Author here. I'm a patent attorney who also builds with agentic coding tools.
The backstory: I used a fitness app I've admired for years as design inspiration and built a working app for a completely different vertical in about five hours. That made the defensibility question very concrete for me — if I can do that in an afternoon, what does "shipping faster" actually protect?
The article walks through why IP strategy (specifically: instructing your coding agent to surface patentable ideas during development and filing provisional patents quickly) becomes relatively more valuable as the cost of building software drops.
The practical piece is an inventions.md workflow — a markdown file your agent appends to when it spots something that might be patentable, with a lightweight weekly triage process.
Honest questions, I'm not trying to troll -- does this stuff even matter in a practical sense? Like, what's the minimum threshold (of revenue?) to where litigation even becomes a possibility?
It's becoming more and more trivial to reverse-engineer via LLM, which could then be used to clean-room a clone product. Wouldn't the value then lie in the brand, the quality of execution and support, etc? What if you get cloned by somebody in Bratislava or China or whatever? Fighting this stuff in court (or courts, plural!) across international borders feels like a quagmire at best and a terrible waste of time and money at worst.
I guess without patents, somebody could clone you and then file for the patents themselves, and sue you into oblivion -- that would be hell on earth obviously lol.
But in the end, what if you just showed up with your product and did an incredible job and ignored the clones and simply "let the best man win"?
All valid concerns! Any business is unlikely to stand or fall based on its patents. Brand, quality of execution, and the team are all where the big part of the value lies.
You're also absolutely right that litigation is expensive, and most early-stage companies will never file a lawsuit. But patents actually create a lot of value for start-ups, and in practice, leverage can come from fundraising and acquisitions, competitive deterrence, and, hopefully, preventing someone else from filing a patent that blocks yours. Very few patents are actually enforced, but they do add value in other ways.
Re the clean-room / LLM reverse engineering point: patents don't protect code, they protect mechanisms. If someone clean-rooms a different implementation of the same underlying technique, the patent still covers it. That's actually the whole point of a patent vs. copyright.
Just executing well and hoping the best person wins is philosophically correct and justifiable. Unfortunately, though, sometimes it's not an even playing field. A big player with significant resources can quickly squash an early-stage start with fantastic ideas. I also say a big part of inventing is just identifying the problem. I've worked with a lot of startups that have really identified a gap in the market or a particular need and worked hard to develop a solution. Just to have a quick follow slipstream on all of that domain experience and hard work can be incredibly frustrating.
Maybe not the most satisfying response, but hopefully of some value. Thanks for commenting.
Author here. I'm a patent attorney who also builds with agentic coding tools. The backstory: I used a fitness app I've admired for years as design inspiration and built a working app for a completely different vertical in about five hours. That made the defensibility question very concrete for me — if I can do that in an afternoon, what does "shipping faster" actually protect? The article walks through why IP strategy (specifically: instructing your coding agent to surface patentable ideas during development and filing provisional patents quickly) becomes relatively more valuable as the cost of building software drops. The practical piece is an inventions.md workflow — a markdown file your agent appends to when it spots something that might be patentable, with a lightweight weekly triage process.
Honest questions, I'm not trying to troll -- does this stuff even matter in a practical sense? Like, what's the minimum threshold (of revenue?) to where litigation even becomes a possibility?
It's becoming more and more trivial to reverse-engineer via LLM, which could then be used to clean-room a clone product. Wouldn't the value then lie in the brand, the quality of execution and support, etc? What if you get cloned by somebody in Bratislava or China or whatever? Fighting this stuff in court (or courts, plural!) across international borders feels like a quagmire at best and a terrible waste of time and money at worst.
I guess without patents, somebody could clone you and then file for the patents themselves, and sue you into oblivion -- that would be hell on earth obviously lol.
But in the end, what if you just showed up with your product and did an incredible job and ignored the clones and simply "let the best man win"?
All valid concerns! Any business is unlikely to stand or fall based on its patents. Brand, quality of execution, and the team are all where the big part of the value lies.
You're also absolutely right that litigation is expensive, and most early-stage companies will never file a lawsuit. But patents actually create a lot of value for start-ups, and in practice, leverage can come from fundraising and acquisitions, competitive deterrence, and, hopefully, preventing someone else from filing a patent that blocks yours. Very few patents are actually enforced, but they do add value in other ways.
Re the clean-room / LLM reverse engineering point: patents don't protect code, they protect mechanisms. If someone clean-rooms a different implementation of the same underlying technique, the patent still covers it. That's actually the whole point of a patent vs. copyright.
Just executing well and hoping the best person wins is philosophically correct and justifiable. Unfortunately, though, sometimes it's not an even playing field. A big player with significant resources can quickly squash an early-stage start with fantastic ideas. I also say a big part of inventing is just identifying the problem. I've worked with a lot of startups that have really identified a gap in the market or a particular need and worked hard to develop a solution. Just to have a quick follow slipstream on all of that domain experience and hard work can be incredibly frustrating.
Maybe not the most satisfying response, but hopefully of some value. Thanks for commenting.
So? Differentiators should have always been about customer service. Most products are enshittified, anyway.