What the readers found in v2.0.0-rc2
39 findings from 5 readers, ordered by what they cost, not by who found them. Each points at a unit of the candidate, so anybody can check it. Every finding gets an answer; declined, with a reason, is one of them, and none is deleted.
1loses the commission6loses the sale8loses the reader9costs trust11worth fixing4minor
| Cost | Finding | Unit | What | Checked | Answer |
|---|---|---|---|---|---|
| loses the commission | catherine-1 Catherine Doyle | Ch. 22: The company itself: the UK, Portugal, Lithuania, the European Union | The book's stated objective is that the reader has started a company by the end, but the chapters that say how (the company itself, and Form your company) are placeholders that only list what is missing. A publisher cannot sign a book whose headline promise is a stub. | holds Placeholder confirmed: ch:22 says "The material is not written yet" and ch:37 says "in this candidate it is not written yet"; both hand over to notes for Dinis (the UK, Lithuania, Portugal and EU memos). A known gap, waiting on Dinis's memos; what the study adds is that it is the gap every reader names. | awaiting the editor |
| loses the sale | aisha-1 Aisha Bello | Ch. 37: Form your company | The chapter that should make the reader a registered company says it is not written, and chapter 22 says the same. The book's objective, to start a company by the last page, cannot be met in this candidate. This is the reason I would not buy now. | holds Placeholder confirmed: ch:22 says "The material is not written yet" and ch:37 says "in this candidate it is not written yet"; both hand over to notes for Dinis (the UK, Lithuania, Portugal and EU memos). A known gap, waiting on Dinis's memos; what the study adds is that it is the gap every reader names. | awaiting the editor |
| loses the sale | catherine-2 Catherine Doyle | Ch. 37: Form your company | Chapter 37 opens by promising a registered company with a bank account, then says the steps are not written. A reader who bought for this would feel misled. | holds Placeholder confirmed: ch:22 says "The material is not written yet" and ch:37 says "in this candidate it is not written yet"; both hand over to notes for Dinis (the UK, Lithuania, Portugal and EU memos). A known gap, waiting on Dinis's memos; what the study adds is that it is the gap every reader names. | awaiting the editor |
| loses the sale | ellie-1 Ellie Marsh | Ch. 37: Form your company | The book promises a registered company by the end, and the chapters that should deliver it (22 and 37) are placeholders. A buyer who came for that would not buy this version. | holds Placeholder confirmed: ch:22 says "The material is not written yet" and ch:37 says "in this candidate it is not written yet"; both hand over to notes for Dinis (the UK, Lithuania, Portugal and EU memos). A known gap, waiting on Dinis's memos; what the study adds is that it is the gap every reader names. | awaiting the editor |
| loses the sale | martin-1 Martin Kowalski | Ch. 22: The company itself: the UK, Portugal, Lithuania, the European Union | The title and chapter 1 promise that by the end I have started a company, but the chapters that say how to form one (22 and 37) are placeholders, and the book offers a note instead of steps. The sold book cannot make this promise as it stands. | holds Placeholder confirmed: ch:22 says "The material is not written yet" and ch:37 says "in this candidate it is not written yet"; both hand over to notes for Dinis (the UK, Lithuania, Portugal and EU memos). A known gap, waiting on Dinis's memos; what the study adds is that it is the gap every reader names. | awaiting the editor |
| loses the sale | nadia-1 Nadia Rahman | Ch. 22: The company itself: the UK, Portugal, Lithuania, the European Union | The book's stated objective is that the reader has started a company, yet the two chapters that say how to form one (22 and 37) are placeholders with two lines on Portugal and Lithuania and nothing on the UK. A reader who came for this does not buy, or puts it down here. | holds Placeholder confirmed: ch:22 says "The material is not written yet" and ch:37 says "in this candidate it is not written yet"; both hand over to notes for Dinis (the UK, Lithuania, Portugal and EU memos). A known gap, waiting on Dinis's memos; what the study adds is that it is the gap every reader names. | awaiting the editor |
| loses the sale | nadia-2 Nadia Rahman | Ch. 17: Be a founder: the expert founders | A 51-year-old with a salary, a mortgage and savings needs to know what leaving a job costs and how long she can last. Nothing in the book addresses personal runway, giving up a salary, or the regulatory duties of a business that sells to insurers. The access argument says starting is cheap but not how a person survives while doing it. | holds Searched the whole candidate: no mention of runway, salary, mortgage or savings, and nothing on leaving a job to found. The expert founder the frame is written for is the reader who would have to. | awaiting the editor |
| loses the reader | aisha-2 Aisha Bello | Ch. 39: Your first price and your first ten customers | The hands-on pricing chapter repeats chapter 19 almost sentence for sentence, down to the same RiskMandate ladder, the same free-is-not-free argument and the same gym study. A reader who read 19 meets it twice and stops. | holds Measured: 8-word runs shared between chapters, as a share of the smaller chapter: ch:13 and ch:30 319 (29%), ch:19 and ch:39 169 (15%), ch:7 and ch:32 113 (11%), ch:7 and ch:33 107 (10%), ch:16 and ch:35 56 (7%), ch:12 and ch:36 48 (4%). The hands-on part restates the argument chapters it follows. | awaiting the editor |
| loses the reader | aisha-3 Aisha Bello | Ch. 1: Gen ZX: who this book is for, and what it is for | The first chapter spends its length on 2018, the two generations, a founder-age study and the name, before saying plainly what the reader will do. A reader on a phone gets the promise in the fourth section. | holds Ch:1's sections are Two generations, Generation ZX, What it is for and The measure: the objective is in the third section, the company list in its fourth paragraph. | awaiting the editor |
| loses the reader | catherine-3 Catherine Doyle | Ch. 30: An agentic workflow from zero | The Contacts folder, five documents and six-line agent policy appear in chapter 13 and again in chapter 30, and the five-lens example in chapter 7 recurs in chapter 32 and the four pricing moves in chapter 19 recur in chapter 39. Repetition pads the length and tires the reader. | holds Measured: 8-word runs shared between chapters, as a share of the smaller chapter: ch:13 and ch:30 319 (29%), ch:19 and ch:39 169 (15%), ch:7 and ch:32 113 (11%), ch:7 and ch:33 107 (10%), ch:16 and ch:35 56 (7%), ch:12 and ch:36 48 (4%). The hands-on part restates the argument chapters it follows. | awaiting the editor |
| loses the reader | ellie-2 Ellie Marsh | Ch. 40: Ship a serverless product in a weekend | The weekend chapter, the one a beginner turns to for something to do, is written for developers and piles up unexplained terms. It is where a non-coder would put the book down. | holds Ch:40's core step is "a vault server of 83 lines, one Worker, one R2 bucket and one Durable Object per vault"; none of the four terms is defined in the chapter. | awaiting the editor |
| loses the reader | ellie-3 Ellie Marsh | Ch. 3: Everything is connected: graphs | The graphs chapter starts well with the detective board and then moves into code review and the author's command line tool, which tells a non-coder the book assumes she codes. | not checked | awaiting the editor |
| loses the reader | martin-2 Martin Kowalski | Ch. 30: An agentic workflow from zero | Several chapters say the same thing twice. Ch 13 and ch 30 are the same workflow, ch 19 and ch 39 the same four moves and the same RiskMandate example, ch 7 with ch 32 and ch 33, ch 12 with ch 36. A reader who notices stops trusting the structure and the length. | holds Measured: 8-word runs shared between chapters, as a share of the smaller chapter: ch:13 and ch:30 319 (29%), ch:19 and ch:39 169 (15%), ch:7 and ch:32 113 (11%), ch:7 and ch:33 107 (10%), ch:16 and ch:35 56 (7%), ch:12 and ch:36 48 (4%). The hands-on part restates the argument chapters it follows. | awaiting the editor |
| loses the reader | nadia-3 Nadia Rahman | Ch. 4: Where do you land? | Chapters 2 to 5 are about 8,500 words of wonder, graphs, Wardley maps and learning before the reader does anything. A reader who wants the first version of her tool is a long way from it, and the practical afternoon is in chapter 13 or later. | holds Measured: ch:2 to ch:5 are 8,751 words; the first hands-on afternoon is ch:13, about 26,000 words in. | awaiting the editor |
| loses the reader | nadia-4 Nadia Rahman | Ch. 40: Ship a serverless product in a weekend | The serverless and privacy chapters (23 to 27) and the weekend product (40) use terms that are never defined for a non-technical reader and are mostly about the author's own products. A non-coder cannot follow the Saturday steps. | holds As ellie-2 for ch:40; the terms are used in part 7 without a glossary. | awaiting the editor |
| costs trust | aisha-4 Aisha Bello | Ch. 20: Sell | Prices, programmes and examples keep being the author's own companies, and the chapter admits there is no paid run yet and the prices are hypotheses. It reads as selling the author's products, not as advice a student can act on. | partly Partly. Counted: sgit 38 mentions in 16 units, RiskMandate 26 in 15, SG/Send 12 in 5, "the newsroom" 112 in 36. RiskMandate is the worked example in ch:19, 20, 25 and 39 (three mentions each), sgit dominates ch:24 (12). There are outside examples too (Stripe's fee, the 2.7 million founder study, Nokia), so "always" is too strong; "the main worked example is usually the author's own" holds. | awaiting the editor |
| costs trust | aisha-5 Aisha Bello | Ch. 17: Be a founder: the expert founders | The central company example is a code review business built on graphs and the author's own tools, for a reader who has just asked job or company. It is hard to tell whether it is advice or a pitch, and it is too big for a first company. | not checked | awaiting the editor |
| costs trust | catherine-4 Catherine Doyle | Ch. 24: Privacy by design: SG/Send, sgit, and the value of constraints | Large parts of the middle and end are the author's own products (sgit, SG/Send, RiskMandate), used as the evidence for the argument. It is disclosed, but the book reads at times as a product tour, and the free-to-read online edition makes it look more like marketing. | partly Partly. Counted: sgit 38 mentions in 16 units, RiskMandate 26 in 15, SG/Send 12 in 5, "the newsroom" 112 in 36. RiskMandate is the worked example in ch:19, 20, 25 and 39 (three mentions each), sgit dominates ch:24 (12). There are outside examples too (Stripe's fee, the 2.7 million founder study, Nokia), so "always" is too strong; "the main worked example is usually the author's own" holds. | awaiting the editor |
| costs trust | catherine-5 Catherine Doyle | Ch. 41: How this book was made, and what you made | The co-writing with Claude and the method are disclosed in the last chapter, after 40 chapters in the first-person voice. The disclosure is honest but late, and the About note and chapter 41 explain the process twice. | holds Searched: the word Claude appears only in ch:41 ("Co-written with Claude, and credited"); the About note describes agents but does not name the model or say co-written. | awaiting the editor |
| costs trust | ellie-4 Ellie Marsh | Ch. 19: Price it, then give it away | The proof is nearly always the author's own products (RiskMandate, sgit, SG/Send), with no example from an ordinary small business. It reads as a book about his companies and the pricing example is hard for a reader to transfer. | partly Partly. Counted: sgit 38 mentions in 16 units, RiskMandate 26 in 15, SG/Send 12 in 5, "the newsroom" 112 in 36. RiskMandate is the worked example in ch:19, 20, 25 and 39 (three mentions each), sgit dominates ch:24 (12). There are outside examples too (Stripe's fee, the 2.7 million founder study, Nokia), so "always" is too strong; "the main worked example is usually the author's own" holds. | awaiting the editor |
| costs trust | martin-3 Martin Kowalski | Ch. 25: Controlling your agent: RiskMandate | The practical examples for security, privacy, pricing and selling (ch 19, 20, 24, 25, 39) are all the author's own products, sgit and RiskMandate. The stake is disclosed, which helps, but a book that meant to teach me would use someone else's example at least once. | partly Partly. Counted: sgit 38 mentions in 16 units, RiskMandate 26 in 15, SG/Send 12 in 5, "the newsroom" 112 in 36. RiskMandate is the worked example in ch:19, 20, 25 and 39 (three mentions each), sgit dominates ch:24 (12). There are outside examples too (Stripe's fee, the 2.7 million founder study, Nokia), so "always" is too strong; "the main worked example is usually the author's own" holds. | awaiting the editor |
| costs trust | martin-4 Martin Kowalski | Ch. 21: Fund it, or do not | The funding chapter argues for a sale of existing shares plus a lend-back, with a tax worked example, from a hypothesis that nobody has tested. The counter-evidence is fairly stated, but the chapter's confident close ('Do not get signed before you have the album') goes beyond its evidence. | holds The worked example in ch:21 is labelled illustrative and the counter-evidence is stated; the closing line "Do not get signed before you have the album" appears twice in the chapter. A judgement on tone, kept as the reader's. | awaiting the editor |
| costs trust | martin-5 Martin Kowalski | Ch. 17: Be a founder: the expert founders | The case that expert founders win rests on a study of 2.7 million founders and three examples, two of them people the author helps and the third himself; it closes on a business plan for his own tools. It is sourced, but it is the book's thesis chapter. | not checked | awaiting the editor |
| costs trust | nadia-5 Nadia Rahman | Ch. 19: Price it, then give it away | Nearly every worked example in the founder chapters is the author's own product (RiskMandate, SG/Send, the newsroom meter). It is honest about that, but the reader cannot tell how the method would work for an insurance tool or a clinic, and the book can seem more about the author than about her. | partly Partly. Counted: sgit 38 mentions in 16 units, RiskMandate 26 in 15, SG/Send 12 in 5, "the newsroom" 112 in 36. RiskMandate is the worked example in ch:19, 20, 25 and 39 (three mentions each), sgit dominates ch:24 (12). There are outside examples too (Stripe's fee, the 2.7 million founder study, Nokia), so "always" is too strong; "the main worked example is usually the author's own" holds. | awaiting the editor |
| worth fixing | aisha-6 Aisha Bello | Ch. 30: An agentic workflow from zero | The actionable hands-on workflows with a promised result are in the second to last part, 30 to 40. A reader who wants to act today should meet that style first, not after 29 chapters of argument. | not checked | awaiting the editor |
| worth fixing | aisha-7 Aisha Bello | Ch. 4: Where do you land? | Wardley maps, explorer, villager and town planner are taught before the reader has a reason to care, and the chapter tells you to find your centre of gravity first. It is the book's frame, but it is jargon for a reader who came with a job question. | not checked | awaiting the editor |
| worth fixing | catherine-6 Catherine Doyle | Ch. 2: You are lucky to be here | Notes addressed to Dinis are embedded throughout the chapters and several chapters admit to unnamed sources or unread articles, which shows how the draft was made but would have to be resolved before print. | holds Holds for the candidate: 47 notes marked For Dinis in 36 of 52 units. Expected in a release candidate (decision 009) and not in a sold book; the readers say so themselves. What it adds: the notes interrupt the reading for every reader, so a reader's edition without them is worth having before the next study. | awaiting the editor |
| worth fixing | catherine-7 Catherine Doyle | Ch. 38: The credits and programmes, 2026 | Dated tables of programmes, credits and prices (chapters 23 and 38) and many statistics (chapter 10) have a shelf life of months. They belong on a web page or in an annual update, not in the printed text. | partly Holds on the facts (ch:38 is dated 2026), but the dating is a decision (013): the hands-on part is in the candidate, dated and renewed each year. The publisher's point is about format: on a web page, not in print. For Dinis. | awaiting the editor |
| worth fixing | ellie-5 Ellie Marsh | Ch. 30: An agentic workflow from zero | The Contacts folder workflow, the six rules and the five exercises appear almost word for word in both ch 13 and ch 30, and the pricing argument in ch 19 returns in ch 39. The repetition makes a 75,000 word book feel padded. | holds Measured: 8-word runs shared between chapters, as a share of the smaller chapter: ch:13 and ch:30 319 (29%), ch:19 and ch:39 169 (15%), ch:7 and ch:32 113 (11%), ch:7 and ch:33 107 (10%), ch:16 and ch:35 56 (7%), ch:12 and ch:36 48 (4%). The hands-on part restates the argument chapters it follows. | awaiting the editor |
| worth fixing | ellie-6 Ellie Marsh | Ch. 1: Gen ZX: who this book is for, and what it is for | Notes to the author and open decisions sit inside the reader's text, and the 'about' unit describes the review process before the book starts. A buyer should not meet them. | holds Holds for the candidate: 47 notes marked For Dinis in 36 of 52 units. Expected in a release candidate (decision 009) and not in a sold book; the readers say so themselves. What it adds: the notes interrupt the reading for every reader, so a reader's edition without them is worth having before the next study. | awaiting the editor |
| worth fixing | ellie-7 Ellie Marsh | Ch. 5: Connect the dots backwards | Chapters 2 and 5 spend their length on wonder, optimism and learning theory before any practical step; the reader reaches something to do only in chapter 8 or 13. | not checked | awaiting the editor |
| worth fixing | martin-6 Martin Kowalski | Ch. 41: How this book was made, and what you made | The last chapter and the notes marked For Dinis make the manuscript read as a working document. In the book that would be sold, the notes must go and the process (roles, board, candidates) shrinks to a page, or the reader ends on the machine and not the argument. | holds Holds for the candidate: 47 notes marked For Dinis in 36 of 52 units. Expected in a release candidate (decision 009) and not in a sold book; the readers say so themselves. What it adds: the notes interrupt the reading for every reader, so a reader's edition without them is worth having before the next study. | awaiting the editor |
| worth fixing | nadia-6 Nadia Rahman | Ch. 30: An agentic workflow from zero | The Contacts folder, five documents and six rules appear in full in chapter 13 and again in chapter 30. Chapters 7 and 33 repeat the tool-call and compression explanation too. Cut one of each pair, and put the practical version first. | holds Measured: 8-word runs shared between chapters, as a share of the smaller chapter: ch:13 and ch:30 319 (29%), ch:19 and ch:39 169 (15%), ch:7 and ch:32 113 (11%), ch:7 and ch:33 107 (10%), ch:16 and ch:35 56 (7%), ch:12 and ch:36 48 (4%). The hands-on part restates the argument chapters it follows. | awaiting the editor |
| worth fixing | nadia-7 Nadia Rahman | Ch. 2: You are lucky to be here | Notes addressed to the author and admissions such as an unnamed podcast and a paywalled article read as scaffolding in a book that would be sold. In a printed edition these must go or be resolved, and the opening claim should not rest on an essay the author did not read. | holds Holds: ch:2 says of the Business Insider essay it opens on, "I could not get past the paywall, so I cannot tell you what it argued." Honest, and still the first argument after the introduction rests on it. | awaiting the editor |
| worth fixing | nadia-8 Nadia Rahman | Ch. 21: Fund it, or do not | The funding chapter is aimed at a founder with a £10 million profitable company, with share-sale and tax mechanics. The reader at the start, with no revenue, has no use for it yet. A short section on what to do before there is any revenue is missing. | partly The worked example is a UK company worth £10 million; the chapter's argument (profitable before raising) does speak to the pre-revenue founder, but there is no section on what to do before revenue. | awaiting the editor |
| minor | aisha-8 Aisha Bello | Ch. 2: You are lucky to be here | Telling the reader to read on paper with a pen sits oddly in a book people may read on a phone, and the chapter spends many words on history. | holds Ch:2's How to read this book opens "Read it on paper, with a pen." | awaiting the editor |
| minor | catherine-8 Catherine Doyle | Ch. 34: Images, infographics and storyboards, drawn by code | Chapter 34 on storyboards drawn by code documents the author's own site and does not help a reader start a company. | holds Ch:34 mentions storyboards 12 times; it is a workflow for drawing images by code, of the kind this site uses. | awaiting the editor |
| minor | ellie-8 Ellie Marsh | Ch. 6: What generative AI is | The 'Where I was corrected' list is a good honest habit but arrives as a wall of technical terms for a reader who only wanted to know what generative AI is. | not checked | awaiting the editor |
| minor | martin-7 Martin Kowalski | Ch. 2: You are lucky to be here | The opening argument against fear rests on a Business Insider essay the author could not read past the paywall. It is honest about that, but the first chapter after the introduction is mood, and a reviewer notes it. | holds Same passage as nadia-7: the paywall admission is in ch:2. | awaiting the editor |