timmsbuilds
NotesfoundercontentNUMBERS

We rebuilt a dead 17,800-subscriber site in a weekend with Claude

Founder Content was our community of founders before we had a product to sell them. It died because every piece was done by hand. We brought it back in a weekend, and now the library and its marketing run themselves for about $80 a month.

Nick Timms · 19 Nov 2026 · 7 min read
The short answer

Founder Content’s library took two years to collect by hand, as part of one person’s job, and it died when that stopped. We rebuilt it with Claude in one weekend: the first version went live on Saturday 7 November at 20:58, with all 1,000 resources and 5,113 old links kept. Claude added five years of new founder writing, and the site now posts one founder a day across six channels. It costs about $80 a month to run.

THE RESULT SO FAR
2 daysto rebuild what took 2 years1,119 resources
7,500+Instagram views in the first 10 days96% new people
8,100old subscribers back in touch33% opened
One founder a day, every day, on
foundercontent.com
Founder Content today: Where founders give a s#*t about other founders. Featured pieces from HubSpot, Canva, Slack and Front, a sign-up box, and the library: 1,119 resources from 318 founders across 25 topics.
Founder Content today, rebuilt with Claude: 1,119 resources from 318 founders.

We built Founder Content in 2021 to bring founders together. Plenty of marketing advice says build the community first and the product second, and this was exactly that: the best of what founders have written and said about building a company, in one place. It launched on Product Hunt, finished fifth of the day, and 17,800 founders signed up.

It was also a lot of work. Every piece was found, read and written up by hand, by one person, on and off for two years, alongside our blog. People wanted it, but we couldn’t keep feeding it and focus on everything else. It slowly died.

This year, reviewing our marketing, one thought kept coming back: we still had a list of 17,800 founders. On 3 November we nearly switched the site off for good.

A task in our to-do list on 3 November: redirect foundercontent.com to nicktimms.com. The old site no longer loads its content; an email has gone to our host provider asking for a 301 redirect.
Our to-do list on 3 November.

We always thought Founder Content was an excellent idea, made stronger by a close-knit community of founders in the same position: the stress, the loneliness, the hours. Hearing from people who’d walked the path, picked with care, was a real help to them. But it was overwhelming to maintain, and we were about to stop. Then we realised how much of it could now look after itself. Even if it never makes money, it’s helping founders in the same position we were in.

Three days later we asked a different question: what would it take to bring it back properly?

Day by day

From “let it go” to live in four days.

  1. Tue 3 NovWe ask our host to redirect foundercontent.com. The plan is to let it go.
  2. Fri 6 NovWe decide to rebuild it instead.
  3. Sat 7 Nov, 20:58The first version is live: all 1,000 resources, with every old link still working.
  4. Sun 8 NovTopics, companies and a page for every resource. I review 243 new pieces in one sitting.
  5. Mon 9 NovCompany facts, founder careers and a welcome email. The first Instagram post goes out.
  6. Tue 10 NovThe first daily founder goes out on every channel at 1pm.The first slide of the Reid Hoffman post: Reid Hoffman was here.Reid Hoffman was hereShipping version one: “If you’re not embarrassed by your first version, you launched too late.”View on Instagram ↗
  7. Sun 15 NovThe first short video clip.The cover of the Tobi Lütke reel: Shopify has no KPIs.Tobi Lütke, in his own words“Shopify has no KPIs”, a clip from the original interview, captioned and linked back to it.Watch the reel ↗
  8. Mon 16 NovThe first automatic Monday update. It failed. More on that below.
The site itself took 26 commits, from Saturday evening to Monday lunchtime.

Then and now

Same name, same links, a very different library.

foundercontent.com
The Founder Content home page in 2022, from the Wayback Machine: Where Founders give a s#*t about other Founders.
The same home page in 2022, when every piece was added by hand.

Every old link still works. Claude matched all 5,113 old addresses to their new pages, so visits from Google, old newsletters and tweets still land somewhere.

Same address, years apart.

Claude finds. I decide.

Curation is the whole product, so the final call is never Claude’s. Claude searched five years of founder interviews, essays and talks, scored every find against a one-page brief, and sent me anything good enough. Then I went through every one myself.

The first review · Sunday 8 November
274pieces Claude scored against the brief, from 2021 to now5 tests, 1–5 each
3.5 hrsfor me to go through every one that passed, in one sittingA approve · R reject
25 of 65of Claude’s top-scored pieces I still turned down123 made it in
localhost:4500 · our review page
Our Founder Content review page on 8 November: Bret Taylor of Sierra on AI agents, outcome-based pricing and the OpenAI board, a 101-minute podcast scored 24 out of 25, with Approve, Reject and Skip buttons and the reject reasons open.
The real review page. Claude scored this one 24 out of 25. I pressed R: too thin.

Anything under 20 out of 25 never reaches me. But the rule that mattered most came from my own no’s: a piece has to be one of the best on one topic. A famous founder on a wide-ranging chat show doesn’t make it, however good the chat. Here’s the start of the brief, if you want to borrow it.

FOUNDER CONTENT · THE SELECTION BRIEF (the first part)

Auto-reject: the link doesn't open or doesn't match; not first-hand;
a compilation or re-upload; a hard paywall; mainly promotional;
before 2021; a duplicate.

Score each piece 1-5:
- First-hand: 5 = they did it and explain how; 1 = opinion about others.
- Substance: 5 = depth, numbers, real decisions; 1 = surface level.
- Specific: 5 = a founder could act on it this week; 1 = platitudes.
- Still relevant: 5 = holds in 2026; 1 = already dated.
- Accessible: 5 = free, clean, easy to take in; 1 = hard to reach.

20+ out of 25 with nothing below 3 goes to Nick or Duda.
The rest is rejected, with the reason kept.

1. One-topic test (most important). The piece must fit clearly into
   ONE topic and be a strong example of it. Wide-ranging conversations
   are out, however famous the guest.
2. Name three events. It must cover 2-3 specific decisions or moments
   (a pivot, a near-death, a repricing, a first sales hire, a layoff)
   with concrete detail. Themes alone = too thin.

One resource, then and now

The old library was a list of titles. In 2021, the whole entry for the Airbnb seed deck was one line: “This is the AirBnbPitch Deck, used by the AirBnb on their Seed Round in 2008. The total amount raised was $600,000.” This is its page today.

foundercontent.com/pitch-decks/airbnb-pitch-deck
The full Airbnb pitch deck page today: Brian Chesky, a link to the original on SlideShare, a summary, five key takeaways, who it’s for, why it’s worth your time, company facts, more from Brian and more pitch decks.
Claude writes 40 of these a week and we check every one. The original is always one click away.

The five missing years

The old library stopped in 2021, so it had nothing on the AI wave. We wrote a one-page brief and sent Claude through everything since. The test: judge the piece, not the person. A founder’s account of what they did is in; the same founder’s take on the news is out.

It found 557 candidates in a few days, and nothing reached the site until it was checked. The library now holds 1,119 resources from 318 founders, with a new way in: a map of every stage of building a company. Pick where you are and it shows the founders who’ve been there.

foundercontent.com/pivoting/des-traynor-intercom-how-we-rebuilt-the-company
One of the new 2026 pieces: Des Traynor of Intercom on how the company rebuilt itself for the AI era, with a summary, five key takeaways, who it’s for, why it’s worth your time, more from Des and more on pivoting.
One of the new pieces: Des Traynor on rebuilding Intercom for AI, from January 2026. Claude scored it 25 out of 25, and I approved it.
foundercontent.com/map
The founder map, from Idea to Exit.

The marketing runs itself too

Every page in the library is also a post. The old Founder Content reached people one way: the newsletter. The new one posts a founder a day on six channels, walking the map from Idea to Exit each week. Claude turns each page into a six-slide carousel and writes the text for each channel, and every link is tagged so we can see what works.

Brian Chesky was here: before Airbnb was Airbnb, he had to show people wanted it.
The deck opens with the problem: hotels are expensive, they cut you off from the city, and there’s no easy way to stay with a local.
Then it shows proof, a market and a model: demand data, a market sized for budget travellers, a 10% commission, and winning big events first.
The takeaway: sell the problem and the proof first.
Thursday 12 November on Instagram: the first four of six slides.
The Buffer calendar for November 2026, filtered to Founder Content’s channels: posts at 1pm every day on LinkedIn, Threads, X, TikTok and Instagram, queued through 24 November.
The real Buffer calendar on 16 November: a founder at 1pm every day, short clips at 6pm a few times a week, queued through 24 November.

What it costs

What it costs to run Founder Content, a month
ThenOne person’s monthly salary. Founder Content was part of their job, on and off.$2,083
NowBuffer $54, Supabase $10, Claude ~$10, Vercel $0–20~$80
US dollars. Claude’s cost is the 40 summaries a week. We release only 40 new pages a week to Google, so 1,520 wait their turn.

What broke

  • Wrong covers. On Sunday night some fresh cover images didn’t match their pieces. We put the 2021 covers back.
  • Odd old links. Some old addresses had spaces in them, and some came from an even older version of the site. All of them now redirect.
  • A sleeping Mac. The first Monday update ran while my Mac was offline and stopped at step one. It now waits for the internet and tries again later in the day.

Week one, in real numbers

Early numbers, all real. The daily posts started on Tuesday 10 November, so this is about a week.

Instagram account insights for foundercontentdaily, last 30 days: 7,523 views, 96% from non-followers, 3,023 viewers. Top posts: the Every great founder was once where you are post with 7.8k views, then Tobi Lütke, Brian Chesky, Melanie Perkins and Dario Amodei.
Instagram, the first 10 days: 7,523 views, 96% from people who didn’t follow us yet.

On the site, visits come from social first, then Google and AI assistants. Google already shows Founder Content in 272 different searches, mostly on page two for now.

Buffer insights for Founder Content’s six channels, 10 to 16 November: 37 posts, 74 followers, 11 reactions, 0 comments, 2.41% engagement rate.
The six channels, 10–16 November: 37 posts, 74 followers from zero, 11 reactions, no comments yet.

And the old subscribers are hearing from us again, in waves, most engaged first. About 8,100 have had this email so far, and the first 756 opened at 33%.

The email the old Founder Content subscribers got: Founder Content is back (and it’s different), from Nick at Founder Content. It says they subscribed a few years ago, it went quiet, and we’re getting back in touch, with a link to our latest article.
The email the old subscribers got. It went quiet, we said sorry, and we started writing again.

Founder Content is at foundercontent.com, with one founder a day on Instagram. I’ll share the 30-day numbers in December.

Know a founder who needs this?

“We rebuilt a dead 17,800-subscriber site in a weekend with Claude. It now runs itself, marketing included, for about $80 a month.”

timmsbuilds newsletter

17,188 following along · Thursdays