Ten tasks I stopped doing. My week came back.

Hey — and welcome to Sunday Signal.

I used to call this cheating. Now it is the reason I have any weekends left.

There are ten things I do not do myself anymore. Email triage, meeting-notes cleanup, first drafts of anything, research briefs, weighing tools I am choosing between, decision framing, editing what I write, Friday reviews, customer support drafts, Sunday planning.

I still make the calls. I still hit send. But everything between a blank page and a decision is not my job anymore.

The audit that found all ten takes about fifteen minutes. The rule that decided which ones make the list is one sentence, and I use it every time a new task shows up. Both are below.

The busiest hours of your week are almost never the important ones.

I ran the audit for the first time a while ago and it was uncomfortable. Not because it revealed hard work I was avoiding. The opposite. It revealed hours of shallow work. Sorting, summarising, drafting from zero, weighing three near-identical options. Work I had been treating as "my job" because I had always done it.

Sorting is not a job. It is a shape. Same input, same output, over and over. Anything with that pattern can be handed off, and every hour spent doing it is an hour you did not spend on the parts nobody else can do.

The rule I use now is one sentence. If a task is repetitive, or if it is a translation from one form to another (notes to a summary, transcript to action items, three tabs to one briefing), it is not mine anymore.

I shipped a pile of small things this week. Three carousels, a delegation post, a video pipeline that now turns a raw talking-head take into a fully branded short with one command. None of them are what this issue is about. What matters is that all three shipped because the audit made room.

Last Sunday I wrote about retiring channels and every place they still live. The follow-up on that is here in reverse. Once you kill the busy work, you have to actively put the freed hours somewhere on purpose, or the vacuum fills itself back up with more busy work by Wednesday.

Honest half: the audit works. Holding the freed hours is the harder problem, and I have not solved it as cleanly as I have solved the finding.

Week 14 of building in the open. What it actually looked like:

The audit is one page and fifteen minutes. Here it is in full.

Open your calendar for the last seven days. Open your task list. Add anything else you know you did that did not land in either. You want the honest picture, not the polished one.

For each item, ask a single question: what kind of work is this?

H — only a human can do this. It needs your judgement, your relationships, your taste. If someone else did it exactly, the outcome would be worse.

A — this is repetitive or it is a translation from one form to another. Same shape every time, or notes-into-summary, or transcript-into-actions. Any competent AI hands this back at 85% the first time.

B — you need to do the deciding, but the machine can do the drafting or the sorting. You direct, it types.

If a task is genuinely mixed, tag it B. B is the leverage list. Do not tag anything H just because you have always done it. That was my first pass and it was wrong on six of the ten items.

Paste this into your favourite AI, with your tagged list underneath.

Here's my last week of work, tagged H, A, or B. For every A and B, tell me: 1) the single prompt I'd paste to get 85% of it done in one shot, 2) what part I still have to do myself, and 3) which of these three, if I actually handed it off this week, would give me back the most time. Be direct, not encouraging.

You will get back a starter kit for your actual life, in about a minute.

The mistake I made the first time was trying to hand off everything at once. It fell apart in three days.

The version that stuck: pick two tasks in week one, run them daily, keep the exact prompt in a note you can find in three seconds. Add two more in week two. Two more the week after. By week four, six of them run on autopilot and the last four happen because you notice the pattern.

The four highest-utility ones out of my ten. Paste any of them into Claude or ChatGPT without editing.

Email triage. Paste your unread subject lines and the first two sentences of each, then:

Here are my unread emails. For each: reply-now, wait, or say-no. One line each. No fluff.

First drafts of anything. For any post, email, doc, or pitch:

I need to write [describe the thing]. My audience is [describe them]. Give me a rough first draft. Don't polish. Bias to structure over prose.

Friday review. Paste your calendar plus your task list:

Here's what I planned for the week vs what I actually did. What worked, what didn't, what should I do differently next week? Be direct, not nice.

Sunday planning. Brain-dump every commitment for the week you are walking into:

Here's everything on my plate this week. Answer three things. Which should I say no to before Monday? What does a realistic week look like, not the aspirational one? What am I not seeing?

Reply and tell me which one you tried and what came back. If your list looks nothing like mine, that is the interesting part. I want to see it.

Sixteen Saturdays in. The arc so far: ship, constrain, build, compound, patch, subtract, ask, remember, price, staff, equip, doubt, show, keep, retire, and now hand off. Deciding what is yours is a different move than deciding what to stop. Both matter, and they cost different things.

The quiet lesson from this one is that I was not lazy. I was doing the wrong work with the right effort. Delegation was not about doing less. It was about doing less of the wrong thing, so the real thing had room to happen.

If any of this was useful, forward it to one person who thinks their calendar is a virtue.

Quietly, deliberately. One Saturday at a time.

— Kalpesh
Founder, Sun Ventures

P.S. Work with me. Everything above came out of building real things for real use.