Entrepreneurship is as much about persistence than about being right the first time.

Cursor co-founder Michael Truell’s Hacker News submission history is a quiet record of that truth. Before Cursor became the fastest-growing SaaS startup in history — before the $29.3 billion valuation, before SpaceX agreed to acquire it for $60 billion, before the four MIT co-founders turned billionaires in their mid-20s — there was a string of submissions under the username mntruell that most people scrolled past without a second thought.

Eight of them, going back to January 2022.

What the Submissions Actually Show

The earliest entries are the most telling. In January 2022, Truell posted “Cursor – An AI Email Companion” from cursor.so, a product that had nothing to do with code editors. He posted it twice: once with 2 points, once with 4 points and 5 comments. These are not the numbers you frame on a wall. The product was an AI tool for email — a reasonable thing to build in early 2022, when the race to put AI into productivity workflows was just beginning. It went nowhere.

Two months later, in March 2022, the team came back with “Show HN: A Copilot for Diffs” — a more developer-focused idea, hosted on cursor.sh. One point. No comments.

Then silence for most of 2022, and a pivot.

By March 2023, the name and the ambition had grown. “Cursor: A code editor built for programming with AI” landed 14 points and 11 comments — small numbers still, but a different category of response. The product had found its real direction. Then in January 2023, Truell had linked to a post by co-founder Aman Sanger titled “Cursor: An AI-Enabled IDE”, picking up 8 points.

By January 2025, the submissions were about specific technical capabilities inside a product that people were already using: “Character Prefix Conditioning” (a blog post about how Cursor handles token-level code completion) appeared twice — once as a quiet 1-point post and once as a 29-point submission with 15 comments. And “A New Tab Model,” also from January 2025, announced the next generation of Cursor’s Tab completion system.

These last few submissions are a different world from the 2022 email companion. The team wasn’t pitching a concept anymore. They were publishing research on a tool that millions of developers were already depending on.

The Through-Line

What the submission history shows, more than anything, is a team that kept shipping. The email companion didn’t stick, so they moved to diffs. Diffs didn’t land, so they went all-in on a code editor. The code editor found an audience, and they doubled down, building proprietary models and publishing technical writing about how those models worked.

Truell has spoken about the early days of the actual Cursor IDE with characteristic understatement. In 2023, the team “lived like monks” and focused entirely on the product, letting it spread by word of mouth. No growth hacking, no press tours — just relentless iteration on something they believed in. In 2021, the co-founders were wrestling with what to do with their interest in AI — whether to pursue it in academia, join an existing AI effort, or start their own thing. The eight Hacker News submissions are, in a sense, the answer playing out in public.

The product that eventually worked was also the one they cared about most personally. Truell and his cofounders were obsessed with GitHub Copilot when it launched for individual developers in 2022, and they believed they could build something far more deeply integrated. What may have made Cursor succeed where earlier projects failed was a simple decision: to go all in. “We had a ton of conviction about that, and we had a ton of excitement about that, and so at some point we just decided to go for it,” Truell said.

That conviction eventually compounded into extraordinary results. Cursor raised a $60 million Series A in June 2024. By the end of 2025, it had closed three more rounds bringing in $3.3 billion, with its valuation going from $2.5 billion to $30 billion in a single year. By April 2025, Cursor was writing 1 billion lines of code a day. More than a third of pull requests at Cursor were being created by autonomous agents. Each of the four co-founders — none older than 26 — now holds a stake worth roughly $5.5 billion after the SpaceX deal.

This Has Happened Before

Truell’s story follows a pattern that has repeated across the history of technology entrepreneurship. The founders who build something important rarely do so on the first try, or the second.

Brian Chesky and Joe Gebbia came closest to quitting Airbnb before they built it. The company launched multiple times in 2008 with minimal traction — once with no luck, and a second time with only two customers, one of whom was Chesky himself. They had pitched the idea to 15 angel investors and been turned down by all of them. To keep the company alive, they sold custom presidential-election cereal boxes — “Obama O’s” and “Cap’n McCain’s” — raising $30,000 to fund operations. The founders of what would become one of the most valuable hospitality companies in the world were literally selling breakfast cereal out of desperation.

What saved Airbnb was persistence long enough to get accepted into Y Combinator, where Paul Graham pushed them to actually go meet their hosts in person, understand what people wanted, and fix the product accordingly. The idea hadn’t changed; the execution had been refined to the point where it finally worked.

There is a long list of companies with similar founding stories — YouTube pivoted from a video dating site, Slack from a failed gaming startup, Twitter from a podcasting platform called Odeo. The pattern is consistent: most big ideas arrive in the wrong packaging first. The founders who make it are the ones still showing up when the right version finally clicks.

Why Persistence Is the Variable Most People Ignore

The Hacker News record matters because it’s unedited. There’s no PR narrative, no origin story smoothed out for a pitch deck. Just a series of timestamped bets, most of which didn’t pay off.

The 1-point submissions from 2022 are exactly as unglamorous as they sound. The team was working out of MIT, posting ideas that nobody upvoted, building products that attracted no users. The gap between that and a $60 billion SpaceX acquisition is not explained by a single moment of inspiration. It’s explained by staying in the game long enough to find the right problem to solve.

Truell started coding at 11, built a popular programming game as a teenager, interned at Google after his first year at MIT, and spent years thinking carefully about where AI was headed before Cursor’s code editor ever shipped. He maintained an unusually low public profile for a founder whose company reached billions in ARR — Cursor’s growth was driven almost entirely by product quality and developer word-of-mouth, not founder celebrity. The eight Hacker News submissions fit that profile exactly. No hype, no viral threads about their journey. Just shipping, learning, and shipping again.

The email companion is long gone. The tab model is now what developers around the world open every morning when they sit down to write code.