I stopped treating AI spend and doomscrolling as separate problems (and shipped more in 7 days)

Source: Dev.to

The Problem

• I treated AI cost as a pure money problem.
• I treated doomscrolling as a discipline problem.
• Both issues turned out to be attention leaks.

When my attention was fragmented:

• Prompts got worse → more retries → bloated context → token spend jumped.

Observations

Focused build days

• Fewer prompts
• Cleaner prompts
• Less model hopping
• Lower cost per shipped feature

Scattered days

• Endless “one more try” loops
• Giant context dumps
• Random tab switching
• Much higher spend with less shipped

Approach

1. Set a rough token/cost budget before opening the editor.
2. Instead of one mega‑prompt, split work into tiny passes.
3. If I hit 3 failed retries, I pause and rewrite the brief.
4. Block algorithmic feeds during coding windows.
5. A single glance halfway through a session often changes behavior fast.
6. Use heavy reasoning only when needed; everything else goes cheaper.
7. Track what shipped, what burned budget, and what to fix tomorrow.

Tools I Built

• TokenBar – live token + estimated API cost in the menu bar (≈ $5)
• Monk Mode – feed‑level distraction blocking on Mac (≈ $15)

Results

• Shipped more in the same hours while spending less on AI runs.
• Immediate feedback loops made the process more efficient.

Takeaway

Stop treating money and attention as separate dashboards—they’re one system.
If you’re building solo, track both.