It speaks about surrendering to the process.

Watch Michael Dieter’s welcome note, David Berry’s keynote, and responses.

Video summary by YT AI

This workshop, titled “Vibe Coding Workshop,” critically examines the emergence of AI-assisted programming (referred to as “vibe coding”) and its implications for computational culture. Hosted by the Centre for Digital Inquiry and the Centre for Interdisciplinary Methodologies at the University of Warwick, the event explores how these new tools reshape software development, labor, and technical knowledge.

Key Themes and Discussions:

The Rise of Vibe Coding (0:22ย –ย 3:02):ย Coined byย Andre Karpathy, “vibe coding” refers to writing software by interfacing with AI through natural language, prioritizing aesthetic or atmospheric iteration over traditional software engineering.

Vector Coding & Media Paradigms (9:32ย –ย 19:04):ย Keynote speakerย David Berryย introduces the concept of the “vector medium,” theorizing a shift from symbolic to statistical computation where media is “dissolved” into high-dimensional vector spaces. This framework helps explain the opacity and probabilistic nature of current LLMs.

The Politics of AI Tools (57:40ย –ย 1:01:28):ย The discussion highlights the “harnessing” of LLMs (such asย Claude Code) and the risk of “productarianization,” where developers become alienated managers of opaque, AI-generated code.

Methodological Challenges (1:15:35ย –ย 1:23:50):ย Janna Joceli Omenaย discusses how these technologies challenge existing digital methods, noting that while they can assist in research (re-assembling intellectual trajectories), they also risk masking technical conditions, requiring researchers to maintain a critical, “guarded” engagement.

Black Box & Accountability (1:57:35ย –ย 2:04:08):ย The participants debate the ongoing relevance of the “black box” concept in understanding AI, concluding that it remains a vital heuristic for addressing political economy and the lack of explainability in machine learning systems.

Workshop Participants:
David Berry:ย Professor of Digital Humanities,ย University of Sussex.
Janna Joceli Omena:ย Lecturer in Digital Methods,ย King’s College London.
Michael Castelle:ย Associate Professor,ย University of Warwick.
Moderators:ย Michael Dieterย andย Carolina Bandinelli

Leave a Reply

Discover more from Janna Joceli Omena

Subscribe now to keep reading and get access to the full archive.

Continue reading