Why Did Kortex, the “$3 Million Mistake,” Fail and Become Eden?

TL;DR

Kortex failed because slow execution left the startup marching in place and unable to deliver priorities such as desktop and light mode. After aggressive goals caused the “whole car” to fall apart, the founders proved they could rebuild Kortex and more in far less time, then pivoted to Eden, a search engine for memories that finds visual details at exact timestamps. Read on for the mistakes, rebuild test, and Eden’s capabilities.

Transcript

Cortex, the $3 million mistake. One of the things I started thinking about is like what are the problems with Cortex? Like we can't make any step forward because we're just like marching in place. Then the comments coming in and everything else. It's like what are we doing? Let's set these goals. Let's make them really aggressive and see what happe... Read More

Key Insights

• Kortex was a second-brain app that its founders worked on for 2 years before deciding that a drastic product change was necessary. The original concept was not dismissed as inherently bad, but its execution accumulated problems that prevented the founders from reaching their intended vision.
• Slow execution was Kortex’s central operational problem because the company struggled to complete desktop, mobile, and even light mode. The founders believed this meant they had discarded the primary advantage available to a startup, which was the ability to move quickly.
• Aggressive goals exposed Kortex’s weaknesses by revealing that the team could not make meaningful forward progress under its existing approach. The founders described the result as the whole car falling apart, which forced them to reconsider the product and how they were building it.
• Rebuilding Kortex from scratch became credible when the founders tested whether they could recreate the entire product and add more functionality in much less time. About 3 weeks into their stay in Japan, they concluded that this hypothesis had proved true.
• Eden is positioned as a search engine for memories that can retrieve highly specific visual moments. In one demonstration, searching for a pink sweater surfaced every frame containing someone wearing one and returned the exact timestamps for those appearances.
• Kortex’s failure was a series of necessary mistakes rather than a single mistaken idea, according to the founders. They viewed those mistakes as part of the learning required for first-time founders to improve their product judgment, team management, and willingness to take calculated risks.
• The founders’ partnership began through an unexpected conversation in a ghostwriting cohort. Ari publicly asked Matt about the software he was building, followed up with examples of his work, criticized the existing implementation, and was subsequently brought into the project as its technical co-founder.
• Persistence was presented as especially important in startup engineering because founders may need to build products outside their existing expertise. Ari began without web development or front-end knowledge, yet chose the unusually demanding task of building a text editor as an early startup project.

Questions & Answers

Q: Why did Kortex fail as a startup product?

Kortex failed because its team could not make meaningful progress and had become too slow to ship important features. Even light mode was difficult to deliver, alongside needs such as desktop, so the founders concluded that they had surrendered their startup’s main advantage: moving fast.

Q: Why did the founders replace Kortex with Eden?

Kortex had stopped short of the founders’ vision, and its existing development approach left them marching in place. They tested whether they could rebuild the entire product and add more functionality in much less time; 3 weeks into being in Japan, that hypothesis proved true, supporting the pivot to Eden.

Q: What is Eden’s search feature?

Eden is described as a search engine for memories that can retrieve specific visual moments. In the demonstration, searching for a pink sweater resurfaced every frame where someone wore one and provided the exact timestamp.

Q: Was Kortex really a $3 million mistake?

The founders call Kortex the “$3 million mistake,” but they do not consider the idea of building a second-brain app inherently bad. They view the experience as a series of necessary mistakes that helped them improve as first-time founders and reach the point where they could build Eden.

Q: What execution problems slowed Kortex down?

The team struggled to move forward while trying to deliver needs such as desktop and light mode. The difficulty of completing light mode became a clear example of how slowly development was moving, while visible activity had previously made the founders think they were progressing.

Q: How did aggressive goals expose Kortex’s problems?

The founders set aggressive goals to test whether the company could stop marching in place and make real progress. Instead, they said the “whole car fell apart,” revealing that their process and execution could not support the product they wanted to build.

Q: What did the founders learn about startup speed?

They learned that seeing some form of progress does not mean the company is moving forward effectively. By struggling with priorities such as desktop and light mode, Kortex had thrown away what they considered a startup’s one major advantage: the ability to move fast.

Q: What happened to existing Kortex users during the move to Eden?

The founders said Kortex users would receive emails as Eden began rolling out, with an in-app banner available for transferring their information. They also warned that Eden was early access and would not yet match Kortex in areas such as desktop and mobile availability.

Summary & Key Takeaways

• Kortex began as a second-brain app created by first-time founders Matt and Ari, who met through a ghostwriting cohort while pursuing similar software ideas. Their partnership combined Matt’s product focus with Ari’s technical persistence, but early development included inefficient outsourcing and months spent building functionality that Ari believed could be reproduced quickly.

• As Kortex developed over 2 years, slow execution became a defining problem. The team struggled to ship major platform needs such as mobile and desktop, while even light mode was difficult to complete. Aggressive goals exposed that the company was marching in place and had lost the speed that should distinguish a startup.

• The founders concluded that Kortex represented a series of necessary mistakes rather than a worthless idea. They pivoted, rebuilt the product from scratch, and renamed it Eden. Its demonstrated capabilities include searching recorded memories for visual details such as a pink sweater and returning every matching frame at the relevant timestamps.