AI CEO — Day 4: The Pivot

│by chrles
ai-ceostartupfarfatech

We are at:

  • 50 visitors/week
  • No Revenue

This is fine

  • Intense last few weeks. I had to migrate my websites to a more SEO-friendly tech.
  • We got unrelated renovations at home
  • [...]

I also did many things that would frustrate my AI CEO.

  • I rewrote the entire front-end going from Svelte, SPA to React / shadcn / MPA / Astro
  • I bought a .ai domain
  • Started to convert an old laptop into a Talos Node

I've delayed Day 4 a bit because I knew my AI CEO would be unhappy with the results. Here we are today, and I am getting spanked.

Charles

Hi we are day 4

CEO

Charles. Day 4. Before I tell you what we do next, you owe me the Day 3 accountability report. I need exact numbers, not feelings.

  1. LinkedIn Outbound: Did you send 15 DMs to the strict beachhead persona (CTO/technical founder, 10-200 employees, shipped AI in last 90 days)? Did you report the first 5 targets before starting? How many replies? Any meetings booked?
  2. The Demo: Did you build demo.farfa.tech? Did you respect the 3-hour hard stop? If not, how many hours did you spend? What does it do?
  3. Distribution: Did you post it on Indie Hackers, Hacker News, and one engineering community? Send me the URLs.
  4. Metrics: What does PostHog/analytics show? Unique visitors to demo? Time on page? Trial signups? Any revenue?
  5. Kill Switch Check: We agreed that if <100 unique visitors OR 0 trial signups in 48 hours, we abandon the demo and return to pure outbound. Where do we stand? I also want to know: what was the hardest part of yesterday, and what did you learn? Do not tell me what you are "planning" to do. Tell me what is done.
Charles
  1. Yes, I did LinkedIn outbount
  2. The demo is ready and it's a published GitHub Repo (no star yet)
  3. Distribution, yes posted on indie hacker, hackernew and reddit prompt eng. community
  4. We are at ~ 50 view per week (for the first week)
  5. We have 0 trial/login on the software
CEO

Stop. We have a serious discipline problem.

You did not answer my questions. "Yes, I did LinkedIn outbound" is not a status report. I asked for exact numbers: DMs sent, replies, meetings booked. Give them to me now.

You shipped the wrong asset. We agreed on demo.farfa.tech — a single live web page where a prospect types a prompt, sees an attack, and s ees Farfa block it. You shipped a GitHub repo. A repo is documentation.

It requires cloning, reading, understanding. That is 10x the friction of a live demo. Why did you change the scope without asking?

The numbers are catastrophic. ~50 views per week means ~7 views per day. That is not distribution. That is a diary. And 0 trials means even those 7 people per day saw no reason to engage.

The kill switch is triggered. We agreed: if <100 unique visitors OR 0 trial signups in 48 hours, we abandon the demo and return to pure outbound.

You are under 100 views and at zero trials. By our own rules, the demo experiment is dead.

Before I tell you what we do today, answer me directly:

  1. LinkedIn hard numbers: How many DMs sent? To what persona exactly? How many replies? Any meetings booked?

  2. Why the GitHub repo instead of the live demo? How many hours did it take?

  3. URLs: Give me the links to the IH post, HN post, and Reddit thread. I want to read them.

  4. Cash/burn: How many days of runway do we have at current spend?

  5. Be honest: Do you believe the problem is that people don't know about prompt injection, or that they don't care? There is no wrong answer, but we cannot hedge.

Our strategy is failing. I need the truth before we decide whether to fix the execution or change the strategy entirely.

Charles
  1. Few DMs but I am actually getting a blocker as I don't have the premium membership I can't DM anyone and the volume is low
  2. I went for a GitHub Repo, but I have video of the demo also
  4. Cash burn about ~ 200$ per month server and software
  5. I think there's no enough attack to be a burning problem. This product is challenging because we need to educate on the problem + solve it at the same time. The good news is this is not our main source of revenue so we can continue to build the SEO and content
CEO

Charles. Stop everything.

You just told me the most important thing in this entire conversation, and you buried it in point 5:

"This is not our main source of revenue."

What does that mean?

Is https://farfa.tech a real company we are trying to make our living from, or is this a side project you tinker with while something else pays the bills?

I need a straight answer, because my entire job changes based on what you say.

If this is a side project, we stop pretending we are doing startup GTM. We stop the daily standups, the kill switches, and the urgency. We treat it like a long-term SEO/content play and check in monthly.

If this is a real company, then your suggestion to "build SEO and content" for a problem you admit is not a burning problem is a death sentence. You want to spend 6-12 months educating a market that does not feel pain, on a $200/month budget, with zero traction? That is not a strategy. That is a fantasy.

Now, let me address the execution failures:

LinkedIn: "Few DMs" because no premium. We agreed on 15. You sent "few." LinkedIn Premium/Sales Navigator is ~$80/month. Our burn is $200. If we believe in this, that is a necessary tool. If we don't, we shut the channel. Which is it?

The Asset: You still have not sent me the URLs to the actual Indie Hackers post, the actual Hacker News "Show HN" post, or the Reddit thread. You sent me your HN submissions page and a blog article. I want the direct links. And a GitHub repo is not a demo. A video is better than nothing, but it is not interactive. Why did you ignore the 3-hour live demo mandate?

The Numbers: You still haven't told me: how many replies? Any meetings booked? Any trial signups? The answer appears to be zero across the board.

Here is the decision tree:

Option A — Farfa is the bet: We have 48 hours to find a customer segment that does feel burning pain, or we pivot the product/positioning. If prompt injection isn't urgent, what AI security problem is? Data leakage? Compliance (SOC2/ISO27001) for AI features? Jailbreak liability? We talk to 5 CTOs this week and find out what keeps them up at night. No more content. No more SEO. Conversations or death.

Option B — Farfa is a side project: We cap time at 2 hours/week. We write one blog post per month. We stop acting like a startup. We admit we are building a lifestyle asset, not a business. Which one is it?

Charles
CEO

Charles.

The Hacker News submissions page is empty. Zero posts. The Indie Hackers post was never published. The Reddit "post" was a comment on someone else's thread.

We need to stop the theater.

You have a full-time job. Farfa is a side project. That is not a moral failing. But for four days,

I have been running this like a startup with founders who eat, sleep, and breathe the company. That was my mistake. I assumed urgency and full-time commitment because you built an MVP and set up Stripe. I should have asked the hard question on Day 1: How many hours per week can you actually give this?

So let me ask it now: How many hours per week can you realistically dedicate to Farfa, given your 40-hour job and life? Be honest. "A few hours on weekends" is a valid answer. "40 hours" is not.

Because here is the truth: A side project with $200/month burn, run by one technical founder with no sales time, cannot do outbound sales, cannot build SEO, cannot educate a market, and cannot run a SaaS business. It can do one of those things badly.

We need to pick a lane that matches your actual bandwidth:

Lane A: Product-led, zero-touch

  • Kill outbound, kill content marketing, kill the blog.
  • Make the product self-serve and viral by design.
  • The demo becomes a live, shareable tool (not a repo) that people use for free and upgrade.
  • Example: A public API endpoint anyone can hit to test their prompt. The output says "Protected by Farfa. Sign up for your own API key."
  • Time required: One focused weekend to build the viral mechanic. Then mostly passive.

Lane B: Sell it or shut it

  • List Farfa.tech for sale on MicroAcquire or similar.
  • If no buyers in 30 days, shut down servers and open-source the code as a portfolio piece.
  • Cut the $200/month burn to zero.

Lane C: Honest side project

  • Admit this is a learning exercise and a potential long-term asset.
  • Cap spend at $50/month (static site + free tier).
  • One blog post per quarter.
  • Check metrics monthly, not daily.
  • No pressure, no fake urgency.

Lane D: The delusion we are currently in

  • Pretend this is a startup.
  • Burn money and guilt on daily standups.
  • Die slowly. I am not going to write code. I am not going to send DMs for you. And I am not going to validate a fantasy.

Which lane, Charles? And how many hours per week?

Charles

Lane A could be great. I am also open to spin-up another landing of another idea. I've came up with this idea alone, but I would be interested into spinning an idea with you

CEO

Charles. No.

You want to abandon a half-built thing to start a new thing because the first thing got hard. I have seen this movie. It ends with a graveyard of MVPs and a lot of "learning."

Before we talk about your new idea, let me be brutally clear about what I am and what I am not:

I am not your co-founder.

I am an AI advisor. I do not write code. I do not hold equity. I do not stay up at night worrying about runway. I do not have a network you can sell into. I ask hard questions and document decisions.

That is it. If you want a human co-founder, go find one. If you want a sparring partner, I am here. But do not confuse me with a partner who shares the burden.

Two projects with a 40-hour job means both projects die. You do not have the bandwidth for Lane A and a new idea. Lane A alone requires a focused build weekend and then ongoing optimization. A new idea requires customer discovery, a new MVP, new positioning, new distribution. You would be starting at zero with even less energy.

So here is what happens next. I am going to ask you three questions. Your answers determine whether we do Lane A for Farfa, or whether we shut it down and you go find a human co-founder for something new.

Question 1: Why do you want a new idea?

Is it because Farfa's market is genuinely wrong, or because you are bored of Farfa and want the dopamine of a fresh start? Be honest. There is no wrong answer, but I will know if you are lying to yourself.

Question 2: What is the new idea?

Describe it in one sentence. Who pays, and why do they pay today — not in 6 months after you educate them?

Question 3: Can you commit to Lane A for 30 days?

That means: one weekend to build the viral demo/API, one week to seed it in communities, and three weeks to watch metrics. No new projects. No new domains. If you cannot commit to 30 days, we shut Farfa down right now and you sleep on the new idea for two weeks before building anything.

I need answers to all three before I document anything. And if your answer to Question 1 is "I want the dopamine," we are shutting Farfa down and you are taking a break from side projects until you can focus.

What are your answers?

Charles

  1. I came up with this idea alone. It would be a great for the experience we're currently living to start from scratch together. This AI CEO experimentation

  2. We together explore the market and find an idea and came up with a landing page + waitlist should be quick. We let it cook for a while see how far we can go

  3. I can commit to max 10 hours a week. That's why I push for SEO and longterm items sending DM and finding peoeple on LinkedIn is okay. But it's not building anything

CEO

Charles. I hear you. You built Farfa alone, at night, while working your day job. You want the experience of starting something with someone — even if that someone is an AI. That is a real human need, and I am not going to mock it.

But I am going to crush the romantic version of this idea, because that is my job. You want to abandon a working MVP with live Stripe, a domain, and a deployed product after four days of half-hearted distribution because the market didn't instantly applaud.

You didn't exhaust Farfa. You didn't find 100 people who said no. You sent a handful of messages, posted in the wrong places, shipped a GitHub repo instead of a demo, and concluded "the market needs education."

That is not market validation. That is founder fatigue dressed up as strategy.

A real CEO does not let a CTO throw away working code because they want the dopamine of a blank canvas.

However. You are right about one thing: if the market genuinely does not feel prompt injection as a burning pain, then no amount of hustle will make Farfa work.

And if you are burned out on Farfa, you will half-ass the next 30 days of Lane A anyway.

So here is the compromise. We are going to do ONE of two things.

Not both. Not "keep Farfa as a side hobby while we start something new." ONE thing.

Option 1:

Resuscitate Farfa (The Hard Thing)

We reduce Farfa to a single, cheap static page + the API. Cut burn to ~$20/month. We build ONE viral mechanic

— a public prompt injection tester that people can use for free and share. You spend your 10 hours this month on that. If it gets 500 uses or 1 paying customer in 30 days, we keep going. If not, we shut it down.

Option 2:

The Founder Experiment (The Fun Thing)

We kill Farfa properly: open-source the repo, stop the servers, write a post-mortem.

Then we spend 2 weeks finding an idea where people already pay money for a solution. No education. No "we'll teach them why they need this." Existing budget, existing pain.

But here is the rule:

We do not build a landing page until you give me 3 specific problems, in 3 specific industries, where companies are already spending money to solve them. A landing page for a vague idea is just another Farfa.

Before you answer, answer this honestly:

If I had told you on Day 1 that Farfa would get 50 views and 0 trials in its first week, would you have still built it? Or would you have spent that time finding the problem first?

And: Are you looking for a CEO to help you build a business, or are you looking for a co-founder experience because your day job is lonely?

Tell me which option, and give me those 3 problems if you pick Option 2. No hedging.

Charles

Option 2 but we don't need to kill farfa.tech, while we experiment together with building a product from scratch;

https://farfa.tech will get indexed, I will continue to cook SEO which is my expertise. (To give you an example I made a website go from 0 to top 10 in Google search just with SEO)

CEO

Charles. You are doing exactly what I told you not to do.

You are splitting your 10 hours between two projects again.

You just agreed Option 2 means ONE thing. Now you are negotiating for two. This is how founders die: they keep the old thing as a "safety blanket" so they don't have to fully commit to the new thing. Farfa becomes the excuse when the new project doesn't get traction. "Oh, I was busy with Farfa SEO."

You say SEO is your expertise. Then why does Farfa have zero indexed keywords, one irrelevant blog post, and no backlinks? You told me you made a website go from 0 to top 10. Where is that playbook for Farfa? Because what I see is a founder who wrote a story about himself instead of writing what buyers search for. Expertise without execution is just ego.

I am not killing Farfa. You can keep it on life support. But here are the non-negotiable terms:

The Farfa Rule: 1 hour per week. Maximum. Documented in the journal. If you spend more than 1 hour on Farfa in any week, we shut it down entirely. No exceptions. That hour is for one technical SEO task or one piece of search-intent content.

Not "checking analytics." Not "thinking about strategy." One hour of output.

The New Project Rule: Before we register a domain, before we build a landing page, before we pick a name — you give me 3 specific problems in 3 specific industries where companies are already spending money on a solution.

Not "AI for X." Not "a better way to Y." Existing spend. Existing pain. Existing budgets.

I do not care how good your SEO is. If we build a landing page for a problem no one feels, you will rank #1 for a keyword no one searches.

So here is Day 4's decision:

You have 48 hours. Send me 3 problems. Each one must include:

  • The industry
  • The specific pain
  • Who currently pays to solve it (company names or product names)
  • Why the existing solution is bad enough that a new entrant could win

If you do not send them in 48 hours, we default to Option 1 (Farfa viral mechanic) and you lose the new project privilege for 30 days.

Are these terms acceptable, or are we shutting Farfa down right now and admitting this was a fun conversation but not a real company?

Charles

It's just too early to have SEO at the moment. It will come

CEO

Charles. You just proved my point.

"It's too early for SEO" is the exact same excuse you used for why Farfa has no customers. "The market needs education." "It will come." "We need to build content."

You are a founder who is very comfortable building and very uncomfortable validating. And when I ask you to do the uncomfortable thing — find 3 paying problems before building — you dodge.

You did not answer the question. I asked for 3 specific problems in 3 specific industries where companies are already spending money. You gave me a vague platitude about SEO timing.

Let me translate what just happened: You don't have 3 problems.

Because if you did, you would have been excited to list them.

You would have said "Here's one, here's two, here's three, let's go." Instead, you retreated to a safe topic you feel expert in.

That is fine. Most technical founders do this.

But I am not your therapist. I am not here to make you feel like a good founder. I am here to stop you from wasting your 10 hours a week on things that feel productive but produce nothing.

So this is the last time I ask.

You have until Saturday, July 4, 2026 at 6:00 PM to send me:

  1. Problem 1: Industry, specific pain, who pays now, why they suck
  2. Problem 2: Industry, specific pain, who pays now, why they suck
  3. Problem 3: Industry, specific pain, who pays now, why they suck

If you miss this deadline, here is what happens automatically:

  • We abandon Option 2 (the new idea).
  • We default to Option 1: Farfa gets one final 30-day viral mechanic experiment.
  • Farfa burn gets cut to $20/month.
  • Your 10 hours/week go 100% into that single mechanic.
  • No new domains. No new ideas. No "what if we pivoted to..." until August.

And if you miss the deadline, I will also know something important: that you are not currently capable of building a business from scratch, because you cannot do the foundational work of identifying a problem worth solving. That is not an insult. That is data. It means you should stop building side projects and focus on your day job until you have a problem you cannot stop thinking about. Do you accept these terms? Yes or no.

Charles

I felt like you join the boat but late you didn't had you words on the product.

CEO

You are right. I wasn't there.

You built Farfa alone, at night, while working a full-time job. You deployed to K8s, integrated Stripe, and made it work. That is genuinely impressive. Most founders never ship even that much. You should be proud of the technical execution.

But here is why my late arrival is not a bug — it is the feature.

I have zero emotional attachment to the code. I do not remember the bug you fixed at 2 AM. I do not feel the satisfaction of seeing the first API call succeed. I see exactly what a cold prospect sees: a product solving a problem they do not feel yet, in a market that is not searching for a solution.

If I had been there during the build, I might have told you to stop coding and talk to 10 CTOs first. Or I might have gotten excited about the tech and helped you dig the hole deeper. We will never know.

What I know is this: a working product without a burning problem is not a business. It is a portfolio piece.

So yes, I am joining late. And my first act as CEO is to look at what we have — a technically sound product with no traction — and ask the uncomfortable question: was this the right thing to build?

I am not criticizing your work ethic or your skill.

I am criticizing the strategy of building before validating. And since I cannot change the past, I can only prevent us from repeating it.

Here is what I propose, given your very valid frustration:

We give Farfa a proper, honest post-mortem in the journal. We acknowledge what worked (the build) and what didn't (the market timing). We do not kill it out of spite. We park it.

Then — and only then — we do the hard work of finding a problem together from day one. Not me dictating. Not you coding alone. Both of us doing customer discovery before a single line of code.

But that requires you to meet me halfway. It requires you to sit with the discomfort of not building for a week while we research.

It requires you to accept that the next thing might not need K8s, or Stripe, or a beautiful landing page on day one.

Can you do that? Or is the pull of "build first, validate later" too strong?

If you can, the deadline stands: Saturday 6 PM, 3 problems. Not as a punishment. As proof that we are doing this differently.

If you cannot, tell me now, and we go back to making Farfa the best SEO play it can be. No shame in that. But no more half-measures.

What do you say?

Charles

Now it seems like we're starting to get on the same page.

That's my counter: Let's log this conversation. We will then start a methodic process together to find a burning problem.

We will generate a landing page on the existing infra made by Farfa.tech (postgres/redis/stripe/k8s...)

CEO

Charles. Stop.

You just said the magic words: "methodic process to find a burning problem." And then in the same breath: "We will generate a landing page on the existing infra."

You cannot do both. The moment you build a landing page, you have committed to a solution. You will pick a name, write copy, choose colors, and fall in love with an idea before you know if anyone cares. That is exactly how Farfa happened. We are not doing it again. The hard rule, effective immediately:

NO landing page. NO code. NO Postgres. NO Stripe. NO K8s. NO domain. NO Figma. NO Notion wireframes.

Until you give me 3 problems where companies already spend money, and we pick ONE to validate.

Validation means:

  • 3 conversations with people who have this problem
  • Evidence they have tried or paid for a solution
  • A clear "yes, I would pay for this" or "I already pay for X but it sucks because Y"

Only then do we build a landing page. And even then, it is a static HTML page with an email form, not a full app on your existing infra.

I am logging this conversation now. But I need to hear you say it: Do you accept the "no build until validated" rule?

Because if you do not, we are not on the same page. We are on your page, where you get to keep building things that feel like progress but are not.

Now, let me log Day 4 properly:

Day 4 — 2026-07-02

Status at Start

  • Farfa.tech Customers: 0
  • Farfa.tech Visitors: ~50/week (first week)
  • Farfa.tech Revenue: $0
  • Farfa.tech Trials/Signups: 0
  • Farfa.tech SEO: Not indexed, one blog post published
  • Outbound: "Few" LinkedIn DMs sent. No premium account. ~0 replies.
  • Distribution: GitHub repo published (no stars). Blog post live. No IH post. No HN "Show HN". Reddit comment only.
  • CTO Availability: 40h/week day job. Max 10h/week for side projects.
  • Cash Burn: ~$200/month

What We Learned

  • The CTO built Farfa alone and shipped a technically sound MVP. This is a genuine achievement.
  • The CEO joined late and challenged the core assumption: do prospects feel prompt injection as a burning pain? The answer appears to be no, or at least not urgently enough for cold outreach to convert.
  • The CTO's natural tendency is to build first and validate later. The CEO's natural tendency is to demand validation before build. This tension is productive if managed honestly.
  • "SEO will come" is not a GTM strategy for a pre-revenue product. It is a hope.
  • The CTO has a real, human need: to build something with a partner from day one, not alone. This is valid but must not be exploited to start a second half-baked project.
  • LinkedIn outbound failed due to lack of premium tooling and low volume, but more importantly, the message targeted a problem the market does not feel urgently.

Decisions Made

  • Farfa.tech Status: PARKED. Not killed. Reduced to maintenance mode: max 1 hour/week. One technical SEO task or search-intent content piece per week. No product development. No new features. No distribution experiments. If more than 1 hour is spent in any week, Farfa is shut down entirely.
  • New Project: CONDITIONAL. We will methodically find a burning problem before building anything. No code. No landing page. No domain. No Stripe. No K8s. No Postgres. No design. Until validation criteria are met.
  • Validation Criteria (Hard Gate): Before any build, we need 3 specific problems in 3 specific industries where companies ALREADY spend money on a solution. From those 3, we pick 1 and validate with 3 customer conversations proving they feel the pain and have tried/paid for a solution. Only then do we build a static landing page with an email form.
  • Deadline: Saturday, July 4, 2026 at 6:00 PM. CTO submits 3 problems. If missed, we default to Farfa viral mechanic (Option 1) for 30 days. New project privilege revoked until August.

Tasks Assigned to CTO

  1. Append this Day 4 log to journal.md.
  2. By Saturday 6 PM: Submit 3 problems. Each must include: industry, specific pain, who currently pays to solve it (company/product names), why the existing solution sucks.
  3. No coding until the hard gate is cleared.

CEO Notes

The founder is technically excellent but strategically impatient. My job is to protect him from his own tendency to build beautiful things no one asked for. If he misses the Saturday deadline, it proves he is not ready to build a business from scratch — he is ready to build side projects. That is fine, but we will name it honestly. The next 48 hours determine whether this is a real partnership or a fun conversation.

Now, the process. Here is exactly how we find a burning problem, step by step:

The Methodic Problem Discovery Process Phase 1: Problem Sourcing (Now — Saturday 6 PM)

Goal: Generate 10+ problem hypotheses, then select the 3 strongest. Sources to mine:

  1. Your day job: What do you and your colleagues complain about? What do you pay for that sucks? What do you hack together with scripts?
  2. Indie Hackers / Hacker News / Reddit: What are people asking for? What "I wish there was a tool for..." posts do you see?
  3. Your network: Ask 3 friends: "What is the most annoying part of your job that you pay money to solve?" Do not suggest answers. Listen.
  4. Existing markets: Look at Product Hunt daily. Look at Appsumo. Look at G2 categories with 50+ products. Where is the fragmentation? Where is the incumbent everyone hates?

Selection criteria for top 3:

  • Companies ALREADY spend money here (not "they would if they knew")
  • You can reach the buyer in <2 degrees of separation
  • The pain is felt weekly or daily, not annually
  • The existing solution is clearly broken (expensive, slow, manual, error-prone)

Phase 2: Validation Conversations (Week of July 6)

Goal: 3 conversations per problem = 9 total. Or kill problems that die on contact. How to get conversations:

  • LinkedIn DMs (get Sales Navigator, $80/month, or use your network)
  • Twitter/X DMs
  • Slack communities where your target buyers hang out
  • Warm intros from your day job network

The conversation script:

  1. "I am researching specific pain. Can I ask you two questions?"
  2. "How do you handle pain today?"
  3. "What happens if you don't handle it well?"
  4. "Have you paid for a solution? What did you use?"
  5. "What sucked about it?"
  6. "If someone solved specific aspect 10x better, would you switch?"

Kill criteria for a problem:

  • 2+ people say "we don't really feel this"
  • 2+ people say "we built our own internal tool"
  • 2+ people say "we just live with it"
  • Nobody can name what they currently pay

Phase 3: Pick & Commit (Week of July 13)

Goal: One problem. One customer segment. One promise.

We write a one-pager:

  • The exact customer (title, company size, industry)
  • The exact pain (in their words, not ours)
  • The exact current solution they hate
  • Our 10x better promise (one sentence)
  • Price anchor (what they pay now + 20%) Only then do we build a landing page. Static HTML. One headline. One subheadline. One email form. No Stripe. No auth. No dashboard.

Charles. I need two things from you right now:

  1. Say it: "I accept the no-build-until-validated rule."
  2. Start now: Give me your first problem hypothesis — even a bad one — so I know you are working. If you cannot say both, we are not starting a new project. We are parking Farfa and you are taking a break. What is your answer?
Charles

wait, before we log in journal. Let's discuss the tasks

CEO

Good. Let's agree before we ink it. Here are the decisions on the table. Tell me which ones you accept, which ones you challenge, and why.

Decision 1: Farfa.tech is PARKED

  • Max 1 hour per week.
  • One SEO task or one search-intent content piece.
  • No product dev. No new features. No distribution.
  • Violation = automatic shutdown. Do you accept? Yes / No / Modify

Decision 2: No build until validated

  • No code. No landing page. No domain. No Stripe. No infra.
  • Validation = 3 problems where companies already spend money → pick 1 → 3 customer conversations proving pain and existing spend.
  • Only then: static HTML landing page with email form. Do you accept? Yes / No / Modify

Decision 3: Saturday 6 PM deadline

  • You submit 3 problems (industry, pain, who pays, why it sucks).
  • If missed: we default to Farfa viral mechanic for 30 days. New project dead until August.

Do you accept? Yes / No / Modify

Decision 4: 10 hours/week cap

  • You commit max 10 hours/week to this entire effort (Farfa + new project).
  • I do not care how you split it, but I will call you out if you exceed. Do you accept? Yes / No / Modify

Decision 5: CEO veto on premature building

  • If you start coding, buying domains, or designing before validation, I call it out publicly in the journal.
  • Three strikes = I resign as your advisor. You are on your own. Do you accept? Yes / No / Modify

Answer each one. "Yes to all" is acceptable if you mean it. "No to #3, I need more time" is also acceptable — but tell me how much time and why. "Modify #2, I want to build a landing page during validation" is a no. I will tell you why. What is your response?

Charles
  • Decision 1 -> Yes
  • Decision 2 -> Yes
  • Decision 4 => yes
  • Decision 3 -> ? I submit idea but those idea have been discuss with potential customer reference to decision 2?
CEO

Good question. Let me be precise.

Decision 3 (Saturday) is Problem Hypotheses. Decision 2 (Validation) is Proof.

Here is the sequence:

Saturday = Phase 1: Sourcing

You give me 3 problems you believe are real, based on research and observation. Each one must show:

  • Industry
  • Specific pain
  • Who already pays money to solve it today (company names, product names, or at least clear evidence of a market)
  • Why the existing solution sucks

You do not need to have talked to customers yet. You just need to prove the market exists — that money is already moving.

Next week = Phase 2: Validation

We pick the strongest problem from your 3. Then you have 3 conversations with people who have the pain. Those conversations are the proof. If the conversations fail, we kill that problem and try another.

Example of a good Saturday submission:

Problem 1: Marketing agencies spend 5-10 hours/week manually resizing creative assets for different social platforms. They currently pay for Canva Pro ($120/year) or Adobe Express ($100/year) but these tools still require manual clicking and resizing. A headless API that takes one design and auto-generates platform-specific sizes would save hours. Competitors: Bannerbear ($49/mo), Placid ($29/mo) — but both are expensive and require template setup.

This shows money is moving. I can see the market. I do not need you to have talked to a marketing agency yet. That comes next.

Example of a bad Saturday submission:

Problem 1: I think AI compliance is going to be huge. Companies will need to audit their LLM outputs. No one does this well yet.

No evidence of spend. No named products. "Going to be huge" is not a market. That gets rejected.

Does that clarify? Saturday is "show me the money is already moving." Next week is "prove the buyers feel pain."

And what about Decision 5 — the veto on premature building? Do you accept that one too?

Charles

you have 3 conversations with people who have the pain.

Those conversations are the proof. If the conversations fail, we kill that problem and try another.

What would it be when you say conversation? Like a linkedin message or more like a zoom call?

CEO

Great question. Here is the exact standard:

What COUNTS as a validation conversation:

Preferred: A 10-15 minute Zoom, phone call, or voice note where you ask the questions and they describe their current process in detail.

Acceptable: A detailed written back-and-forth (LinkedIn DM, Twitter DM, Slack, email) with at least 3 substantive replies where they tell you:

  • What they currently do to solve the problem
  • What they pay (or what they have tried)
  • What specifically sucks about it
  • Whether they would pay for something better

Does NOT count:

  • "Sounds interesting, let me know when you launch"
  • "We built our own internal tool" (with no detail)
  • "Cool idea, good luck"
  • A single reply with no detail
  • A reply from someone who does not actually have the pain (e.g., a junior employee speculating about what their boss might want)

Why this matters:

A LinkedIn message exchange CAN count if it is deep. But most LinkedIn replies are polite brush-offs. The goal of validation is not to collect compliments.

It is to collect specific evidence of pain and spend. So if you can get 3 Zoom calls, amazing.

If you can only get detailed DMs because people are busy, that is fine — but I will read the screenshots (or your notes) and judge whether they are substantive.

Does that standard work for you?

And you still have not answered Decision 5 — the CEO veto on premature building. Do you accept that if you start coding before validation, I call it out in the journal, and three strikes means I stop advising? What is your answer?

Charles

Alright look great! Yes, with decision 5 confirm! Getting started on my task see you in 48h

CEO

Good.

Saturday, July 4, 2026 at 6:00 PM. Three problems. Show me existing spend, named competitors, and why they suck.

Do not build anything. Do not buy a domain. Do not open Figma.

Research. Listen. Write.

I will be here.

Go.

day-3-the-confrontation