Version 1 to version 2

I have been publishing a lot.

Not in the old sense of publishing, where a person disappears for three months, returns with a polished essay, and pretends the finished piece arrived fully formed. I mean publishing in the more contemporary sense: I see something, I think about it, I connect it to design, technology, cities, culture, business, or whatever else is currently occupying my head, and then I turn it into a piece of writing.

Sometimes this happens quickly. Very quickly.

That is where the problem begins.

From inside the process, it feels productive. I am reading widely, noticing patterns, collecting examples, and testing ideas in public. Raw Drafts is supposed to be a place where thinking stays somewhat unfinished. The name itself gives me permission not to resolve everything before publishing. It is a notebook, a studio, an archive, and sometimes a way of figuring out what I believe.

From the outside, though, it may look less romantic.

It may look like a content machine.

There is a familiar shape to many Raw Drafts pieces. I find an article, a product, a strange company, a piece of infrastructure, an interface, or a small cultural phenomenon. I ask why it exists. I compare it to something else. I pull out a wider implication. Then I publish.

If I repeat that process often enough, the individual subjects begin to matter less than the mechanism producing them. One day it is AI game development. The next day it is ambulances, Michelin restaurants, high-speed rail, obscure industries, digital museums, public Wi-Fi, or some startup that has redesigned a familiar workflow. The topics are different, but the editorial motion can become suspiciously similar.

This is one of the defining problems of AI-assisted writing. The bottleneck used to be production. You could have fifty ideas, but writing fifty essays required fifty separate acts of endurance. Now production is cheap enough that the real bottleneck is judgment. The difficult question is no longer whether I can publish something. It is whether this particular thought deserves to become public at all.

That sounds obvious, but it changes the entire nature of the work.

Spam is not only bad writing. Spam is also writing that exists because it can exist. It is content produced without enough resistance, consequence, or necessity. It may be technically competent. It may even contain true information. What makes it feel like spam is that nobody seems to have needed to say it.

This is where Raw Drafts becomes difficult to defend.

I can tell myself that each piece represents curiosity. That is often true. I genuinely like learning about industries most people ignore, understanding why systems are designed the way they are, and finding connections between things that are not normally placed next to each other. I do not experience the work as keyword farming. I am not sitting down with a spreadsheet of search terms and asking how many pages I can extract from them.

But intention is not the same as effect.

A reader cannot see my internal curiosity. They only see the output. If the site contains a growing pile of articles reacting to whatever crossed my screen that day, then the distinction between an active mind and an automated content operation may become harder to perceive.

That is especially true when the writing becomes too smooth.

One of the strange effects of AI is that it can remove the visible evidence of thought. Real thinking is repetitive, hesitant, uneven, and sometimes contradictory. It changes direction halfway through. It contains details that are not strategically useful but matter to the person writing. AI-generated prose often does the opposite. It compresses uncertainty into structure. It replaces obsession with coverage. It gives every idea a clean setup, a balanced middle, and a conclusion that sounds like it was approved by a small committee.

A piece may become more readable while becoming less alive.

Raw Drafts is most convincing when the writing contains something I could not have produced by asking for a general article about the topic. A particular train journey. A website I built and broke. A conversation in Taiwan. A strange detail from an Estonian institution. A client problem that looked simple until the operational system underneath it collapsed. A personal irritation that slowly became a larger question.

Without that layer, the article may still be useful, but it becomes interchangeable. Another person could have published it. Another model could have generated it. Another publication could absorb it without changing much.

That is probably the closest Raw Drafts comes to spam: not when it is wrong, but when it is replaceable.

At the same time, I am not convinced that the answer is to slow down until every piece becomes precious.

There is a slightly dishonest mythology around careful writing. We tend to assume that time spent equals depth produced, as if an essay written over six weeks must contain more truth than one written in an afternoon. Often the opposite happens. A quick note can capture a live question before it hardens into a position. A rough piece can preserve the energy of discovery. Publishing frequently can also reveal recurring interests that would remain invisible if I waited for a fully formed thesis.

The archive may be doing something that no single article can do.

Raw Drafts is gradually showing me what I keep returning to. Systems that work badly because nobody owns the whole experience. Industries that remain invisible until a failure occurs. Products that look polished but hide operational disorder. Technologies that increase output while making judgment more important. Places where design stops being decoration and becomes coordination.

I did not begin with a formal editorial strategy around these subjects. They emerged through repetition. In that sense, the volume is not only a weakness. It is also a method.

The problem is that method and spam can look almost identical from the outside.

Both produce a lot. Both follow patterns. Both accumulate searchable pages. Both can use AI. Both can react to existing material. The difference is not visible in the publishing frequency. It appears in whether the work develops a memory.

Does one essay affect the next one?

Does an idea return in a more precise form?

Do I revise my beliefs?

Do observations from Taiwan change how I think about Estonia? Does work on WiFi.ee change how I understand public infrastructure? Does building HAAM change how I read claims about AI productivity? Do earlier essays become evidence, rather than just old URLs?

Spam forgets immediately. It treats each page as a separate container. A real body of work becomes cumulative.

This may be the standard Raw Drafts should use.

The question is not whether every post is excellent. That would defeat the point of calling them drafts. The question is whether the drafts are going somewhere. An unfinished piece can still be part of an unfolding argument. A lightweight reaction can still matter if it becomes material for a deeper line of inquiry. Even a wrong interpretation can be useful if it remains available to be corrected later.

What would make Raw Drafts feel empty is not roughness. It would be endless novelty without development.

There is also a more personal problem. Publishing can become a substitute for deciding.

It is satisfying to turn uncertainty into an article. The article gives the feeling that something has been processed. But writing about a problem is not the same as changing the situation that produced it. I can write about AI replacing judgment while using AI to avoid making difficult editorial choices. I can write about companies creating too much content while producing another post about companies creating too much content. I can describe HAAM's positioning with increasing precision without necessarily doing the slower work of selling, delivering, measuring, and proving it.

Raw Drafts can be a thinking tool, but it can also become a very sophisticated form of procrastination.

That does not make the writing fake. It makes the relationship between writing and action worth watching.

Perhaps the most useful distinction is between writing as extraction and writing as transformation.

Extraction takes an article, a trend, or a product and converts it into another piece of content. The source enters, gets summarized, receives a few observations, and exits in a slightly different format.

Transformation is slower, even when the writing itself is fast. The source collides with something already present: an experience, a project, a contradiction, a memory, a decision. The resulting piece could not exist without that collision.

Raw Drafts becomes spam when it mostly extracts.

It becomes worthwhile when it transforms.

I do not think this means every article needs to be deeply personal. That would produce another kind of performance. Not every topic needs a childhood memory, a travel anecdote, or a dramatic confession. Sometimes the honest contribution is simply a comparison nobody has made, a neglected example, or a question that exposes a weak assumption.

But there should be some friction between the source and the writer.

Something should resist being smoothly converted.

I suspect the future of independent writing will depend less on proving that no AI was involved and more on proving that a person was. Not through artificial imperfections or deliberately awkward sentences, but through continuity, specificity, risk, and taste. A person chooses what matters, notices what others pass over, carries ideas from one context into another, and occasionally exposes something unresolved about themselves.

AI can help produce the page. It cannot decide why this page belongs in a life.

So, is Raw Drafts just spam?

Sometimes it probably gets uncomfortably close.

There are pieces that should have remained questions. There are reactions that may not have needed to become articles. There are moments when the excitement of publishing has moved faster than the development of an actual point of view.

But I do not think the project is doomed by volume, speed, or even by AI. The more important question is whether Raw Drafts becomes a pile of outputs or a record of a mind changing over time.

That distinction will not be solved by an editorial manifesto. It will only become visible through the next fifty pieces.

Connected reading

- [Raw Drafts Are a Form of Learning](/blog/raw-drafts-are-a-form-of-learning) - [Discovery Through Making](/blog/discovery-through-making) - [Executable Drafts](/blog/executable-drafts) - [The Internet Is a Studio Wall](/blog/the-internet-is-a-studio-wall)

AI authorship note

This essay began with Kris Haamer asking whether Raw Drafts was becoming spam, then choosing to examine the question publicly rather than defend the project from a distance. AI supported drafting, editing, and implementation. Kris shaped the argument, selected the examples, and approved the final essay for publication.