Oblique Strategies is a set of cards created by musician Brian Eno and artist Peter Schmidt in 1975, designed as a creative tool to inspire and guide individuals through moments of artistic or decision-making difficulty. The cards serve as a creative catalyst, encouraging users to interpret the vague, poetic instructions in a way that sparks inspiration or re-frames their viewpoint. Since it relies on ambiguity, the strategies avoid prescriptive advice, instead fostering intuition and open-ended interpretation.
What makes this cue card technology agnostic is that it relies on association, information, and interpretation. The mind is a permeable membrane soaking up informational elements in a time and place as a passenger, or by conscious directed effort to explore conceptual abstractions or information. Abstraction and ambiguity can be a weapon against a sea of sameness. Complexity and information can feed ambiguity, it strengthens it. Uniformity is decline in an AI future, divergence is the way forward.
You navigate your thoughts to achieve an ordered organisation, guiding your intent towards actionable steps. Part of that step involves deriving observations or interpretations from the information. With artificial intelligence, every thought is an exploding invisible sun. Thoughts move as with Brownian motion—reactive to events, to what we perceive in the external world or in our inner space. In contemplation, the chatter is good and begs to be put on a blank canvas—a trusty notebook/word doc. In performative action, the quiet mind is the Zen mode required to focus on the task for control.
These cue cards apply strategies through creative prompts to address decision-making challenges, learning to interpret abstract prompts in a way that inspires actionable outcomes. The deck is often used by artists, musicians, writers, and anyone seeking to break through creative blocks or approach challenges in unconventional ways. Each card features an abstract phrase or instruction, such as “Honor thy error as a hidden intention” or “Use an old idea,” intended to provoke associative thinking and offer new perspectives on a problem. While it is stated that creatives derive the most value from the cards, you should not assume that if you were in business that this would not be a viable approach.
You might begin with an abstract thought, scribble your notes, and identify a few points that seem more conceptual. You then conduct your market analyst validation, and if the market appears to be under the radar, you may not bother. You get back to your notes to find the answer here or change course to something else entirely. The deck has still worked since it got you to take action. You might start with a place of ambiguity, but you will, in the end, refer to the market analyst for decisions. The deck primes you to overcome creative blocks and unwillingness to experiment, to take risks, and view constraints as opportunities for innovation, all this without AI systems in the beginning. You have control over prompting yourself through these cards before you prompt the AI systems. The chain of thought is expanded when you prime yourself before usage. Frankly speaking, any piece of information, no matter how nebulous, if you pass through the AI system, I believe you will be able to ground it to a real-world application through a reasonable query.
Application in Vibecoding for experimental software development
What is vibecoding?
Vibe coding, a term coined by Andrej Karpathy, involves interacting with a codebase through prompts or conversations with AI to develop programs. Using Oblique Strategies alongside vibe coding works best when you come with no expectations. These strategies offer a system for incorporating randomness and a non-linear approach to the process. It’s often best to ignore business validation at the outset—entering vibe coding with specific criteria can limit the creative trajectory of your prompts. You might shut down ideas that stray from initial notions of product fit, even though those detours could lead somewhere unexpected and valuable.
You can say this is like the writer who types everything down in bullet points in the first draft, with all the mess of typos. The second viewing is to edit, neaten, expand, and refine. The convention of having a document laid out is not bad practice in any regard. I do think that sometimes to arrive at a line of thought that would surprise you can only happen in a kind of serendipity. This is not something you can plan for.
I am writing this as a way to illustrate how you use a system-based approach to utilize spontaneous, improvisational, and chance in a directed decision for vibecoding, which Oblique Strategies does answer for.
Strategies in the practice of Oblique Strategies for Vibecoding
You can use Oblique strategies in vibecoding in a group setting/collaborative, or personal. Propose team vibe coding sessions where each member draws a card and pitches a feature or bug fix. The team votes or combines ideas to foster collaboration
Process:
- Each team member draws an Oblique Strategy card
- Present the current technical challenge
- Each person interprets their card in the context of the problem
- Use AI to rapidly prototype each interpretation
- Vote on the most promising directions
Be aware that production deadlines can suffer due to excessive experimentation
Codebases: Stable vs. Experimental
Your peers have collaboratively vibe code with Oblique strategies. It’s been agreed that the final software is going to be what will be worked on further. You could make two alternate versions through divergent coding, use the second one as the experimental code base to add, and change things. You have a stable “production” version and an experimental “vibe” version. Generate multiple code versions inspired by different Oblique Strategies prompts, then converge on the best solution during refinement. You could do something entirely different from the codebase you settled on as stable and break the branch to do something else entirely. Change course. The road not taken?
Mutagenic Software: Build Big, Then Prune Down for Minimalism
Vibe coding with AI prompts can generate radical variations in an experimental codebase, safely tested without risking the stable version. You’ll use this as a template—either something you’ve found or a branch from the stable code of a previous session. This is “mutagenic software,” as code evolves unpredictably, bloated with experimental features. But unlike biological mutations, you can rest easy—software won’t spiral into ecological chaos or trigger outbreaks. You know how that goes. Go wild with the prompts. Vibe coding with AI can escalate complexity fast, especially when guided by Oblique Strategies-style prompts like “Add more.”
A two-phase process:
(1) Build a “useless” prototype with maximal features
(2) less is more, keeping only what serves the core vision. Every prompt is a layering on the commands for code generation.
After vibe-coding a prototype, use a card like “Take away the elements in order of apparent non-importance” to remove code, revealing the essential systematically.
E.g. A developer builds a feature-heavy app with AI-generated bloat (e.g., redundant APIs), then uses “Simplify” to strip it to a lean MVP.
Use a Scaffold
Starting with a source code, you can use as a scaffold or a blank canvas. Repurpose code, you have some fragment or archived code somewhere, maybe it is written in a different programming language. You rewrite in a new language, then you tell it
Make a sudden, unpredictable change
Observe what happens. Tell it to introduce some bugs. Let’s see your next cue card
Honor thy error as hidden intention
See if you can make something out of those bugs and turn it into features. Next card
riff
Maybe you riff off over a conversation with AI or your peers, and then sporadically take action on prompting. I want to call your attention that for this example, I interpret riff as a call and response. Some people might see the word “riff” and associate that with a guitar riff, in which case I see you. You can choose to interpret this differently in any direction you decide on.
Journal-Driven Development (slow incubation/big step)
A reflective vibe coding practice. Log every AI prompt, Oblique Strategies card, and coding decision in a journal (digital or physical). This slows development but deepens insight, like a journal. Walk, don’t run. Journals become a creative artefact, documenting the “abstractions” of software development through an array of research materials like market snapshots, browsing unrelated articles, images, or code repositories. Now you use the cue card, just one card, state various actions or approaches you can consider, think it through and note it down before you prompt forward. Yes, you can also use each cue card to prompt yourself to respond by action as we saw in the scaffold section.
But consider if you use one cue card and keep the obsession of drawing out what that ambiguity means. It forces you to interpret ambiguous information through your unique context and expertise, one cue card, is a world within a world expanded by your thoughts—a world in a grain of sand. You can use the AI systems to draw out your thoughts into a long chain of thought bubbles. You reflect on a single cue card prompt, and you obsess over that line “Use an old idea” on all the different approaches you can consider to respond led to reviving a deprecated library, noting emotional highs and lows of the process. How did this card change my code? What felt intuitive vs. forced? Document every strategy, prompt, and creative decision.
Software as Art: The Process
Experimental improvisation at its finest. Software development undergoes a process like an art piece that can explore any number of given directions before it’s considered the final software. Vibe coding, guided by Oblique Strategies, treats code as a living artwork, not just a functional product. Each version is a brushstroke — a line of code, shaped by intuition and chance. Maybe you need a system on how to use chance. Can’t decide between two options, flip a coin. Any thought in your head you pass through the AI system just to see what it produces. You are painting software. It might look like a Cubist art piece. Settle down. You are still not Picasso.
Environment
Urban planning in proximity to economic output
If people were products of cities, then they might reflect the characteristics of those cities. Imagine a cluster of companies overlapping in a specific area—an ecosystem where distinct ideas coexist. Industries like gaming, commerce, art/creative, cybersecurity, and technology intersect across economic sectors to paint a vivid scene. Some miles away, a cutting-edge research facility quietly observes the cluster, taking note of its autonomy. Every so often, a layer of their research—a piece of information—is shared. The denizens of this city, enthusiastic, curious, and occasionally suspicious, respond to this little wonder of insight through direct application, inquiry, or reinterpretation. How does the introduction of such input shift the direction of these companies and products—and what does it reveal about market potential?
For autonomy of information representation and generation, the system must maintain continuous independence proliferating from its output once it attains a state in which the components of growth — individuals and companies have anchored a stable establishment. Cities evolve as ideas do, through collaboration, constraints, and dynamic influences. You can implement policies around how cities are developed. You can’t account for the characteristics of culture and values as it paints a place through the agency of groups and individuals just by planning alone. You cannot simulate a culture, you can only produce a representation of the culture for which it must take on properties of its own to be distinct, in order to be distinct the agency of groups and individuals maintain control for the expression of its values.
You can create conditions that make authentic innovation more likely. San Francisco’s tech scene emerged from decades of counterculture, academic research, and economic opportunity converging in specific ways. The real cities in the world that convey the points above are San Francisco, Tokyo, Berlin, Los Angeles and Austin blending overlapping industries. These cities have an overlap of gaming, commerce, art scene and technology industries. Berlin has its cultural output, from museums to street art, fueling creativity. A startup hub with e-commerce players like Zalando and a strong retail scene in Mitte. Berlin’s creative-tech fusion grows from its artistic history and experimental culture. It has a rich modern distinct historical backdrop but at this time seems to be in a transitional stage. I hope some German official reading this somewhere in Germany will give some thought to considering options on how to revitalize the potential of Berlin for commerce, finance and culture.
The strategic planning of transportation systems around active commercial areas can be the points needed for urban planning by proximity. The ease of mobility and movement in industrious regions can play a factor in the maturity and advancement of commerce. Tokyo’s creative-tech scene emerges from a unique tension between traditional craftsmanship culture and digital experimentation. The same city that perfected centuries-old woodworking techniques also pioneered karaoke, Pokemon, and emoji. Austin deliberately cultivated its “weird” identity to attract diverse creative industries. The annual South by Southwest festival exemplifies this approach. Music, film, and interactive technology converge in the same physical and temporal space.
Some of these cities already have the components that were established as a backdrop of history, culture and development. A small change for optimizing the planning of urban development would be enough to nudge it in a particular direction for GDP output. This is the part where I point the reader to how Oblique strategies and vibecoding relate to urban planning? Vibe coding and Oblique Strategies mirror this: they’re tools for creative collision, stirring ideas in a city-like feedback of system emergence. The stimuli is a feedback for groups and individuals for collision that differ by the city’s characteristics.
You are now a process. The process is organic, structured, and chaotic, you determine the extent of the process making checkpoints along the way with every step forward. At any point, you can use ambiguity to take you further away like a reflecting infinity mirror as you document your capacity for thought in an ocean of everything. The process itself is, after all, part of the art. Before you crack the code like a routine morning wake, store up with comfort before controlled chaos with some coffee, bacon or eggs and the essential bread for carpe diem. Here’s one for you geeks, creatives, and ghouls out there to vibe code with Oblique Strategies. Eno in, run free — code-on
When the going gets weird, the weird turn professional — Hunter S. Thompson
References
https://github.com/joelparkerhenderson/oblique-strategies
