Humanity once prided itself on the greatness of the creative process, the ultimate preservation of taste and judgement.

Poetry, plays, paintings, religious works.

All are artefacts of human thinking, emotions, opinions and reflections of social zeitgeists. That is, the defining mood, collective mindset, and cultural climate of a specific era.

The letter of a war-torn soldier writing back to his spouse, DaVinci's Mona Lisa, Shakespeare's Othello, the journal entries of an individual during the Covid-19 pandemic, they are the purest exercise and rationalisation of our thoughts and experiences.

Heralds once would yell on the streets of medieval and renaissance societies, spreading news of the town- whether real or bribed, For most of human history, communication was judged not only by whether it was correct, but by whether it moved people, carried meaning, and fit the moment. We were concerned with what sounded great.

As humanity and society progressed towards the rise of academia, there was a shift, an inversion of human taste and judgement. We began to prioritise what was correct over what sounded great.

Societies, countries, governments were no longer defined by their cultural phenomena, but whether they were able to successfully launch rockets to the moon with the help of human computers- every individual concerned with being correct with their mathematical calculations.

Where we are now:

In the present, we have reached a new third state:

With the help of Artificial Intelligence, we are able to pull from academia, sources of truth and every opinion that can span the internet.

This is the maximal extension of what modern societies are prioritising; correctness.

This spans the news articles we consume, social media posts, educational content, text messages, emails, almost every surface we absorb daily.

Probabilistically and metaphorically, artificial intelligence exposes us to the average of all outputs.

Although we train models on sources like Shakespeare's Othello, we look only at the final output- we can never trace through Shakespeare's creative process, the thousands of lines discarded, burnt amongst a pile of many more creative outputs. That is where taste and judgement is encapsulated.

Previously, I was both a full-time student and a full-time employee at a seed-stage startup. A student in Melbourne, working from San Francisco.

While AI has made this combination practical and manageable for the first time, there is little to no preservation of human taste in across artefact of work.

AI tools have become central to all work: Claude Code for engineering and product, Claude for writing, Claude for quite literally anything.

My thoughts are injected at the input level, but the output is a result of layers of AI slop.

I once prided myself on high taste output, great product designs, insightful write-ups for marketing and sales decks.

Now, exposure to AI content on the internet and an over-reliance on tools have led to a personal frustration regarding quality of output- also much to the realisation of the teams that I work with.

Questions always tread along the lines of:

Donald what happened to your old writing style?

Whether you wrote this or not, I definitely perceive this as AI slop.

I have drifted into a tarpit of prioritising speed of execution > quality of execution.

The lack of human judgement and taste was noticeable within a month.

If this is the present state of AI-first operators like myself, what does this have to say for the future where every individual is nurtured with AI from first-principles?

The convergence problem:

In a world rapidly converging toward a homogenised AI dialect, the preservation of genuine human voice becomes not just valuable, but essential.

The AI dialect is a derivative first of english, and technically every language thereafter, where humans begin to speak in the patterns, conventions, and thought processes of an AI model. Whilst it’s still English, it is a deviation far from how humans actually speak, and it is an abstraction away from our deepest thoughts towards the world. For the first time, we lose a language to an intelligence that isn’t our own. One that was seeded by ourselves, which we succumbed to without realising.

Think of the following as a chain of events in the bullish case:

Industries, institutions and the individual begin to mass adopt AI.
Every output is either entirely dependent on AI, or directly influenced by AI
The shows we watch are scripted with AI, the Instagram captions we read are written with AI, the literature we read is written by AI
As consumers of content, the 'AI dialect' trickles into every output we produce: text messages, emails, even the way we speak
Every piece of text flowing between systems and people are now influenced by AI whether we know it or not, we are thinking from an AI's perspective and not one of our own.

Human thought is the catalyst for judgement and taste, when either ends of this equation fall out, the others suffer. When the way we think, speak and write changes, every human system built on the foundations of judgement regress.

When we are conditioned only on being correct, rudimentary systems change:

Delivery of healthcare; we will empathise less with the patients we see, fail to articulate diagnosis with care and precision.

Customer service; agentic order enquiries are fulfilled, but customers are never left wanting to come back for more, retail becomes purely transaction based.

TV shows; Bluey starts speaking in irregular patterns, we feel like something is off, but we can't pinpoint it.

We are wiring the next generation of humans to think from AI-first principles.

Everything we do, stems from consuming artificial intelligence, everything we do thereafter, converges towards the same average.

This is where the problem space for Tact (let's call this the company name for now) emerges.

Tact: the skill of handling delicate situations or delivering difficult truths in a considerate and respectful way.

The fundamental question we seek to solve:

If artificial intelligence is the future of correctness, how do we ensure that the delivery is tasteful and anchored on human judgement?

How can we be both CORRECT and GREAT?

We need to propagate taste and human judgement as outputs across frontier models and emerging AI products globally.

That's where the solution ties in.