At 04:11 UTC on November 9, Ryan Dahl, creator of Node.js and Deno, posted on X that it is shocking how quickly software engineering has moved to "100% AI written", and more shocking that engineers carry on as if nothing happened. He said AI writes C++ better than he ever could, but the problem of making sure it works did not magically disappear; teams are simply shipping far more code at higher quality. Within about six hours the post had roughly 1,370 likes, 236 bookmarks, 83 replies, 25 quotes and nearly 90K views. Quote posts split into camps: some say the bottleneck has moved from writing code to verifying it, so testing and review are worth more, while others question the "100%" and "higher quality" claims. This is a personal opinion post; "100%" is not a measured statistic.
Key Takeaways
- ✓Reach: @rough__sea posted at 04:11 UTC on November 9 and within about six hours had about 1,370 likes, 236 bookmarks, 83 replies, 25 quotes and nearly 90K views
- ✓Claim: software engineering has moved to "100% AI written", AI writes C++ "better than I ever could", but making sure it works has not gone away
- ✓Conclusion: engineers are shipping far more code at higher quality, yet "continuing on as if nothing happened"
- ✓Split: some quote posts say the bottleneck has moved from generation to verification, making testing and review more valuable; others question the "100%" and "higher quality" claims
- ✓Caveat: this is a personal opinion; "100%" is not a measured figure, and the post cites no data or specific projects
Key Decision Metrics at a Glance
Turn your technical choice into a development budget
Compare 40 dev plans & simulate token costs vs $20/mo subscriptions
Project Links & Resources
Direct AccessIn-Depth Technical Analysis
Core Background & Industry Pain Point
Ryan Dahl created Node.js and Deno, and his X bio says he is now at Cloudflare. On November 9 he posted that software engineering has turned "100% AI written" in a very short time, and that engineers are carrying on as if nothing happened. The line matches what many teams see: coding agents now handle most implementation, but verification, review and accountability before release still sit with people.
Architecture Highlights & Underlying Mechanism
His argument has two parts. First, AI now writes code better than senior engineers; his example is that it writes C++ better than he ever could. Second, the problem of making sure it works did not magically disappear; teams are now shipping far more code at higher quality. In other words, generation got cheaper and verification became a larger share of the cost. A product manager's quote post put it this way: the cheaper generation gets, the more testing and review are worth, so the next good products should bet on verification.
Authoritative Benchmarks & Measured Comparisons
Not published. The post gives no statistics or project data, and "100%" is a personal judgment, not a measurement. The only verifiable numbers are engagement: about 1,370 likes, 236 bookmarks, 83 replies, 25 quotes and nearly 90K views in about six hours. Pushback is real too: one reply said it is shocking people believe it is 100%, and others questioned the "higher quality" claim.
Developer Hands-on Guide
If your team relies on coding agents such as Claude Code or Codex, shift effort from writing to checking: pair every agent change with automated tests (for example Deno's built-in deno test), make tests, type checks and lint a required CI gate before merge, and ask PRs to state what a human actually verified. As code volume grows, review should focus on behavior and edge cases rather than reading every line of the diff.
Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.