Threads of Intent

Startup stories are much more than just the numbers

🔌 Cursor user? Try this new extension to capture, search and learn from your journey.

---

Startup stories begin and end by the numbers, don’t they? Instead of funding, lets survey failure.

At the start of COVID-19 in March 2020, Roger Lee created Layoffs.fyi which began tracking tech companies and their adjacencies. Now a digital graveyard of ambition.

Of the 2728 companies it has tracked, 291 (~11%) have had 100% layoff events (25 of which were acquired).

This equates to ~$30B of invested capital loss.

Notable flops in that data include BritishVolt ($2.4B), Fisker ($1.7B), Quibi ($1.8B), Convo ($1.1B), BlockFi ($1.1B).

1000s of smaller companies without a news-bite to capture have closed in that same time period based on data from Carta at a rate accelerating YoY into 2024.

Pitchbook estimated that 3200 startups failed in 2023 alone.

The data is fragmented, but the facts—no matter how you piece them together—are indisputable: out of roughly 55,000 VC-backed companies operating in the U.S. at any given time, success is anything but guaranteed.

This is hardly news. Historically, the average startup’s chances of reaching that vaunted IPO have always been well under 1%.

It’s always good to remind yourself that survivorship bias is why we call them unicorns.

This meditation is not about inevitable failure

If you know anything about startups, you know they’re far more than just aggregate statistics.

They're thousands of small decisions, countless lines of code, and infinite "what if" moments and forks in the road.

They’re a process of discovery where you have to figure out how to remove problems — progressively by creating software — that the market cares enough about to pay you for.

In a way that both scales and is profitable enough to keep you default alive so you can keep removing more. The infinite game personified.

Benn Stancil’s How A Startup Feels describe the nature of work like:

Sitting in front of a mixing console worth of controls and not knowing what to do. It is important to configure the machine well, and you will have no idea how to do it.

In this process of discovery you learn about yourself.

I’ve described the story of my experience at The League — a VC backed startup ultimately acquired by Match Group — as the most enlightening experience that I would never want to repeat.

Its benefit?

You learn which knobs matter. Which signals to trust. You learn that some of your most impressionable moments come from conversations - with customers, with teammates, with mentors, and now increasingly, with AI.

Threads of intent

We’re all spending, and will continue to spend, far more time interacting with AI.

In the last three months alone, I've had over 400 conversations with ChatGPT alone.

Each one a thread of intent - not just questions to a model, but an expression of some possibility. Each prompt, each response, each exchange fits into piece of the story we're trying to tell ourselves and increasingly to each other.

Not everything is worth capturing. Much of it is intentionally disposable. But in the chaos of creation, pivotal moments of clarity—stories that unveil new possibilities—often slip away before we can learn from them

Lost to time, buried in chat histories, or forgotten entirely.

Insights that could, if well structured, light the way forward for others vanish into ether.

Earlier this week we launched the first SpecStory product iteration. If you’re using IDEs like Cursor to create software our intent is to be your AI conversation time machine - preserving the context, intent and invisible threads that connect idea to implementation.

What’s the most fun is exploring this uncharted territory.

Intent isn't just a command to a machine but rather the essence of how we express possibility. It's what lies between the lines of code, in the spaces between decisions. We're building the yet to be fully defined infrastructure to capture these crucial in-between moments, these sparks of creativity that too often slip through our fingers.

Perhaps that's the real story here

While the startup graveyard keeps filling with aggregate statistics, the true lessons - the ones that matter - live in these moments and more-so between conversations with increasingly advanced “personified” models.

We started with the Cursor IDE because every journey needs a first step. But the vision extends much beyond just capturing these conversations.

As

Jake Levirne

wrote in our blog:

It’s a bridge between traditional software engineering and a new way of creating software. We're one of the earliest teams to successfully blend software engineers and software composers, working together in harmony.

This isn't about replacing traditional engineering – it's about augmenting it, making it more accessible, and allowing more voices to contribute directly to a product's evolution.

To do that well, we need better institutional memory.

And it’s time to measure something different: the quality of our decisions, the clarity of our intent, and our ability to learn from every step of the journey.

That’s the story we’re here to help you tell.