We are building a consumer AI startup to solve for AI's lack of taste — and sharing our story as we go.
Two serial co-founder’s experiment in elevating taste, with AI.
Hi, I’m Neil, [and I’m Az]. And this is our first Substack post….inhale….
Hopefully most of you will remember the genre-defining podcast, Startup from Gimlet - the brilliant, meta podcast about starting a podcast company in 2014. It didn’t just tell a story, it captured a moment when podcasting shifted from fringe to mainstream and entered the zeitgeist.
Az and I feel like we’re standing at a similar, but much more significant inflection point. Except this time, the shift is happening with AI.
So, inspired by Gimlet, and with assistance from our AI Co-Founder (meet Tex below), we’re documenting our journey back into startup life - this time in the AI age.
Say hello to Tex (he’s our 10X multiplier - super-smart but clearly lacking in taste)
Not our first rodeo
A little bit of background for context; Aaron & I are both long-time builders, brand shapers, cultural nerds, and curious observers. Between us, we’ve built, scaled and exited Australian media and technology companies like Junkee Media and Lexer.
Earlier this year, after wrapping our respective chapters, our paths crossed once again and we started having some exciting conversations about this once-in-a-generation moment. We were both naturally drawn to the spaces where AI doesn’t have all the answers. The spaces where there will always be a need for a human-in-the-loop to apply a critical lens.
A founding truth emerged:
AI has no taste.
[a friend texted this to me when they heard about my hypothesis 😊]
On a recent trip to the UK, AI helped with logistics, but when it came to uncovering my London, it completely missed. It didn’t catch the scent of the off-menu dish wafting past at Ottolenghi. ChatGPT’s recs were no match for my friends’ - because they’d actually been there, for real.
Large language models are trained on an enormous and diverse corpus of online data, but that data is skewed, heavily toward content that is widely available, easily crawled, and highly duplicated - In other words, mass culture.
This is fine, if your idea of a perfect holiday is a top-10 TripAdvisor list. But not if you’re the kind of person who seeks out third-wave coffee roasters hidden in Kyoto backstreets.
That’s where taste plays its fundamentally human role. Taste lives in discernment, in the beautifully subjective edge-cases, in the quiet confidence of knowing when you’ve made a great choice. In the goosebumps and tingles. In breaking with convention and pattern, because it makes the experience unique.
And yet AI is everywhere. Eating the internet (and repeating many of its mistakes). The ground beneath us is shifting in terms of how we create, curate, and consume.
So Aaron and I are doing what founders do when the ground starts shifting:
We’re building again.
What we are building exactly, will reveal itself in time. We are starting with the why, the why now, and the how.
And this time, we are not just using AI - we are building with it.
This Substack is our build log.
But it’s also an experiment.
What happens when two founders and an AI, co-author a journey in real time?
That led us to a simple but non-negotiable principle for our startup:
Start EVERYTHING with AI. Not as a bolt-on. But as the central nervous system.
This means rethinking everything: the product, the process, the people, and yes, even the storytelling. And this doesn’t diminish the creative output, far from it. It sticks a 🚀 up it!
AI-native is not a new concept. But this Substack isn’t just about building with AI.
It is building with AI.
Tex - our AI co-founder helped shape this very post. Refined it. Challenged it. Nudged it where it needed to go. Unashamedly so.
And if we are serious about elevating human taste in an AI-native world, it starts here: with transparency, collaboration, and the courage to be seen figuring it out in public and with anyone reading this being the ultimate taste-filter on the end product.
What to expect from us:
Expect imperfections. We will use AI but speak in our own voices.
I come at this as a strategist, a builder, and a lifelong observer of shifts in human behaviour. That lens will shape everything I explore here.
Az comes at this as a creative technologist. He’s driving the foundational technology, data and product choices we make and he will share them with you.
We will capture what we learn about building an exceptional team, match-fit for the AI era and share the highs and the lows of fundraising and the state of VC in Australia.
We’ll unpack how we are using AI to think faster, go deeper, and challenge ourselves harder and importantly - where AI's limitations are. Our AI Co-Founder Tex may even break the fourth wall on occasion and join the narrative.
We’ll reflect on what it means to hold onto and elevate taste and human discernment, even as the noise gets louder.
We will throw in a few good (and slightly time-embellished) tales from previous startup adventures (the wins and the disasters).
And we’ll drop the ideas, links, and perspectives that are fuelling our thinking in real time.
The what will come. But for now, we’re starting with the why.
Let’s go!
Neil [and Az]
A parting provocation: How much of this post is us, and how much is Tex? I’ll let you decide if it even matters.
Great to be on the journey with you Neil!
Very cool! I am jumping to solution mode but anything you train and be taught the habit of looking for the unique or off the beaten track in data. Very cool problem to solve.