Tokens. Tokens. Tokens.

Now that I have your attention — let’s talk about why you’re hitting limits faster than you should.

Most people burn through tokens because they type like they’re explaining something to a stranger at a party. Long, rambling, over-explained.

Here’s a quick benchmark worth saving:

• Gemini: ~60-80 words = 100 tokens
•  ChatGPT: ~75 words = 100 tokens
•  Claude: ~70 words = 100 tokens

Different vendors, different math. Same problem if you’re not paying attention.

I was listening to Lupe Fiasco’s Dumb It Down and it clicked. He spent a whole song telling his audience to simplify. Felt like AI advice before AI was a thing.

So here’s your prompt strategy rewritten as Lupe:

You typing way too much, Lu — Dumb it down

Model says you’re near your limit, Lu — Dumb it down

Break it up and keep it tight, Lu — Dumb it down

Them long blocks ain’t it, Lu — Dumb it down

The actual fix is simple: stop treating one session like a novel. Break your work into focused chunks. One task, one context, one goal per session.

Try it this week and let me know what changes.

If the response is good, I’ll make this a biweekly — every two weeks — newsletter of whatever pops in my head at the intersection of the human experience and AI.