# Founder review: the altitude ladder experiment

**Date:** 22 August 2026 · **Source:** voice memo, transcribed by Otter.ai, reproduced
verbatim below (per the house rule: the founder's voice is source material and is not
edited) · **Status:** Review **r004** on [the reviews register](../reviews/index.html),
normalised into six items with the agent's comments and proposals at
[/reviews/r004.html](../reviews/r004.html). Unlike r001 to r003, this review's central
item was **executed rather than proposed**, because the memo asks for an experiment and
an experiment that is only described has not been run. It adds a new artefact beside the
book at [/altitudes/](../altitudes/index.html); no book content was changed.

**What it says, in one paragraph:** build five or seven versions of the book at
different altitudes, where the current book is the bottom level. The top is a paragraph
or a few paragraphs, an executive summary, and each of its phrases hyperlinks down into
the next level, which does the same again, all the way down. It has to be built
**bottom-up**, because each layer is a compression of the layer below that captures its
core ideas. Do it as an experiment rather than a commitment, and not necessarily for the
whole book: maybe get the top three levels right and take a couple of sections all the
way down, to see how it plays out. The number of pages should fall at every step. Two
concepts are in play and must stay apart: **the graph** (the facts, the evidence, the
structure, the ideas being conveyed) and **the projection** (the text itself and its
links). Each layer needs its graph; not every layer needs its projection. The structure
is tree-like as concepts merge going up, though the relationships can be many-to-many.
This also opens up translation, cultural adaptation, and versions aimed at different
stakeholders: technical, business, graph-literate. And the interesting part: once you
can see all the facts, observations and hypotheses a paragraph or a section is carrying,
you will find repeated content and contradictions between them, which then need a way of
being visualised.

---

## Verbatim transcript

Dinis Cruz 0:00

So, continuing the book, here's an experiment. Can we try to create five or seven
versions of the book, which are all at different altitude? So, come here by. this is
that let's say that the current book that we have is level seven or level five, right?
So what we need is a graph-like structure where it's kind of like this sort of idea of
that summary, that hyperlinked set of evidences, where we have the top level, which is
almost maybe a paragraph or a set of paragraphs that basically describe almost like the
executive summary, right, of the book, right, and then that hyperlinks. If you think
about visually, it means that each of the phrases that we have, and we actually have a
vote that did this very well, each of the phrases that it has hyperlinks to the next
level, and then hyper, and then that has the same thing, and then each of the areas
hyperlinks to the next level, to the next level, to the next level. So, so we can
approach this from the top down and from the bottom up, and well, actually needs to be
from the bottom up, right? But the logic is that each layer is a compression of the
previous layer that captures the core ideas of this, and and this is also very
interesting when we then start looking at translations and adapt them to different
cultures, or even adapting adapting to different stakeholders, right, or different
audiences-a more technical audience, a more business-like audience, a more graph
knowledge audience-is can we create? And let's maybe do this as a revision, as an
experiment. Can we create this level of abstractions, so that we can kind of see how it
works, and maybe don't don't do for the whole book, right? Maybe do it for a couple of
sections, so we can see how this play out, and so and maybe you can, yeah, see see if
works out, or see see if it's better to have the top three levels correct, and then
maybe just for a couple you go all the way down, or if you can do the whole thing right.
But the logic, and in a way, if you think about it, the number of content and pages
should be going down and down and down, but almost abstraction. And I think, and here's
also the other interesting thing to do here is whether we have a projection or not. So I
think what we might want to do is, and this this is important, is because if you think
about it, there's already two concepts here. There's the concept of the graph, which has
all the facts, the evidence, the structure, the points, the ideas, what you're trying to
convey, and then there's the execution of that, which is the text itself, right? And
then the evidence and all the links. So each of these layers needs to have those graphs,
and then needs to have the projection. That's what maybe don't you can you don't need to
do the projection for all of them, because what it means, it means that the text kept
getting smaller and smaller and smaller, but you need to be base it on specific graphs
and specific concepts, and there should be a tree-like structure of this as you go up
and as the concepts merge, and and they of course they could be many-to-many
relationships. In fact, it's interesting because I think with this we actually will find
some repeated content and some contradictions in the content, right? Which would be
quite interesting. Which is basically once you look at all the facts, observations,
hypotheses, then a particular paragraph or particular section is doing, then that we
should find contradictions with others, and yeah, and then find a way to visualise

Transcribed by https://otter.ai