ITK: Not the center of the universe

Read past editions of “Inside the Kitchen” →

Welcome back inside the kitchen. 🔥

I’m glad you’re here, because 99% of people are missing what’s going on beneath the surface right now. Let’s dive in together.

This week’s take: We are not the center of the world we are building. But AI might be.

That sounds dramatic, but I mean it in a very practical way.

Whenever a powerful new technology shows up, we do not just use it. We slowly reshape the world around it.

Cars are the easiest example. At first, cars were just a new way to move around. But over time, we rebuilt entire cities for them. Roads widened, parking lots appeared everywhere, drive-thrus became normal, suburbs expanded — and traffic lights, highways, gas stations, garages, insurance rules, and speed limits became part of everyday life. It feels natural to us now, but it wasn’t natural until we redesigned the world around the car.

The same thing happened with elevators. Once buildings could move people vertically, cities started building upward. The skyscraper is not just a building style. It is what happens when architecture adapts to a machine, and people no longer have to take the stairs.

And don’t even get me started on how smartphones have changed everything from buying concert tickets to managing finances and finding new friends …

The point is not that these changes are all bad somehow. A lot of them are useful. Cars made travel easier. Elevators made modern cities possible. And smartphones can make daily life more convenient.

But the pattern matters:

At first, technology adapts to us. Then, slowly, we adapt to it.

I think AI is entering that second phase.

At the beginning, AI felt like a tool you opened when you needed help writing an email, summarizing an article, creating an image, or understanding a new idea. But now the world is starting to bend around it.

Websites are being written so AI systems can understand and recommend them. Companies are changing how meetings are structured because AI can summarize them. Teachers are altering assignments because students have AI. Customer service is being redesigned around chatbots. Even job applicants are writing resumes for AI screeners instead of humans.

That is the part worth noticing.

AI is no longer just sitting inside the world. Instead, the world is being formatted for AI.

And once that starts happening, the ethical question gets bigger than “Is this tool helpful?” The question becomes: what kind of world are we building so this tool can work better?

Because sometimes making the world easier for machines makes it harder, colder, or stranger for people.

Think about it.

A warehouse that is optimized for robots may become less comfortable for human workers. A hiring process optimized for AI screening may reward people who know how to write for the machine instead of people who are actually better for the job. A school system built around AI detection may make students feel like suspects rather than learners. A workplace where every meeting is recorded and summarized may become more efficient, but also less human, less private, and less forgiving.

None of these are sci-fi examples. They are small, real adjustments. That is what makes them powerful. Individually, each change seems minor, but together, they start to reshape the world we live in.

This is where I think the “humans are not the center” idea matters. We like to imagine technology is built around human needs. But often, once a technology becomes powerful enough, humans begin organizing their behavior around the needs of the technology.

We drive where roads allow us to drive. We work in buildings shaped by elevators. We communicate in ways shaped by smartphones. And now we may begin thinking, writing, teaching, working, and even remembering in ways shaped by AI systems.

That does not mean AI is evil. But it does mean that AI is becoming infrastructure. And infrastructure has a way of disappearing into the background until we forget we ever had a choice.

The futuristic version of this could get strange.

Imagine cities designed so delivery robots can move more easily than pedestrians. Schools where every assignment is structured mainly to be readable by AI graders. Workplaces where every conversation is recorded because the AI assistant needs a complete memory. Websites written less for human readers and more for machine summaries. People changing how they speak, write, apply for jobs, create art, or explain themselves because they know an AI system will be the first judge.

That future probably will not arrive all at once. Instead, it will arrive through convenience: one small optimization at a time.

So the danger is not that AI suddenly takes over the world in some obvious way. The danger is that we slowly rebuild the world to make AI more useful, and only later realize we made it less human.

So the question is not just, “How can AI fit into our lives?”

The harder question is, “How much of our lives are we willing to reshape so AI can fit more easily?”

Because history shows that once a technology becomes important enough, the world starts making room for it.

This time, we should be careful about how much room we give away.

(This is for educational purposes only… not investment advice)

This week’s take: The Robot Games showed something important about AI: the next bottleneck is not just making robots smarter. It is helping them understand messy physical reality.

The videos are fun to watch. Humanoid robots running races, playing soccer, doing obstacle courses, and more. But honestly, the most interesting moments are when they fail.

A robot misses a step, runs into a wall, or malfunctions. A machine moves well in one setting, then gets confused when the world is slightly different. The funniest moment I’ve seen was a robot knocking itself out with a kick!

But this is the real physical AI problem:

Chatbots live in text, which is often clean. But robots live in the world, which is messy.

A robot does not just need a bigger model. It needs to process cameras, radar, depth sensors, motion, heat, obstacles, lighting changes, weird angles, bad data, and unexpected movement in real time.

Most of that data is not useful. A self-driving car or warehouse robot can generate tons of sensor data, but much of it is repetitive, noisy, or irrelevant. The valuable part is knowing what to ignore, what to keep, and what needs to be handled immediately.

That is why edge AI matters: AI that operates on-device.

Instead of sending every bit of raw data back to the cloud, devices need to filter and understand the world locally. A robot needs to know what matters right now. A car needs to process the road instantly. A factory camera needs to spot the defect without waiting on a data center.

So the next AI infrastructure layer may not just be training bigger models. It may be cleaning, filtering, and interpreting physical data at the edge.

$QCOM ( ▼ 0.36% )

Qualcomm is still mostly known as a smartphone chip company, but that undersells what it is becoming.

The company is increasingly positioned around edge AI, which means running AI directly on devices instead of relying completely on the cloud. That matters for physical AI because robots, cars, drones, industrial cameras, and smart machines often cannot wait for a faraway server to tell them what to do. They need fast, local intelligence. Qualcomm’s advantage is that it already knows how to build low-power chips that handle sensing, connectivity, and on-device AI — exactly the kind of architecture physical AI needs.

If AI keeps moving from screens into cars, robots, factories, drones, and physical devices, then the important question becomes: where does the intelligence actually run? For a chatbot, the answer can be a data center. For a robot in a warehouse, a car on the highway, or a machine on a factory floor, the answer often has to be much closer to the action.

That is where Qualcomm gets interesting. Physical AI needs chips that can process messy real-world data quickly, efficiently, and locally — and Qualcomm is one of the companies built for exactly that kind of world.

So while you watch highlights and fails from the viral Robot Games, let it be a reminder that robots do not just need brains. They also need perception — they need to understand the world as it changes. That’s the next step.

Hi — I’m Hunter, a PhD candidate whose work has appeared in major academic journals and popular tech outlets. I founded FryAI to make staying ahead of AI clear, accessible, and fun.

📧 Email me directly if you have any questions or want to chat!