AI Is a Bore

Oh No, We Are

A post for anyone who has felt their soul evaporate while scrolling yet another “10 Prompt Hacks to 10× Your Productivity” thread.

The Day the Future Started Repeating Itself

Open any professional feed. Within ten seconds you will encounter a claim, a panic or a promise you’ve seen before:

• “AI won’t replace you, but someone using AI will!”

• “We urgently need regulations.”

• “Look at this prompt—absolute game‑changer!”

Refresh. The same carousel re‑loads, rearranged like fridge‑magnet poetry. We live inside an infinite GIF: the technology is frame‑advancing at blistering speed, but the accompanying conversation is stuck on loop.

Meanwhile, the models sprint from GPT‑3.5 to GPT‑o3 to whatever‑comes‑after‑summer‑holidays. They learn entire corpora between coffee breaks. We, in contrast, rehearse the same talking points with the doggedness of amateur actors who fear opening night.

If you sense cognitive dissonance, an accelerating horizon married to a stagnant discourse, you are not alone. The dissonance is the story.

Why Our Discourse Fossilised While the Code Compiled

The content economy rewards déjà vu.
Algorithmic platforms prize predictability: familiar hooks, known dopamine triggers, repeatable calls to action. Nuance dies where click‑through thrives. Thought leaders soon discover that “prompt of the week” posts convert better than awkward meditations on epistemology, so they comply.

We outsource curiosity to the machine.
Large language models can now summarise a 400‑page white paper before the kettle boils. Wonderful, but the shadow side is a creeping atrophy of first‑hand inquiry. If the app can regurgitate, why sweat the synthesis? The result: recycled insights tarted up as new because nobody remembers the primary sources.

Comfort in crisis mode.
Apocalyptic or utopian extremes give us emotional oxygen. They spare us the harder labour of navigating a grey zone where outcomes hinge on governance, design and most lethally our own willingness to adapt. So we stay on the moral high wire, arguing abstractions.

The Economics of Boring

Boredom is not neutral; it is profitable. Boredom scales. It packages easily into listicles, webinars and glossy PDFs.

True experimentation, by contrast, incurs opportunity cost, reputational risk and the possibility of failure in public. You can spend a month building a genuinely novel AI‑augmented workflow for your library or you can dash off “Five Ways Librarians Can Harness ChatGPT” before lunch and still make the afternoon management meeting.

Predictable noise crowds out signal and the platform metrics applaud the noisemakers. We shouldn’t be surprised that many choose the applause.

Lost in the Possibility Space

While we litigate last year’s anxieties, whole categories of collaboration blossom unnoticed:

• Micro‑expert markets freelancers who rent niche expertise to fine‑tune agents for local contexts: a Basque‑speaking academic, a rare‑disease community manager, the city archivist in your regional town.

• Ambient decision engines lightweight models embedded in everyday objects, nudging our behaviour in subtle ways.

• Synthetic mentors personae fine‑tuned on living thinkers who consent to licence their intellectual style, creating a renewable apprenticeship market that breaks the tyranny of geography.

These examples are neither speculative nor distant. They are quietly shipping, ignored by feeds obsessed with the sixty‑second hot‑take cycle.

A Brief History of Forgetting

Remember GPT‑4’s vision feature? The collective gasp, the avalanche of demos? Two weeks later the noise faded, replaced by the next shiny thing. Our attention behaves like venture capital on amphetamines: deploy, hype, abandon. The memory resets; the models do not.

Consequently, our institutional learning is shallow. We rediscover “Wow, large context windows!” every single release, refusing to build layered expertise. The door opens; we admire the doorknob; we close it and walk away.

The Bottleneck With a Pulse

Here is the unflattering hypothesis: the limiting reagent in the AI revolution is not hardware, not training data, not even regulation - it is the human capacity to transcend performative discourse.

We face a challenge of cognitive operations:

1. Asking weirder questions. Most prompts are linear requests for extraction or speed. Few probe the model’s ability to suggest alternative ontologies, to remix disciplines, to interrogate the user’s assumptions.

2. Designing for negative capability. Keats defined this as the ability to remain in doubt without irritably reaching after fact or reason. Our interfaces could foster exploratory ambiguity rather than immediate answers, but only if we demand it.

3. Institutionalising forgetting, deliberately. Imagine R&D sprints where the rule is no inherited prompt libraries. Teams must rederive methods, surfacing tacit knowledge and revealing stale assumptions baked into old macros.

Five Provocations (Not a Framework, Promise)

• Build something uselessly beautiful. Artists understood centuries ago that frivolity is R&D for the soul. Train a model solely to generate fictitious taxonomies of clouds. Watch your strategic imagination enlarge.

• Hold a Bias Funeral. Schedule a meeting whose sole purpose is to retire one well‑worn bias anecdote (“the Amazon résumé filter!”) and replace it with a fresher, more context‑specific example drawn from your own data.

• Run a One‑Day Hermitage. Disconnect all feeds, lock yourself with the latest model and pursue a single problem to exhaustion. Observe how quickly the absence of social discourse alters the shape of your inquiry.

• License your thinking. Offer a limited‑scope dataset of your notes, essays, or podcast transcripts for yourself or others to fine‑tune. Monitor what emergent insights come back. Think of it as intellectual composting.

• Audit for questions asked. Instead of tracking KPIs on answers delivered, measure the novelty and range of questions your team poses to the system each quarter. Reward breadth, not just throughput.

Boredom Is a Governance Choice

Electricity did not change the world because people wrote thought‑pieces about electricity. It changed the world because individuals wired strange new devices into dark corners and refused to be bored by their potential.

We sit at an analogous threshold. If our discourse feels tedious, it is because we have chosen safety, familiarity and algorithmic applause over exploration. The models are not the culprits; they mirror the conservatism of their operators.

So let us retire the Groundhog Day carousel. Brake the dopamine scroll. Open the possibility space like a field guide and wander beyond the well‑trodden trails.

When future historians examine the Anthropocene code‑base, may they note that in the mid‑2020s humankind finally lost patience with its own clichés and got wonderfully, dangerously interesting again.

See you off the feed.