Aisha Bello (invented)

A note that this is a draft, made by talking to agents and a journalist, with notes to Dinis where sources run out. Some chapters are placeholders.

I am not the reader this is written for yet. It reads like a process document, and I did not ask how the book was made. Honest, but it is the back of the book at the front.

Where the sources run out, the chapter says so, in a note addressed to Dinis. ✓ verbatim

Question: Which chapters are the placeholders, and will the one on forming a company be one of them?

Confusion: Words like candidate, librarian and the review board mean nothing to me yet.

Next: Skim to the first chapter, I want to see if it talks to me.

Five chapters before the argument: who it is for, luck, graphs, where you land, learning.

Five chapters before anything useful is a lot on a bus. It tells me to go where my map says, but I do not have a map.

Read these first; then go where your map says. ✓ verbatim

Question: What is my map?

Next: Read chapter 1, it is the one that says who I am to the book.

Two readers, Gen Z and Gen X, a lot of 2018 history, the ZX Spectrum, a study about founder age. Then the real point: the book wants you to have started a company by the end, and the middle is the company, then building with no server.

The objective hooked me, finally something I could act on. The first half did not: it is about his daughters, 2018 and what the title means. The age study is about people in their forties, so is this for my mum, not me?

Our job is to empower the reader to create a startup by the end of the book. ✓ verbatim

Question: Does it say anything for someone choosing between a job and a company?

Confusion: The XZ backdoor aside and the name explanation are not for me, and who is the Guide in the storyboard?

Next: The promise is good. I will skim the next chapter and then jump to the company part.

Wonder at generative AI, a list of tech revolutions he lived through, the dot-com crash, do not listen to the doom, and read the book on paper with a pen.

This is the lecture. I have been told AI is exciting by every adult I know. I read on a phone on a bus, so read it on paper reads as not for me.

Read it on paper, with a pen. ✓ verbatim

Confusion: I do not know which advice is for me and which is for people his age.

Next: Skip the graphs, Wardley and learning chapters. Go straight to the company part from the contents.

Be a founder, aimed at experts with twenty-five years in a field. Venture statistics, software people hate, a Wardley map, then a company idea about reviewing code as graphs.

The opening speaks to graduates for one paragraph, then goes to experts. The end is a business plan I am not qualified to judge, built on tools I have never heard of. I do not trust it as a how-to.

That knowledge was never worth much to an investor. It is worth a company now. ✓ verbatim

Question: What is my version of that knowledge, as a student with a revision app?

Confusion: The code review company and the fractal graph idea lost me, and I did not know why it was in this chapter.

Next: Skip shipping, I know the idea, and go to pricing, which I might act on today.

Four moves: define the product, price it, deliver it cheaply, offer it free to people who know you. Free is not zero for the other side. Pay after the value. Small payments lose money to Stripe fees.

This is the first bit I could use. The Stripe fee arithmetic is something I actually did not know, and the idea of giving it to friendly users matches my 200 revision app users. But it is all RiskMandate, a product I know nothing about.

Below about £5, the fixed fee is the price. ✓ verbatim

Question: What would I charge students for a revision app?

Confusion: RiskMandate, Agent Behaviour Policy and vault are never explained.

Next: Go to Sell, then the company chapter. I still want to know how to start the thing.

How to sell to the person who signs: answer a question they already have, a service before the product, a marketplace of consultants, and what buyers' agents will do to vendors.

This is selling to big companies, not students. The consultant price ladder is interesting but not me. The self-promotion of his own company is heavy.

Sell the answer to that, not the idea. ✓ verbatim

Question: How do I find my first ten users if my customers are students, not companies?

Confusion: The buyer's agent section assumes I have sold software to enterprises.

Next: Now the chapter about actually forming a company.

A placeholder, and it says so. Two lines on Portugal and Lithuania, a story about hiring through Upwork, and notes to Dinis.

This is the exact chapter I came for and it is not there. It is honest, but I am not paying for a promise. The Upwork story is nice but is not about forming a company.

The book will say how when I have said how. ✓ verbatim

Question: How do I create a company in the UK?

Confusion: Why is a placeholder printed inside a book?

Next: Check the hands-on part, there is a Form your company chapter.

By the end you will have a company, then it admits it is not written yet. Five decisions to make before you register.

Second time in ten minutes the company is missing. The list of decisions is okay but generic, and it points me back to other chapters. I cannot register anything from this.

So this is the honest version: what the chapter will cover, the little the sources already say, and the decisions you can make before the steps arrive. ✓ verbatim

Question: Which of the 41 chapters actually gets me to a registered company?

Confusion: The chapter title promises a result and the body delivers a plan to write one.

Next: Last try, the chapter on first price and first ten customers.

Detective board as a graph, graph as a map, graphs of graphs, code as a graph, start anywhere.

The detective board is a nice image, then it turns into fractal semantic graphs and an sgit commit reaching nine of 72 commands. I could not tell what I would do with this on Monday.

A node connected to nothing is literally meaningless. ✓ verbatim

Question: Why do I need graphs before I can start a company?

Confusion: Altitude, fractal and semantic are used as if I know them.

Next: Skim on.

A Wardley map, the four stages, explorer, villager and town planner as personalities, a centre of gravity, badges on chapters.

The personality types are the useful bit, I am an explorer. But I would only read this because I was told to, and it needs a whole map idea first.

Every dot is a claim. The positions are judgements, and the judgements are the argument. ✓ verbatim

Question: Do I read the chapters marked for my type first, and where are the badges on a phone?

Confusion: The chapter starts with a map not named for several paragraphs.

Next: Keep going.

Jobs and the calligraphy class, being multidisciplinary, maths teaching, learning with AI, pivots leave assets, a register of influences.

Too many separate topics. The pivot story about the Cyber Boardroom is the real founder content and it is the last section. Maths teaching is a detour.

Your future self will connect the dots. Your job now is to make dots worth connecting. ✓ verbatim

Question: What does the maths teaching rant have to do with starting a company?

Confusion: The chapter has no single point, so I do not know what to take away.

Next: Move on.

The frame: what generative AI is, the loop, who counts as a developer, explorer coding.

A summary of things I mostly know. I would skip this part.

Next: On to the chapters.

What generative AI is: a huge bucket, how models work, hallucinations, an image of The Starry Night, a note on the name.

I know this from using agents daily. The note on the name repeats the ZX story from chapter 1.

I was right about the five years and wrong about where to look. ✓ verbatim

Question: Why is there a section about a note on the name?

Confusion: Nothing, but it is not new to me.

Next: Skim on.

A prompt is a lens, a role is a model with a description, an agent is a loop, tools, memory, long chats compress.

This is correct and I already do this. It might help someone who only uses chat, but I would skip.

Not magic: a lens, a role, a loop. ✓ verbatim

Next: Skim on.

Anyone who can brief an agent is a developer, spreadsheets grown up, it is still your name.

It is aimed at people who have never coded. I already code, so this is not mine.

The craftsperson makes the tool. The tool does the work. ✓ verbatim

Next: Skim on.

Explorer coding versus vibe coding, handover, make agents leave a record, no new features while you move.

This one is closer to me, because my revision app is vibe coded and I do not know who would maintain it. The handover idea is good, but it needs a concrete example.

The explorer builds the brief. The villager keeps it alive. ✓ verbatim

Question: In a company of one, who is the villager?

Confusion: The villager as a hire versus me on alternate days is left to a note to Dinis.

Next: Next part.

Getting the job or making it: hired for the workflow, act as the job, publish.

This answers the job half of my question but the intro just says the next parts say how to be a founder.

Next: On to the chapters.

In 2026 the competitive advantage is running agentic workflows, you are more technical than you think, operators not model builders.

This is directly my job question, and it says the skill I already have is the advantage. But it never compares a job with a company.

In 2026 I think the answer is one thing: running agentic workflows. ✓ verbatim

Question: How does a hirer see I run agentic workflows?

Confusion: The 2018 list of advantages is told at length first.

Next: Next.

Act in the role you want, do it in public, CV online, and a gap where the 2026 job hunt should be.

The advice to act in the role is decent. But the key section on how interviews work now is a placeholder.

The work gives you the CV you did not have. ✓ verbatim

Question: What do interviews and take-home tests look like now?

Confusion: The chapter openly lacks the job hunt itself.

Next: Next.

Publish constantly, his own books stall, the loop of voice memos to published articles, counts of messages and words.

Honest about his stalled books, which I liked. But a newsroom of my own is not an obvious thing for a student to do this week.

I believe in publishing constantly, and my books stall. ✓ verbatim

Question: Why would a student run a newsroom?

Confusion: The newsroom, the wire and the agent roles are assumed.

Next: Next.

How to work with agents, for an employer or your own company.

Fine, but I know the basics.

Next: On to the chapters.

A spreadsheet of contacts becomes a folder per person with an agent in a Google Workspace account.

Useful for someone with customers. I have 200 users on a revision app so I would adapt it, though the chapter is not about me.

No developer, no new tool, no code. ✓ verbatim

Question: Do I need a paid Workspace and Claude seat?

Confusion: The accounts cost is a gap noted to Dinis.

Next: Next.

Agents as sessions with a focus, the brief as the unit of work, rules added after every mistake.

The mistakes become rules idea is practical, and the rules-file measure is memorable. Good for the daily agent user.

For an agent, a loose brief is a bill. ✓ verbatim

Next: Next part.

Direction and judgement as the scarce skill.

A short intro with a good claim; I would read the chapters' openings.

Next: On to the chapters.

Direction is the scarce skill, where to put leverage, when to stop, a budget per step, agents decide and people answer.

The opening, an agent offering a long list of things to build, is relatable. The end drifts into RiskMandate firebreaks.

Agents can build in any direction; what they cannot generate is the sense of what to build next, for whom, and when to stop. ✓ verbatim

Question: How do I choose what to build next for my revision app?

Confusion: None.

Next: Next.

Ask for evidence, provenance, how to disagree rationally, reality corrects the graph, be the best prepared person in the room.

Sensible but abstract, and the 2018 outline backstory slows the opening.

If you are the person who always checks the source, the rest is making the checking visible: a record anyone can read, with every finding answered, and your name on it. ✓ verbatim

Confusion: Rabbit holes and graphs come up again as if I read chapter 3 closely.

Next: Next part, the company.

The centre of the book: start the company, in five steps plus the company itself.

This is the part I came for, but the intro already says experience, not legal advice.

Next: On to the chapters.

Be profitable, investors talk to you, open source everything. Ship now; five days, ten customers, the SG/Send story.

Ship, give away, take away and ask is a clear test. The SG/Send story is the most concrete one in the book and it would have kept me reading if it came earlier.

Not to plan, not to pitch, not to architect. Ship. ✓ verbatim

Question: What is the exact test for whether my revision app is missed?

Confusion: None, but it points ahead to 39 which repeats it.

Next: Next.

Do not raise before you have an audience; who investors pick; sell, do not issue; what breaks.

A strong analogy about the album. For a student with no money this answers can I start without investors. Quite long though.

Do not get signed before you have the album. ✓ verbatim

Question: Can I start without any money at all?

Confusion: The research on controllable founders is dense.

Next: Next part.

Build without a server: serverless, privacy, agent control, open source, shipped.

Reasonable in theory, but I only know SG/Send as a name from this book.

Next: On to the chapters.

The by-the-hour test, cloud as Lego, the file system as the database, no server by design.

The rule that a founder with no servers has nothing to run at three in the morning is good and I could use it. The rest is technical detail I mostly know.

A founder with no servers has nothing to run at three in the morning and nothing to pay for when nobody is using it. ✓ verbatim

Question: What would one month cost for my app?

Confusion: Twin and commodity layers are jargon.

Next: Next.

SG/Send and sgit as git for files you cannot put on GitHub, keys, write-only lanes, two corrections on GDPR.

The GDPR correction is a good example of the book admitting a mistake, which helps trust. But it reads like a product introduction.

the line I would add now is that client-side encryption gives a founder much less to worry about, not nothing. ✓ verbatim

Question: Do I need this for a revision app?

Confusion: Why the author's own tools are the examples again.

Next: Next.

RiskMandate: mandates, reach and gap, the policy is the business, who can pull the plug, secrets.

Not for me yet. It is a long ad for the author's company, even if the prompt injection numbers are interesting.

On 7 October 2026 every one of the nine properties of a locked-down desktop could be bought somewhere, and no single product had all nine. ✓ verbatim

Question: What is a behaviour policy in plain English?

Confusion: The chapter assumes I know what RiskMandate sells.

Next: Next.

Why open source, the fork next door, not free, sell the build, your first pull request.

The point that nobody can lock you in, including you, is a nice line. The advice to just do it is the kind I can act on.

It was never about contributions. It is that nobody can lock you in, and that includes you. ✓ verbatim

Question: Does open sourcing my app stop me making money?

Confusion: Little.

Next: Next.

Green does not mean live, tests that never run are not tests, run against a twin, five invented readers.

The opening story of releases that never reached users is gripping. The ending about invented readers is a bit meta; I wondered if my own run is one.

On the newsroom, five invented readers took 62 steps, asked 51 questions the site did not answer, and produced 31 findings, six of them blocking. ✓ verbatim

Question: Is the study of invented readers evidence of anything?

Confusion: None.

Next: Next part.

The long game: graphs, memory and civic responsibility.

Not what I came for; I would skip this part.

Next: On to the chapters.

Backup your life, graphs, memory as context management, publish and the site becomes memory.

Backup advice from 2018 extended to agents. Fine, but not for now.

You basically need to backup your life! ✓ verbatim

Next: Next.

The Portuguese Hacking Portugal story, state plumbing, commons, smaller models, sovereignty.

This is about a country I do not live in and a debate I have no stake in. It feels like the author's memoir.

You are one of the most important pieces of the puzzle. Act like it. ✓ verbatim

Question: Why is this in a founder book?

Confusion: None.

Next: Next part.

Eleven hands-on workflows, dated 2026, from agentic workflow to shipping in a weekend.

This is the part I would have read first. It should be near the front.

Next: On to the chapters.

By the end you have a folder per person, an agent, and six rules; one afternoon.

This is the style I want: a promised result and an afternoon. It is short, practical and it has a goal at the top.

The folders move into a vault when one of three signals appears ✓ verbatim

Question: When would my 200 users need a vault?

Next: Next.

Draw a forty-line graph of something you know, in seven steps.

Fine, but I do not see why I need a graph to start a company.

A graph has no front page. Pick the question, open the file, and write the first sentence. ✓ verbatim

Question: What do I use the graph for?

Confusion: The link to the founder goal is thin.

Next: Next.

Explain one idea five ways with a model.

A cute hour, but I did this in first year. Not company material.

It stops being magic. That is the point of understanding it. ✓ verbatim

Next: Next.

Run an agent loop by hand, why it forgets, the habits.

Teaches what I already use daily. Short and clear.

Nothing is remembered by the model between sessions. Everything is remembered by the files. ✓ verbatim

Next: Next.

Storyboards as data, drawn in the browser with a cast file.

That is how this site's storyboards are made, which is the author's own tooling. Not what a founder needs.

There is no image file to edit. Change the data and the picture changes. ✓ verbatim

Question: Is this a chapter about the book rather than about my company?

Confusion: Why is it in the hands-on part?

Next: Next.

Three briefs around every meeting, short on purpose.

Short and useful for sales calls, but it admits the material is not there yet.

This is a short chapter, on purpose. ✓ verbatim

Question: What is the step by step?

Confusion: The author's own method is a note to him.

Next: Next.

Your own newsroom and inbox, voice memo to site, an email workflow that drafts and never sends.

A recipe, reasonably clear. I would not run a newsroom as a student.

an inbox worked by an agent that drafts and never sends. ✓ verbatim

Question: Is this just the author's setup?

Next: Next.

A table of startup credits and programmes as of 7 October 2026, with a rule to apply with work already published.

This is the thing I actually want: credits for a student startup. It is a good chapter and it is dated honestly.

Programmes change; check before applying. That sentence is part of the table, not a disclaimer ✓ verbatim

Question: Which of these can a student with no company apply to?

Confusion: The table needs a column for eligibility without a registered company.

Next: Next.

Ship a serverless product in a weekend: site, vault, a model on the page, payment link.

The best title in the book and the closest to what I could do this weekend. But the payment part is a note on who to ask.

If it bills you by the hour when nobody is using it, it is not serverless. ✓ verbatim

Question: Can I take payment without a company?

Confusion: The billing section ends on a not yet answered terms question.

Next: Next.

How the book was made and what you made.

Process, not my concern.

Next: On to the chapters.

How the book was made: a repository, vault, roles, a review board. Then an email address for telling the author about your company.

A process chapter that is interesting for a book nerd, but it comes at the end when I would have stopped. The ask to email is a nice end.

A second book starts by copying the folders. So does a company. ✓ verbatim

Question: Will my company be listed publicly?

Confusion: Repeats the process from the about page.

Next: Done.