Do you still believe this?

There is more startup advice on the internet right now than at any point in human history.

There has also never been a worse time to blindly follow it.

For 20 years, venture firms have been turning their wins into playbooks. “How to ___” hire/build a sales team/scale.

The premise is - someone already made the mistakes, so you do not have to.

Except there might be a problem with that.

While digging through 400+ firms website data behind last week’s Market for Words, I found 107 venture firms have published or promoted some kind of “playbook”.

BUT, I also found 33 firms explicitly telling founders that “there is no playbook.”

So which is it!?

Well, a playbook is just experience written down, but now experience has a rapid half-life.

The company that taught your investor how to scale was built with different tools, different costs, different distribution channels, different competitors and, increasingly, a completely different idea of what a company needs to look like.

For founders, that creates a problem. The venture firms publishing the advice, have a different one:

What is worth publishing when you cannot know how long it will stay true?

VC advice or assumptions?

In 2014, First Round told first-time founders that if they were not spending at least half their time hiring, they were "already on the road to failure."

A year later, Insight Partners told software founders that turning a good product into a big company meant scaling an inside-sales team.

Neither was bad advice - both firms were publishing hard-won lessons from companies that had scaled.

But buried inside both was an assumption so obvious… to scale a company, you scale the organisation.

12 years later, that assumption is dying. Blitzscaling is out of fashion and tiny teams are ripping past $100M.

The excellent Saul Klein was on the Odin podcast and said the methods, tools and infrastructure used to build businesses now change every 6 to 12 months, which makes the half-life of being a useful adviser - based on past experience - short, and shrinking.

Some principles are timeless, but the technical stack and tactics of a startup today are unrecognisable from 5, 10, 15 years ago.

He said it’s why Local Globe builds internally, as useful now requires continually refreshing your exposure to the production environment.

Experience from 2014-2022 is not automatically additive anymore. In tactical domains, old experience can be a prior that needs updating rather than an advantage.

Timeless is Usually Generic

The obvious pushback is 'just publish timeless content.' But “timeless” is usually just generic…

Hire great people. Talk to customers. Focus on product etc etc These sentences survive for decades because they are abstract enough to survive anything. They are true without being useful.

Yes, Do Things That Don't Scale is 13 years old and still circulating. But Paul Graham is the exception everyone cites - even over-analysing PG content is its own genre of expired content.

Meanwhile, something extremely specific and tactical can be enormously valuable exactly because it captures how things work right now, even if it is useless in 3 years.

So longevity is not actually the thing to optimise. The question is what kind of claim a piece is making about its own decay.

Content Decays at 5 Speeds

Venture content decays at different rates depending on what it describes.

1. Human Nature: Decays slowly (founder psychology, incentives).

This is where the PG essays live.

2. Business Physics: Mechanism stays, parameters move.

Default Alive or Default Dead still works - but the burn rates and round sizes in the examples do not.

3. Technological Regime: Needs active versioning.

Right until the regime shifts, then rapidly wrong. e.g. Bessemer's Cloud Laws. The PLG canon.

4. The Frontier: Expiration is the feature (Mobile Is Eating the World).

Its expiration is almost evidence that it did its job - recency is the point.

5. Historical Record: Expired playbooks preserved as primary-source evidence of past regimes.

some content can acquire a second life as primary-source history. A 2013 First Round piece about how a startup built its sales org will eventually be bad advice and excellent evidence - of how great companies were built under a previous regime… which in itself might distill new lessons applied to a new context.

Content is therefore not binary between EVERGREEN vs DATED.

But a better question is what is a piece is trying to preserve?… a principle, a playbook, a prediction, a benchmark, or a record.

LLMs compound the urgency

The half-life problem used to be partially self-correcting. Old posts sank in search results or advice aged visibly (outdated site designs/logos etc).

But model weights do not work like that. LLMs can retrieve past thinking with present confidence. I found this when asking 8 AI models 2,000 questions about venture.

A 2014 sales-team playbook served to a founder today arrives with no temporal metadata - indistinguishable from current advice, stripped of its date, and preserved at full strength.

Founders now consume venture knowledge through a layer that does not version. Which means the versioning has to happen at the source, or it does not happen at all.

Publish like software

So the practical answer. Most fund content runs on 1 loop:

write → distribute → archive → forget.

But if founders keep discovering your advice through Google, ChatGPT and your own resource centre years later, you have an intellectual product in circulation - whether or not you are maintaining it. The loop should be:

write → distribute → observe → revisit → version.

This isn't an unprecedented problem. When I was writing about the history of VC marketing I found this 1968 banger from BCG. The Experience Curve, argued costs fall predictably with experience and prescribing that "market share must be maintained at all costs." They revised this in 1973 and again in 2013 (45 years later) with a fundamentally different conclusion - markets got volatile, cycles got faster, and experience was no longer enough.

But they never deleted the original.

Much like this, venture advice should be more like software, where it does not pretend v1 is eternally correct.

Imagine the genuinely excellent 2019 article, How to Build Your First Sales Team, treated like a product:

• How to build your first sales team - 2019.

• Updated 2022. (PLG? Maybe remote sales?)

• 2024 Revisited (AI-assisted?)

• Rewritten 2026. What we got right. What stopped being true. What we would tell founders now.

This is not hypothetical anymore… literally as I was writing this and 12 years after the hiring advice this piece opened with - First Round launched the Sales Combine.

A free 2-day program that trains people for AI-era sales, no experience required, then places them directly into port cos that are hiring.

The 2026 version of First Round's sales playbook is not an article at all. When the half-life of advice gets short enough, the playbook isn’t something a firm publishes and becomes something the firm runs - which is exactly Saul Klein's point… build!

They have form here too (the PMF Method, Angel Track), and the training preview still cites their decade of sales articles underneath it. The old advice was not deleted. It became the foundation the new program stands on.

This software mindset unlocks a content moat that cannot be easily copied - versioning. Anyone can publish a hot take today, but only firms with real history can show how their thinking evolved across cycles.

To build that moat, I think firms must be braver about grading their own work in public and ask...

• "Do we still believe this?"

Take your own 5 year old piece, grade it in public - still true, directionally true with the mechanics changed, obsolete, or inverted. Then say why.

I don’t mean apologising for mistakes, just self referential content is an insane content moat - New content = opinions. Graded content = process.

So, what do you publish?

Back to the question from the top… what is worth publishing when you cannot know how long it will stay true?

A good concept is to think about 2 clocks…

1 publishes the slow stuff - human nature, business physics - updated only when the ground has really shifted.

The other works the tactical frontier, and accepts a short shelf life as the price of being useful right now. But date it, build on it, scope it, version it when it breaks.

Which also settles the 107 vs the 33. The playbook firms are right on the day they publish, the no-playbook firms are right a few years later.

But if what you publish in 2026 still feels completely relevant in 2030, you probably kept it too abstract to help anyone.

Laurie, Refinery Media

If you made it all the way through, thanks so much for reading! A thousand VCs now open this every week. If it's helped you think differently about marketing, Venture, or storytelling, please send it to someone in your orbit.

If you enjoyed this, read more from our top posts:

• Most VCs are Not Playing The Money Game

• a16z's $360M/year media operation

• Why Founder-First is Not a VC Content Strategy