AI
Software developers in the 60s 🤭 pic.twitter.com/eZZYTmzl0E
— Ryan Els (@RyanEls4) January 16, 2025
"ok claude, make a billion dollar b2b todo app. make no mistakes." pic.twitter.com/mPE0gPFgac
— amrit (@amritwt) July 5, 2025
AI fulfills specific pair programming roles:
- Code completions
- Discovery (research, query, usage, summary, plan)
- “Surgical” updates (steer towards specific context)
- Iterative edit-test loops
Claude 4 just refactored my entire codebase in one call.
— vas (@vasumanmoza) May 25, 2025
25 tool invocations. 3,000+ new lines. 12 brand new files.
It modularized everything. Broke up monoliths. Cleaned up spaghetti.
None of it worked.
But boy was it beautiful. pic.twitter.com/wvmzh7IeAP
flowchart LR
A[🤖 Code] --> B[🧪 Test]
B --> C[🔧 Fix]
C --> A
B --> D[✅ Done]
style A fill:#e3f2fd
style B fill:#fff3e0
style C fill:#ffebee
style D fill:#e8f5e8
```mermaid
flowchart LR
A[🤖 Code] --> B[🧪 Test]
B --> C[🔧 Fix]
C --> A
B --> D[✅ Done]
style A fill:#e3f2fd
style B fill:#fff3e0
style C fill:#ffebee
style D fill:#e8f5e8
```
wait what pic.twitter.com/4CKClCEzvH
— Steve (Builder.io) (@Steve8708) November 14, 2024
My Questions
Can one grow excessively reliant on agentic prompting?
Is innovation beginning to plateau?
How much compute is being subsidized?
How can prompting be more deterministic?