Last Tuesday, a patient in clinic stared at the small brass wheel on my desk, the one I keep because it still turns with a satisfying resistance, and asked why anyone still cares about old mechanisms when everything now lives inside a phone. She was not really asking about the wheel. She was asking about usefulness, about memory, about whether anything made by hand still matters in a world that seems to reward frictionless software and disposable novelty.
I told her I had been thinking about that same question while watching a reel-to-reel deck I am building on my workbench at home. One spool was missing a guide, the tape path was temperamental, and I had spent the better part of an evening tracing a fault that a cheaper, faster gadget would have hidden from me. The machine was irritating. It was also honest.
That honesty is what keeps pulling me back to kinetic art, cassette futurism, and the strange new comfort of AI for people who make things with their hands. The obvious reading is that AI belongs to the clean, efficient future while analog machines belong to nostalgia. The clinic teaches otherwise. The most interesting thing happening right now is not that AI is replacing makers. It is that AI is making more people into makers, including people who were never trained to think of themselves that way.
The old gatekeeper was not talent
I used to think the barrier to making serious things was mostly talent, or formal training, or access to expensive tools. That was the story many of us told ourselves in medicine too. Then I watched students, clinicians, and hobbyists use generative tools to draft enclosures, sketch circuitry, generate parts lists, or iterate on visual concepts they could never have described cleanly in the first place. My view changed because the bottleneck was not imagination. It was translation.
AI lowers the cost of translation. A person can now say, in plain language, “I want a brushed aluminum faceplate with a 1970s broadcast-console feel, but with a softer glow and a quieter motor,” and get a usable starting point in minutes. That matters because many potential makers are not blocked by lack of taste. They are blocked by jargon, by CAD fluency, by confidence, by the shame of not knowing the right terms.
That is the real liberalization here. AI is not democratizing genius. It is democratizing iteration.
The maker who does not yet know she is a maker
A few months ago, a colleague brought me a rough concept for a wall piece that combined rotating mirror elements with light and sound. She had the taste already. What she lacked was a fluent way to move from image to object. With AI, she could test form, proportion, palette, and motion logic before spending money on material. The first version was clumsy. The third was promising. By the fifth, she had something she could actually build.
I have seen this pattern enough times now that I think of it as the maker’s loop. The loop is simple: curiosity, prompt, prototype, correction, confidence. It sounds trivial. It is not. For many people, the first successful loop is the moment they stop saying, “I am not technical,” and start saying, “I can figure this out.”
That psychological turn matters more than most of the tech press wants to admit. In a 2026 paper, Naito and Shirado reported that faith in AI can narrow the futures individuals consider, which is a useful warning because the same tool that opens one door can also close others if it becomes a crutch. The lesson for makers is not to worship the tool. It is to use it to widen, not flatten, the range of possible outcomes. Source: Faith in AI can narrow the futures individuals consider.
Why cassette futurism keeps coming back
Cassette futurism is funny because it looks like a failed prediction and a viable aesthetic at the same time. It is the idea that the future could have been chunky, tactile, warm, and visibly mechanical. I like it because it preserves friction. You can hear the gears. You can see the cassette window. You know where the bodies are buried, metaphorically and mechanically.
AI fits this aesthetic better than the polished abstractions of the last decade. It makes design accessible without making design invisible. A person can use a model to plan a front panel, generate a user interface mockup, or sketch a control scheme, but the finished object still benefits from the grain of the material. The model is a collaborator, not a replacement for the hands that finish the work.
That is why I am drawn to kinetic art in this moment. Kinetic pieces refuse the flatness of the screen. They insist on motion, weight, timing, and error. They remind me of medicine in that way. Real work is always happening in a three-dimensional environment with constraints.
What the studies actually suggest
The research on human-AI co-creation is starting to match what I have been seeing in practice. In Generative artificial intelligence models outperform students on divergent and convergent thinking assessments, researchers reported that GenAI models outperformed students on both divergent and convergent thinking tasks. Another study, “From humans to AI: understanding why AI is perceived as the preferred co-creation partner,” found across four empirical studies that co-creating with AI increased perceived novelty and usefulness, which in turn increased co-creation intention. Those are not small effects in the real world of people trying to begin.
There is also a broader creative picture. A 2025 systematic review of generative AI and creativity summarized a pattern I recognize from lived experience: AI can help with early ideation, but it can also compress variation if everyone leans on it the same way. The tool opens the studio door. It can also nudge everyone toward the same furniture if they do not keep their own hands on the work. That warning should not scare us off. It should keep us honest.
I would not use AI to make the final aesthetic judgment for a piece of art, a clinical workflow, or a maker project. I would not outsource taste, because taste is where values show up. I would not ask a model to decide whether a work should feel eerie, intimate, or nostalgic, because those are human choices, not output optimizations. I will happily let it accelerate drafts, generate options, and reduce the blank-page tax. I will not let it own the intention.
Clinical life taught me to respect tools that fail honestly
One reason this topic lands so hard for me is that medicine trains you to distrust overconfident instruments. A machine can be useful and wrong at the same time. That lesson is not abstract. I have seen clinicians over-rely on systems that were helpful in aggregate and misleading at the bedside. I have also seen the opposite, where a skeptical human caught what the software missed because the patient in front of them had a story that did not fit the template.
That is the same balance I want in maker culture. AI should lower the cost of trying. It should not lower the cost of thinking. The person who uses a model to generate a thousand terrible variants and then chooses carefully is doing something serious. The person who accepts the first polished answer is not making. They are renting plausibility.
In that sense, AI resembles the reel-to-reel deck I am building. The machine helps, but it also demands attention. A loose belt reveals itself. A noisy transport stage does not politely disappear. The artifact tells the truth about the system that produced it.
What changes when access widens
When access widens, the culture changes before the artifacts do. More people enter the conversation. More ideas become testable. More failures become survivable because the first draft is cheap. That is why I think AI is making makers in a deeper sense than the usual productivity rhetoric suggests. It is creating a larger population of people who can move from taste to prototype without waiting for permission.
That has implications for art, for engineering, and for medicine. A clinician who prototypes a patient-education tool. A resident who builds a simulation around a procedure they keep seeing done badly. A hobbyist who designs an analog instrument because they could not find one that feels right. These are small acts. They accumulate into professional culture.
I also think this matters because the culture of making is how professions keep themselves honest. Medicine used to have more room for craft talk. We talked about hands, judgment, timing, feel. AI can either erase that language or revive it by giving more people entry into making, including people who are not engineers, programmers, or full-time artists.
For readers who want the biographical version of where I write from, my credential page is here: Dr. Sina Bari, MD, Stanford-trained surgeon. I mention it because perspective matters here. I am not describing this from the outside.
The danger is sameness, not capability
The strongest counterargument to this optimistic view is also the most serious one. If too many people use the same tools in the same way, the result can be homogenization. The 2025 and 2026 creativity literature keeps circling that concern, and I think it is correct to do so. Convenience can erase local style. Defaults can become dogma. The model’s taste can become the room’s taste if nobody notices the handoff.
That is why the future I am interested in is not AI art in general. It is AI-assisted singularity, the stubbornly personal kind. A cassette deck with a faceplate no one else would design. A kinetic sculpture that reflects one person’s childhood memory of a hospital waiting room. A reel-to-reel built because somebody wanted the experience of threading tape by hand and hearing the transport settle into motion.
The machine helps you begin. The human decides what is worth preserving.
Back at the workbench
Three nights after that clinic conversation, I returned to the reel-to-reel project and finally got the tape path aligned. Nothing dramatic happened. The machine just behaved the way it was supposed to, with a soft mechanical certainty that felt almost ceremonial. I thought again about the patient who had asked whether anything made by hand still matters.
Yes, it matters. It matters because making still teaches judgment. It matters because AI has made the first pass easier without making the work less human. It matters because more people now have access to the rough draft of competence, and that changes who gets to enter the studio, the garage, the lab, and maybe even the clinic with confidence.
That is what the maker’s loop gives us: not perfection, but entry. Not final answers, but a faster path to the moment when a person says, I can build this. I can repair this. I can imagine more than one version of the future.
The reel-to-reel still sits on my bench. It is not finished. That feels right.
FAQ
How is AI actually helping artists and makers start projects faster?
AI helps by translating rough intent into usable drafts, whether that means layout ideas, visual references, materials planning, or code snippets. The practical gain is not instant mastery, it is getting from blank page to first prototype without waiting for perfect technical fluency.
What is the main risk when makers rely on AI too much?
The main risk is sameness. If people lean on the same prompts, defaults, and model outputs, work can converge toward a narrow visual language, which is exactly the kind of flattening the creativity literature warns about.
What is Dr. Sina Bari's view of AI in creative work?
Dr. Sina Bari’s view is that AI should widen access to making while leaving judgment, taste, and intention in human hands. He sees it as a powerful drafting partner, especially for people who have ideas but lack fluency in technical tools.
Can AI help someone build something analog, like a reel-to-reel or kinetic sculpture?
Yes. AI can help with enclosure concepts, parts lists, visual mockups, motion logic, and troubleshooting ideas, even when the final object is fully analog. The build still depends on the maker’s eye and the physical constraints of the materials.
Why do cassette futurism and kinetic art matter in the age of AI?
They matter because they keep friction, motion, and materiality in view. In a time when so much can be generated instantly on a screen, analog-looking objects remind us that useful work still has weight, texture, and failure modes that need human judgment.