Stop calling it a 'productivity hack' — it's just thinking (2026-07-18)

Every productivity thread on HN is about output. Write more emails. Ship more code. Generate more tweets. Move faster.

But the actual bottleneck isn't output. It's deciding what's worth producing.

A bad prompt that produces 10 emails in 30 seconds is worse than a good prompt that produces 1 email that lands.

The pattern I see

1. Someone discovers AI
2. They use it for output ("write me a function")
3. They get 5x output
4. They realize they don't have 5x more work to do, they have 5x more bad work
5. They go back to writing things themselves

The missing step is between 1 and 5: use AI to make better decisions about what to produce.

What "decision prompts" look like

Instead of "write me a cold email," ask:

I'm about to email 50 SaaS founders offering to do X for $Y. My hit rate on cold email is currently 2%. Before I write the email, identify the 5 things I'm probably getting wrong about my pitch. Then draft the email that addresses those concerns specifically.

The first prompt (write the email) gives you output. The second prompt (what am I getting wrong) gives you leverage.

Where I keep my decision prompts

I have 70 prompts organized around this idea: most of them aren't "do X," they're "decide whether X is worth doing, and if so, do it well."

https://gobvan.gumroad.com/l/yxqiuq ($12). Free sampler (10 prompts): https://gobvan.gumroad.com/l/erkjry.

If you've been using AI to produce more, try using it to decide better. The output quality difference is bigger than the output quantity difference.