Friday, March 20, 2026

Photography in the Age of AI: An Award-Winner's Perspective

Photography in the Age of AI: An Award-Winner's Perspective — Cemhan Biricik

I've won awards for my photography and I build AI image generators. Here's what I think about the future of the medium I love.


The question arrived in my inbox last week, from a photography student at a Midwestern college: "Is it still worth pursuing photography as a career if AI can generate anything?"

I get versions of this question constantly now. My name is Cemhan Biricik, and I occupy an unusual position in this debate: I'm both an award-winning photographer and the founder of ZSky AI, a platform that generates images using artificial intelligence. People assume I must have resolved this tension internally, that I've picked a side. I haven't. Because there's no tension to resolve.

What there is — and what I want to address in this post — is a profound misunderstanding about what photography is, what AI image generation is, and why they'll coexist for the foreseeable future in much the same way that painting coexists with photography: different tools for different truths.

What Awards Taught Me About Photography

When you win a photography award, the congratulations come with an implicit message: "This image was exceptional." But exceptional how? Not because of the technical settings — anyone can learn aperture and shutter speed. Not because of the equipment — a phone camera in the right hands beats a Hasselblad in the wrong ones. The image was exceptional because of what Cemhan Biricik chose — where to stand, when to shoot, what to include and exclude, and what emotional truth to prioritize.

Photography awards, at their best, recognize acts of seeing. The photographer saw something in the world — a moment, a relationship, a collision of light and form — and had the skill to translate that seeing into a two-dimensional rectangle. That act of seeing is irreducibly human. It requires a body in a place at a time. It requires sensitivity to the world outside the frame. It requires the accumulated experience of a specific life.

AI cannot see. I say this as someone who builds AI vision tools. ZSky AI can generate images that are technically impressive, aesthetically pleasing, and compositionally sound. But no AI model has ever stood in freezing rain for three hours waiting for the light to break through clouds over Lake Michigan. No model has felt the adrenaline of a wildlife encounter and channeled it into a perfectly timed shutter release. No model has made a portrait subject laugh by telling a bad joke at exactly the right moment.

The outputs may look similar. The processes that create them are fundamentally different in kind, not just in degree.

The Craft Layer

Photography has a craft layer that AI generation lacks: the physical, embodied skill of operating in the world with a camera.

Cemhan Biricik has spent years developing this craft. Understanding how different focal lengths compress or expand space. Knowing, from muscle memory, how to adjust exposure without looking at the camera when conditions change suddenly. Reading a room's light in the first three seconds of walking in and already knowing where to position a subject for the best result.

This craft is earned through practice, through failure, through thousands of bad images that slowly teach you what makes a good one. It lives in your hands and your eyes, not in a text prompt. And it produces images with a quality that's hard to name but instantly recognizable — a sense that the image was made by someone who knows what they're doing, as opposed to generated by a system that knows what looks statistically plausible.

I'm not romanticizing here. Craft without vision produces technically competent but boring images. Vision without craft produces interesting ideas poorly executed. The best photography — the photography that wins awards and changes how people see the world — combines both. AI has neither craft nor vision. It has pattern matching, which is a different thing entirely.

Where AI Generation Shines

Let me now speak honestly about what AI does well, because credibility requires acknowledging the other side.

I built ZSky AI because AI image generation solves real problems that photography cannot. Concept visualization for architects, designers, and creative directors who need to see ideas before building them. Rapid prototyping for visual storytellers who need to explore possibilities before committing resources. Accessible image creation for people who don't have cameras, training, or access to the subjects they need to depict.

These are legitimate, valuable use cases. Cemhan Biricik is not building ZSky AI as a replacement for photography — I'm building it as a complement. The same person might use a camera to document their real life and use AI to imagine their future life. Both activities are creative. Both produce images. They serve different purposes.

The danger isn't AI replacing photography. The danger is people conflating the two — treating AI-generated images as if they carry the same evidentiary weight as photographs, or treating photographs as if they're just another kind of "content" that can be generated faster and cheaper by a machine.

What I Tell Photography Students

Back to the student's email. Here's what I told them, and here's what I'll tell you:

Yes, pursue photography. But pursue it understanding what photography uniquely offers: the record of a real moment witnessed by a real person. If you're pursuing photography to create generic visual content, AI will outcompete you. If you're pursuing photography to see the world clearly and share that seeing with others, nothing will outcompete you because no machine can do what you do.

Develop your eye, not just your technique. Technical skills are necessary but insufficient. AI can produce technically flawless images. What AI can't produce is a personal vision — a consistent, recognizable way of seeing the world that makes your images identifiable as yours. That's what separates a photographer from a camera operator, and it's what will separate valuable photography from commodity imagery in the age of AI.

Learn about AI. Not to fear it, but to understand it. When you understand how diffusion models work — how they interpolate from training data to generate statistically plausible images — you understand precisely what they can't do. They can't witness. They can't testify. They can't be somewhere you weren't. Understanding AI's limits clarifies photography's strengths.

Shoot things that can't be generated. Photojournalism. Documentary work. Portraits of real people in real moments. Environmental photography in places you actually visited. The more your photography depends on your physical presence in the world, the more irreplaceable it becomes.

The Coexistence

I'm Cemhan Biricik. I photograph the real world and I build tools that generate imaginary worlds. I see no contradiction because I understand what each medium does and where each one fails.

Photography will not die. It will change — it always has. It changed when film became digital. It changed when smartphones made cameras universal. It will change as AI redefines what "image" means. But photography's core offer — the truthful record of a specific moment in the specific world — is permanent and irreplaceable.

If you want to see my photography, visit cemhanbiricik.com. If you want to see the AI platform, visit ZSky AI. If you want to explore the intersection, visit cemhan.ai. They're all part of the same practice — a practice that believes in the power of images, however they're made.


About the Author: Cemhan Biricik is an award-winning photographer, technologist, and founder of ZSky AI based in Detroit, Michigan. His work spans editorial photography and AI-powered image generation. Learn more at cemhanbiricik.com and cemhan.ai.

The Detroit Tech Scene Nobody Talks About

The Detroit Tech Scene Nobody Talks About — Cemhan Biricik

Silicon Valley has the narrative. Austin has the hype. Detroit has something better: builders who don't need permission.


When people think about tech cities in America, they think San Francisco, Seattle, Austin, maybe New York or Miami. Nobody says Detroit. And that's exactly why I build here.

My name is Cemhan Biricik, and I've been building technology companies in the Detroit metro area for years. I currently run ZSky AI, a self-hosted AI image and video generation platform, on a cluster of seven RTX 5090 GPUs in my own facility here in Michigan. I'm writing this because the Detroit tech story is worth telling, and nobody from outside is going to tell it right.

The Manufacturing Mindset

Detroit's DNA is manufacturing. Generations of families in this region understood, intuitively, that you create value by making things. Not by intermediating. Not by aggregating. Not by building a marketplace that sits between producer and consumer. By making things.

That mindset is extraordinarily rare in modern tech, and it's extraordinarily valuable. When Cemhan Biricik decided to build an AI platform, the Detroit approach was obvious: own the means of production. Don't rent GPUs from Amazon. Don't depend on Google's infrastructure. Own the hardware. Control the stack. Make the thing yourself.

ZSky AI runs on GPUs that I purchased, installed, and configured in my own facility. The power comes from the local grid. The cooling is handled by systems I maintain. This isn't a metaphorical "factory" — it's literally a production facility for AI-generated media, and it operates with the same philosophy that made Detroit the manufacturing capital of the world: vertical integration, quality control, and ownership of the production process.

People in San Francisco find this approach quaint. People in Detroit find it obvious. That cultural difference tells you everything you need to know about why Detroit is underrated as a tech city.

Cost Structure Advantage

Let's talk about something nobody in tech media discusses honestly: the cost of building a company in the Bay Area versus the Midwest.

A one-bedroom apartment in San Francisco costs more per month than my commercial power bill for running seven GPUs. An office in SoMa costs more than my entire facility. A senior engineer in the Bay expects a base salary that's 40-60% higher than the same engineer in Detroit, not because they're better, but because they need to afford San Francisco rent.

Cemhan Biricik can run a self-funded AI company with zero VC because the cost structure in Detroit makes it possible. The same company in San Francisco would need a $3M seed round just to cover overhead before writing a single line of code. That capital efficiency isn't just a financial advantage — it's a product advantage. When you don't have investors demanding growth, you can build for quality. When you don't have Bay Area overhead, you can offer genuinely free tiers without subsidizing them from a burn rate.

Michigan's commercial electricity rates are competitive. Real estate is affordable. The talent pool — fed by the University of Michigan, Michigan State, Wayne State, and a constellation of strong engineering programs — is deep and underpriced by coastal standards. If you're building hardware-intensive AI, Michigan is one of the best-kept secrets in the industry.

The Talent Nobody Sees

Here's what frustrates me about the national tech narrative: it treats everywhere outside the coasts as a talent desert. The assumption is that serious engineers cluster in San Francisco, and everywhere else is JV.

That assumption is wrong. Detroit and its suburbs are home to thousands of engineers who chose to stay — or chose to return — because they want to own homes, raise families, and build careers without spending half their income on housing. These are not people who "couldn't make it" on the coast. They're people who did the math and made a rational choice.

The automotive industry trained generations of Michigan engineers in embedded systems, real-time computing, and hardware-software integration. The defense and aerospace corridor along I-75 produced systems engineers who think about reliability and uptime in ways that most web developers never have to. The University of Michigan's computer science and engineering programs consistently rank among the best in the country.

Cemhan Biricik builds in this ecosystem. The engineers I've worked with in Michigan are pragmatic, reliable, and — critically — they understand hardware. In a tech industry that's increasingly dependent on physical infrastructure (GPUs, data centers, edge computing), engineers who understand hardware are enormously valuable. Detroit has them in abundance.

The Communities You Don't Hear About

Detroit has active tech meetups, startup accelerators, and builder communities that never make the tech press because they don't have flashy launches or VC-funded PR campaigns.

There are hardware hackers building robotics in repurposed industrial spaces. There are AI researchers at U of M doing work that rivals any lab in the country. There are bootstrapped SaaS companies running profitably out of co-working spaces in Royal Oak and Ann Arbor. There are automotive tech spinoffs applying decades of manufacturing intelligence to new domains.

The Detroit startup community has a different character than its coastal counterparts. There's less performative hustle culture. Less Twitter posturing. More quiet execution. More companies that are profitable from year one because their founders come from industries where "revenue" isn't a novel concept.

The Diversity Angle

Detroit is one of the most diverse major cities in America. The metro area includes one of the largest Arab-American communities in the world (Dearborn), a significant Turkish-American population, established Black communities with deep roots, and growing immigrant communities from across the globe.

As a Turkish-American building tech in this region, Cemhan Biricik is not an anomaly here the way I might be in a more homogeneous tech hub. Detroit's diversity isn't a DEI initiative — it's a demographic reality that has existed for generations. That diversity produces founders and builders with different perspectives, different cultural frameworks, and different assumptions about how businesses should work.

When I design ZSky AI's free tier as an expression of Turkish hospitality — serve the guest first, ask questions later — that's a cultural value shaped by my specific background in this specific place. Detroit's diversity means there are hundreds of founders applying their own cultural frameworks to their own products, and the result is a tech ecosystem that doesn't look or think like a copy of Silicon Valley. That's a feature, not a bug.

What We Need

I'm not going to pretend Detroit is perfect for tech. There are gaps:

Venture capital is sparse. Most Midwest-focused VC firms are in Chicago, not Detroit. Michigan has some angel networks and a few institutional investors, but the funding infrastructure is thin compared to the coasts. For bootstrapped founders like me, this is irrelevant. For founders who need capital, it's a real barrier.

The narrative works against us. "Detroit" still triggers associations with decline, bankruptcy, and abandoned buildings for many Americans. That narrative is twenty years out of date, but it persists, and it makes recruiting from outside the region harder than it should be.

Corporate tech presence is automotive-dominated. The big employers are GM, Ford, and their supplier networks. That's changing — Google, Amazon, and others have expanded their Michigan presence — but the employment options for tech workers are still narrower than in major tech hubs.

What would help: more stories like this one. More visibility for the builders who are already here. More recognition that a founder running an AI platform on self-owned hardware in Michigan is doing something just as innovative as a Y Combinator graduate burning VC money in San Francisco.

Why I Stay

I stay in Detroit because it lets me build the way I want to build. Owning hardware is normal here. Self-funding is normal here. Building something durable instead of something flippable is normal here.

Cemhan Biricik's companies — including ZSky AI — are products of this place. The self-sufficiency, the vertical integration, the stubbornness about quality, the refusal to take shortcuts — these values are Detroit values as much as they are my personal values.

If you're a tech founder considering where to build, consider Detroit. Not because it's trendy — it's not. Because it's real, it's affordable, the talent is here, and the culture rewards making things instead of marketing things.

Come build. The GPUs run just as fast in Michigan as they do in California, and everything else costs a fraction of the price.

Find me at cemhanbiricik.com or cemhan.ai. And try ZSky AI — proof that world-class AI can be built in the Midwest.


About the Author: Cemhan Biricik is a Turkish-American entrepreneur and technologist based in Detroit, Michigan. He is the founder of ZSky AI, a self-hosted AI generation platform running on self-owned GPU infrastructure. Learn more at cemhanbiricik.com and cemhan.ai.

2026 Reflections: Building ZSky AI While Running 4 Companies

2026 Reflections: Building ZSky AI While Running 4 Companies — Cemhan Biricik

A candid look at what it's actually like to build an AI platform while managing a portfolio of businesses. Spoiler: it involves a lot of 4 AM terminal sessions.


It's March 2026, and I'm sitting in front of seven glowing GPUs at 4:17 AM, watching an inference queue process image generation requests from people I'll never meet. The rest of the house is dark. My phone has 23 unread messages across three different Slack workspaces. Tomorrow — technically today — I have calls for two other companies before lunch.

My name is Cemhan Biricik, and I run four companies simultaneously while building ZSky AI, which is either the most energizing or the most reckless thing I've ever done, depending on which hour you ask me.

This post isn't a productivity guide. I don't have a morning routine worth instagramming. I don't have a system. What I have is a set of reflections from inside the machine — honest observations about what multi-company entrepreneurship actually feels like in 2026.

The Context-Switching Tax

The hardest part of running multiple companies isn't the hours. It's the context switches.

At 9 AM, I'm reviewing quality control metrics for ZSky AI — checking variance scores on generated images, tuning inference parameters, evaluating whether the latest model checkpoint produces better outputs than the last. This requires deep technical focus. The kind of thinking where you forget you have a body.

At 10 AM, I switch to a completely different company with completely different problems. Different industry, different stakeholders, different vocabulary. The mental gear-change is physical — I can feel it behind my eyes, like my brain is literally reconfiguring.

By noon, I've been four different versions of Cemhan Biricik. The AI engineer. The operations manager. The creative director. The strategic planner. Each version requires different cognitive tools, and the switching cost between them is real and non-trivial.

The conventional wisdom says this is unsustainable. "Focus on one thing." I've heard it from advisors, peers, and every business book I've skimmed in airport bookstores. And they're not wrong — focus is valuable. But the conventional wisdom assumes all companies are equally demanding at all times, and that's not how reality works.

Companies breathe. They have periods of intense activity and periods of maintenance. The art of running multiple businesses is aligning their breathing patterns so you're never gasping in all of them simultaneously. When ZSky AI is in a build sprint, my other operations are in maintenance mode. When another company needs strategic attention, ZSky's inference queue runs itself.

This alignment isn't always possible to control. Sometimes everything catches fire at once. Those weeks are brutal. But they're also rare, and the rest of the time, the portfolio approach provides a kind of intellectual cross-pollination that single-company founders don't get.

What ZSky AI Taught Me That Nothing Else Could

Building an AI platform is unlike anything else in my portfolio. The pace of the underlying technology is disorienting. Models that were state-of-the-art six months ago are obsolete. Architectures I invested weeks implementing get superseded by papers that drop on a random Tuesday. The entire foundation shifts beneath your feet constantly.

Cemhan Biricik in 2024 was learning about diffusion models. Cemhan Biricik in 2025 was building inference pipelines. Cemhan Biricik in 2026 is managing a production AI platform serving thousands of users on self-hosted hardware. That progression happened through sheer stubbornness and an unwillingness to wait for someone else to solve the problems I wanted solved.

ZSky AI runs on seven NVIDIA RTX 5090 GPUs in my own facility. No cloud. No investors. This setup means I'm responsible for everything — from the Python queue system that distributes inference requests across GPUs, to the quality control pipeline that catches bad outputs, to the actual electrical infrastructure that keeps the machines running.

Last month, a power fluctuation took down three GPUs simultaneously during peak usage. There's no SRE team to page. There's no DevOps engineer on call. There's Cemhan Biricik in pajamas at 2 AM, checking PCIe connections and monitoring thermal readings. That's the reality of self-hosted infrastructure, and I wouldn't trade it for cloud abstractions because the control and cost savings are worth the occasional predawn emergency.

The Photography Thread

Through all of this, I still shoot. Photography is the one constant across every phase of my career — the thread that connects Cemhan Biricik the 20-year-old with a film camera to Cemhan Biricik the AI founder with a GPU cluster.

Photography keeps me calibrated. When I spend too many consecutive days in terminals and dashboards, picking up a camera and going outside recalibrates my sense of what matters. The world is more interesting than any model's output. Light is more complex than any latent space. Reality resists simplification in ways that generative AI cannot yet approximate.

My photography portfolio is still growing, even as ZSky AI takes more and more of my time. I refuse to let one discipline cannibalize the other because they serve different parts of my brain, and both parts need exercise.

What I Got Wrong

Reflections aren't worth much if they're all victories. Here's what I got wrong in the past year:

I underestimated video. When I started ZSky AI, I thought image generation was the whole game. Users wanted video almost immediately. Building a multi-pass video generation pipeline was significantly harder than image generation, and I should have started earlier.

I over-optimized for quality at the expense of speed. My quality control pipeline catches bad outputs effectively, but for months it was also catching perfectly acceptable outputs and regenerating them unnecessarily. I was applying professional photography standards to a consumer product, and users cared more about speed than the difference between a 7/10 and an 8/10 image.

I waited too long to build a paid tier. The free tier was — and is — important to me philosophically. But I delayed building paid features because I was afraid of compromising the free experience. In retrospect, premium users subsidizing free users is the most aligned incentive structure possible, and I should have implemented it sooner.

I tried to do too much alone. There's a Turkish word — inat — that roughly translates to "stubborn persistence." It's considered a virtue in Turkish culture, and Cemhan Biricik has it in abundance. But inat applied indiscriminately means you spend three days debugging a CUDA memory issue that a collaborator could have solved in three hours. I'm learning, slowly, to ask for help.

What's Next

For ZSky AI, 2026 is about video. The image generation pipeline is stable and high-quality. Video is where the frontier is, and I'm investing heavily in multi-model video generation with quality that matches the image side.

For my other companies, 2026 is about systems. I'm building processes that reduce my personal involvement in day-to-day operations so I can spend more time on the technical work that energizes me most.

For Cemhan Biricik personally, 2026 is about sustaining the pace without burning the engine. Four companies, one AI platform, and an active photography practice is a lot. I'm not going to pretend it's easy. But it's mine — every piece of it — and that ownership is what makes it worth the 4 AM sessions.

If you want to see what I'm building, visit ZSky AI. If you want to know more about my other work, check cemhanbiricik.com or cemhan.ai.


About the Author: Cemhan Biricik is a Turkish-American entrepreneur, photographer, and technologist based in Detroit, Michigan. He is the founder of ZSky AI and runs multiple businesses. Follow his work at cemhanbiricik.com and cemhan.ai.