The economics of hiding code have changed
For most of software history source code was treated as the asset, and companies guarded it because writing it was slow and expensive and hard to reproduce. A competitor who wanted your product had to hire engineers and spend months rebuilding what you had already built, so the secrecy was rational; the code carried real scarcity, and scarcity was worth protecting. That assumption no longer holds. The cost of producing working code has fallen so far that what once took a team of engineers a full quarter can now be drafted in a few days with the help of capable models.
As the cost of writing code approaches zero the value of merely possessing it approaches zero with it, since a thing that is cheap to make is cheap to remake. A competitor who wants to copy your implementation can now do so at a fraction of the historical cost whether you publish it or not. This reframes the whole question of secrecy, because if your code can be reconstructed by anyone with a model and a clear goal then guarding it produces very little protection while you go on paying the full cost of keeping it closed.
Replication was always coming
Closed source assumes that hiding the implementation slows imitation, and in a world of expensive engineering it genuinely did, but the moat that secrecy buys has grown thin and keeps thinning. Think about what actually happens when a product succeeds: competitors observe the behavior, infer the structure, and rebuild it, and they were doing this long before models existed; the only thing that has changed is the speed. The window between your release and a credible clone has collapsed from years to weeks, and secrecy does not close that window so much as delay the moment you admit it is already open. If imitation is going to happen regardless then the defensive value of hiding becomes a cost you carry for almost nothing in return.
The real strength was never the code
Code is the description of what your company does rather than the doing of it. The durable advantage of a company lives in its operations, in how reliably you run the system and how you respond when something breaks and how you handle the edge cases that never appear in any repository. None of that is contained in the source, and you can read every line of a payments company and still be entirely unable to run a payments company. This is why publishing your code costs you far less than it first appears, because the part a competitor cannot copy by reading is the part that actually matters, and that part was never sitting in the file. What you give away is the cheap layer, while the expensive layer stays exactly where it was.
Open code builds trust
When code is open it can be analyzed, and anyone can verify what it does instead of trusting a claim about what it does. For any company whose product touches money or health or safety or compliance that is a direct asset, because it turns a promise into something a customer or a regulator can actually check. There is a second effect that matters more over time, because publishing your implementation is itself a statement of confidence; it says that you are willing to be examined, that your advantage does not rest on concealment, and that you would rather compete on how well you operate than on what you manage to hide. A company that opens its code is signaling that its strength is in execution, and that signal is hard to fake, since a company with a weak operation has every reason to keep the doors shut. Trust earned this way compounds. Customers who can verify your system extend more of it to you, regulators who can inspect your controls move faster, and engineers who can read your work form a clearer view of your competence that then travels on its own.
What this points toward
Put these forces together and the direction becomes clear. The cost of writing code is falling, the protection offered by secrecy is falling with it, and the advantages that remain, operations and trust, are the advantages that openness strengthens rather than erodes. The rational response is to publish, and we should expect to see far more open source in the coming years, not as an act of generosity but as a recognition that the old reasons for hiding have quietly expired. The companies that understand this early will spend their secrecy budget on the one thing that still pays, which is running their operation better than anyone else can. Open source is in again, because the cost of keeping code closed has never been higher relative to what that secrecy now actually buys you.