TL;DR: Humanity spent exactly 100 years solving the problem of "Software." With the release of Claude Opus 4.6 and Codex 5.3, we have officially solved coding. The recent crash in SaaS stocks isn't a market error; it's a correction of reality. Legacy SaaS companies are now vulnerable to 3-person teams who can clone their products in a week. The era of "Software as a Service" is ending; the era of "AI as Software" has begun.
James here, CEO of Mercury Technology Solutions. Tokyo - February 9, 2026
It took humanity exactly one century to "solve" software. From the early punch cards to cloud-native architectures, the smartest brains on Earth dedicated themselves to one goal: building digital worlds with code.
As of Q4 2025, that mission is complete. We are witnessing the final stage of the "Software Era" as we know it.
The Panic in the Markets
If you have watched the markets this week, you’ve seen the bloodbath in SaaS stocks. The trigger? The god-tier duel between Claude Opus 4.6 and OpenAI Codex-5.3. These models have demonstrated coding capabilities so advanced that Wall Street is terrified. The Fear: If AI can write software instantly and infinitely, does the business model of selling software still exist?
Jensen Huang (Nvidia) says the sell-off is absurd. He is technically right, but he is only telling half the truth.
- The Vendor's Truth: He needs software to remain valuable so people keep buying GPUs.
- The Brutal Truth: Legacy SaaS is in trouble. A massive, entrenched software company can now be cloned by a 3-person startup using AI to rewrite the entire codebase in weeks.
The Irony of the Singularity
There is a dark humor in this moment. Software Engineers spent 100 years automating everyone else’s jobs. It turns out, the first major profession to hit the AGI Singularity—where AI perfectly replaces human output—is Software Development itself.
It is a bitter pill for many. The goal was always to make ourselves obsolete, but actually achieving it feels like a punch to the gut. But here is the reality: "Coding" is no longer a competitive advantage. AI codes faster than you. It doesn't sleep. It doesn't make syntax errors.
The Pivot: From "How" to "What"
If your value proposition is "I can translate a spec sheet into Java," you are unemployed. But if you can pivot from Coder to Architect, this is the Golden Age.
Engineers must immediately shift their focus from "How to code" to "How to architect" and "How to talk."
Here are the three layers of value in the AI Era:
- Definition Precision: Can you deconstruct a fuzzy business problem into logic so precise that the AI understands the requirement instantly?
- Architectural Breadth: Once the AI builds the modules, can you design the high-availability, scalable cloud infrastructure to hold them together?
- Integration Depth: Can you translate raw model capability into actual business value for a retailer or a bank?
The Punishment of the "Old Way"
The market crash is punishing "Old Software Thinking." It is punishing tools that have Features but no Intelligence. (I called it digital paper)
AI-Native is the new standard. Future software that isn't backed by an AI model will look like a smartphone that can't connect to the internet—a brick.
We are moving from SaaS (Software as a Service) to "Service as a Software." The software is fluid. It is generated on the fly. It is bespoke.
Conclusion: The Strong Get Stronger
Do not let the volatility scare you. The civilization of software isn't disappearing; it is being Reforged. We no longer need to reinvent the wheel. We stand on the shoulders of giants (Claude, OpenAI) to solve problems that were previously too complex to touch.
Coding is becoming a "General Education" skill—like driving a car. In 1900, you hired a specialist to drive. In 2000, everyone drove. In 2020, you hired a specialist to code. In 2026, everyone codes.
Humanity has solved the "Software Problem." Now, we can finally focus on the "Value Problem."
Mercury Technology Solutions: Accelerate Digitality.
