What a day! Presenting in front of the pioneers who defined the industry and whom I looked up to all these years. It was an absolute honour to be there talking about the next big shift.

I am so grateful to Tom Mullany and the Silicon Project at Stanford for having me.

How many of you wrote by hand for more than 30 minutes this week?
I still write often, but not as much as I used to, and lately I’ve started to feel a different kind of stress in my wrist when I do.

Do you want to take part in a small reading experiment?
Visit this link niteeshyadav.com/journal/readin…. When you land on the page, you’ll see a short pop-up explaining the study and inviting you to opt in.
From there, you can customise your reading conditions and use them while you explore the articles on my site.

Try to 404 on niteeshyadav.com/404 for a surprise 😉

There's something deeply satisfying about building a team you can learn from.

Since 2023, I'd been trying to hire a design engineer for my team. I reached out to several people I knew, explored different possibilities and finally, last year, I came across @Harsshavardan3. And what a ride it's been since then.

We've built so many things that had previously existed only on idea boards. But what made this experience special wasn't just what we built. It was getting to work with someone I could learn so much from. I've never really told Harsha this, but he's had quite an impact on how I think about building and experimenting. I guess he'll find out through this post :)

In early 2025, I started an initiative to carve out a few hours every week for the team to explore ideas beyond their regular projects. This wasn't easy, especially in a fast-paced startup environment where there's always something urgent to work on. But some interesting ideas came out of it, and what I enjoyed most was seeing people take ownership of things they were genuinely curious about.

As Harsha started bringing ideas to life after joining, I realised how much I could learn from the way he worked. During reviews, I'd often find myself asking him how he'd built something or made it work. Instead of just looking at the output and saying "Good job," I wanted to understand the process. How did he approach it? Could I try building something like that myself?

And I think that's something we sometimes lose as we grow into leadership roles.

You spend more time making decisions, managing people and setting direction, and less time actually making things. It can get a little lonely at times. And somewhere along the way, that feeling of not learning enough starts to catch up with you.

And it wasn't just Harsha. I've learnt so much from my team at Pocket FM, be it designers, researchers or content designers. Everyone brought a different perspective, challenged my assumptions and taught me things I probably wouldn't have learnt on my own.

In fact, I often found myself happier when someone proved me wrong. Because that meant there was something new for me to learn. It's something I've always tried to encourage within the team as well. To question things, challenge each other and be comfortable with being wrong.

Over time, I also started looking for more avenues to learn. There were things I wanted to understand better, skills I wanted to develop and research questions I'd been carrying around for years, particularly my work on adaptive typography and interfaces. But I never quite found enough time to explore them the way I wanted to.

That was one of the reasons I decided to take a short break. My family needed me at the time, and I wanted to be there for them. I also saw it as an opportunity to return to being a student for a while, explore things without an immediate business goal and finally give shape to some of my longest-running research interests.

I'm grateful to Prateek Dixit for encouraging me to take this break in the first place and to @hvpandya for stepping in and taking on the responsibility of leading the team. Without his support, taking this time off wouldn't have been possible.

Over the past few months, I've been reading, experimenting, building and, in many ways, rediscovering the joy of learning without an agenda. I'll be sharing more of that journey soon, before I get back to work.

If there's one thing this experience has reinforced for me, it's that growing as a designer isn't about reaching a point where you know enough. It's about staying curious, surrounding yourself with people you can learn from and being willing to become a beginner again.

I’d nudge in the opposite direction.
I encourage designers not to optimise for token-maxing, but for life-maxing, because that’s where real ideas usually come from. I’ve spent the last few months in token-maxx mode and it left my brain feeling oddly stale.
If your whole day is just “how do I squeeze a bit more juice out of AI”, you eventually stop thinking about anything else and the work starts to reflect that.

Love this framing. In design, most people don’t build their own taste, they follow trends. Designers and even design leaders keep copying what big orgs do instead of deepening their understanding and shaping a point of view.

Building taste often means doing something that looks “off” at first, because it’s not what people are used to. And when you’re answerable to many stakeholders, it’s easier to choose what’s already been seen than to back something new. And that is what gives you an opportunity to build something memorable.

The outrage over a designer’s video portfolio, framing it as somehow setting a new bar for hiring, is so petty.

Creating things is hard. A lot of thought, effort and experimentation goes into making something good. And instead of appreciating that, people are busy complaining about how unfair it is.

Yes, some people have more resources, better connections or happen to be born in the right place at the right time. Of course that matters. It matters for individuals, organisations and pretty much everything in life.

But I think we're missing the point. It's not always about resources. Sometimes it's about thinking differently and going the extra mile. I landed my internship at 2×4 in Beijing with a hand-lettered résumé. It didn't take extraordinary resources, just some extra effort and the willingness to try something different.

And guess what? Life is unfair. It always has been and probably always will be. You can spend your time complaining about what others have, or put that same energy into building something with whatever you have.

I'm not saying I've never had these thoughts. I have them too. I've compared myself to others, felt frustrated about opportunities I didn't have and wondered how much further I could have gone with better circumstances.

But the point is to acknowledge those thoughts and move on. Get back to making things. Whatever little I've managed to achieve so far has largely come from doing just that.

Of course, we should question unfair systems and unequal opportunities. But pulling someone down for making something good isn't going to make things any fairer.

The sooner you accept that life isn't a level playing field and start focusing on what you can control, the better off you'll be. Otherwise, it's easy to spend years stuck in an endless loop of complaining.

Someone else's good work isn't your disadvantage. Appreciate it, learn from it and go make something yourself.

Love this question.
I don’t think AI is one fixed thing for designers, it really depends on how we choose to work with it.
Right now, a lot of its use is about speed: getting from a blank screen to something, generating variations, moving projects forward faster. That’s useful, but it mostly lives in the iteration space.

I like thinking of AI as a camera. You can leave it on auto mode and take hundreds of decent photos quickly, or you can slow down, understand light and framing, and wait for the right moment to capture something that feels genuinely new. Same camera, very different outcomes.

AI feels similar. It can absolutely support innovation, but only when designers bring their own questions, judgment and curiosity into the frame, instead of just accepting whatever the “auto mode” gives back.

Designers in tech, spend some time looking at what’s happening in graphic design. The level of experimentation, expression and play is refreshing. Meanwhile, tech design has become so stale that it stinks.

A lot of people are celebrating the productivity gains of AI in design. But if you look closely, most of it is just speeding up the path to enshitification because yesterday’s baseline quickly becomes today’s junk. When volume scales faster than judgment, we don’t get better work, we just get more of what already wasn’t good enough.

I’ve always wanted to build AI that augments my learning while leaving the thinking to me. After setting up my bot workflow, I extended it towards ambient learning: an assistant that understands my work and brings ideas back when they matter. This is v0.1, fully local, so my work stays on my device.

We keep talking about taste like it is some ethereal, untouchable thing.

Lately, though, the tech crowd has a very different take: taste is just another engineering benchmark on a roadmap. The argument says taste is merely high-level pattern recognition, an intuitive grasp of balance, color palettes, rhythm, and tone. Because large language models can chew through centuries of culture and spit out balanced designs or lyrical sentences in seconds, people are starting to say that taste is not a permanent boundary at all. It is just an optimization problem, and the machines are already crossing it.

I completely disagree.

That view mistakes the canvas for the painter. It treats taste as a technical output, when in reality it is a deeply human relationship between an expression, a story, and an audience.

What actually is taste?

At its core, taste is a kind of self expression. That self can be an individual artist, or it can be the collective ethos of a brand or an organization standing up and declaring: we want to be like this, not that.

Society has always placed its highest regard on originality. Whether we are talking about a tailored jacket, a canvas, a piece of furniture, or any other object of desire, we do not fall in love with statistical averages. We gravitate toward deliberate, risky departures from the norm.

An AI model has no self to express. It has no reputation to defend, no lived childhood, no broken heart, and no stake in the outcome. It operates on pure probability distributions, predicting what fits based on precedent. It can produce visual symmetry or effortless cadence all day long for near-zero cost.

When clean aesthetic harmony becomes dirt cheap and infinite, the cultural value of that harmony collapses. You do not win points anymore just for making something that looks balanced. Rarity moves elsewhere.

The story behind the process matters more

Taste is never evaluated in a sensory vacuum. It is decided at the receiving end, by the viewer looking at both the object and the story of how it came to exist.

Take mechanical watches. A five dollar battery-powered quartz watch keeps better time than a handcrafted Patek Philippe. Yet people willingly spend tens of thousands of dollars on the mechanical watch. They are not paying for utility. They are paying for the friction: the resistance of the metal, the heritage, the physical patience of a watchmaker bent over tiny gears.

Or look at what happened when cameras automated realistic imagery. Painters did not close their studios and give up. They pivoted inward and invented impressionism, cubism, and abstract art. The value shifted from copying reality to interpreting it through a subjective, mortal human mind.

Think about art forgeries, too. A master forgery can match an original painting atom for atom. The visual beauty is identical. But the second the lab proves it is a fake, its value drops to zero. The object did not change at all. Only the story about its origin changed.

Taste is a proxy for human trust. When you connect with a work, you are reaching for the presence of another conscious being who wrestled with constraints, made a hard choice, and stood behind it.

Can machines build their own stories?

This does not mean automated systems cannot have a narrative.

Imagine an apparel brand that skips traditional artisan workshops entirely. Instead, an engineer builds an autonomous pipeline of AI agents that scans cultural shifts, generates efficient zero-waste clothing silhouettes, and coordinates production directly with factories.

Is that bad taste? Not necessarily. It is a brand-new kind of story.

Much like the Bauhaus movement in the 1920s celebrated tubular steel and the reality of the assembly line instead of apologizing for it, modern audiences can find genuine aesthetic power in radical computational efficiency. But notice where the meaning comes from: it still comes from human intention framing the system as a cultural statement. The code generates the shapes, but human perspective turns the pipeline into a point of view.

Why skin in the game is the real barrier

Could an AI eventually develop an intrinsic point of view all on its own, without a human steering it?

I think it can. Just like AlphaGo played Move 37 against Lee Sedol, defying centuries of human game theory with a move that seemed alien at first, a machine can develop its own internal logic and non-human aesthetic patterns.

The real question is: will humans care?

It is up to humans to accept that output as valuable, and looking at how our culture works, genuine taste requires the creator to have skin in the game. Intrinsic coherence alone is never enough.

To see why, consider a thought experiment. Suppose a creative director builds an entirely AI-generated fashion brand, with zero hand drawings and zero human-written copy. But this person takes out a second mortgage on their family home, quits their day job, and gambles their life savings to manufacture the line. If it fails, they are bankrupt.

Does that financial risk mean the collection now possesses genuine taste?

It makes for a provocative story, but it is still not genuine taste.

A financial bet is not artistic conviction. A gambler placing their life savings on a roulette wheel has immense skin in the game, but nobody confuses placing a bet with making art.

If that synthetic clothing brand flops, the investor’s bank account takes a hit, but their soul was never on the line. They did not expose their private grief, their awkward obsessions, or their own vulnerable view of the world. They can easily walk away, shrug, and blame the algorithm.

True artistic skin in the game does not offer an escape hatch. It requires you to put your personal perspective on the line and say: this is how I see reality, and if you think it is shallow, you are calling a part of me shallow.

A model cannot do that. It cannot bleed, it cannot face social embarrassment, and it has no mortal life to lose.

The paradox of this essay

That brings this whole conversation to an interesting mirror.

What is the value of this essay you are reading right now?

These words were synthesized by an artificial intelligence. The prose was organized by a model that has no fear of rejection, no reputation on the line, and no soul to bare. By the very definition I have laid out, the software that generated these sentences cannot claim genuine taste.

Yet this essay did not appear out of thin air. It is the artifact of a direct conversation between an AI and me, Niteesh. It exists because I brought a conviction to the screen, challenged the tech consensus, and pushed back against easy answers to pinpoint where human meaning actually lives.

The machine supplied rapid synthesis and structure. The human supplied the friction, the direction, and the core idea.

So, does this essay carry real value, or is it just another piece of automated mimicry?

The algorithm cannot answer that. That judgment belongs entirely to you, the person on the receiving end, deciding your own definition of taste.

✨ We’re hiring AI Designers at @PocketFM_App ✨

If you use AI & Code day in day out in your Design role at work, this would be a great fit.

👉 Use SOTA models & tools
👉 Don’t just design, build & ship your work
👉 Unlimited ownership

Here: pocket.design

If you care about the planet you live in, then you should watch this youtube.com/watch?v=h3mRsv…

Is anyone facing an issue with ChatGPT here generated responses are vanishing and then appearing on killing the app and relaunching with a bit of modifications in the response?