PTENES
MODULE 6.1

⏰ Now, in this project, or every week?

So far, you’ve told your digital employee to do things when you asked for. Now the time dimension comes in: the work starts happening on its own, at the right time. There are two tools for this — they look very similar and serve very different purposes. This module makes the distinction clear once and for all.

6
Topics
25
Minutes
Beginner
Level
Foundation
Type
0 of 60%
1

🎯 Two similar tools, very different results

In the previous tracks, everything started with you. You opened the program, wrote the request, waited, looked at the result. We call this kind of work on demand: it happens because someone asks for it, when they ask. It works very well — until the day the same task needs to happen every morning, without fail, including the morning you forgot.

That's where the axis that gives this track its name comes in: the time. The question is no longer just "what do I want it to do?" but also "when e how often should this run on its own?” There are two possible answers, and they have names: scheduled task e loop.

Here’s the module’s trap. Both do the same kind of thing—repeat work without you asking again—which is why many people choose the wrong one and then think “the automation doesn’t work.” It’s not the automation; it’s the choice. One was built for routines that span months; the other, to keep track of something that’s happening now and wraps up in a few days.

💡 Main Concept

Choosing between the two tools isn’t a technical decision; it’s a matter of timing. Ask yourself: will this be over this week, or will it still be happening three months from now?

  • •This will keep happening: scheduled task.
  • •Ends along with the current topic: loop.
On demand
Tracks 0 to 5
Time
The new dimension here
2
Tools
1
Decisive question
2

📅 What is a scheduled task

New here? One scheduled task (in English, scheduled task) is an instruction you write once, and the computer runs it on its own at a time and frequency you choose—every day at 8 a.m., every Friday at 6 p.m., or on the first of every month. You don’t open anything, click anything, or remember anything. When the time comes, it runs.

The most honest analogy is an alarm clock. You set it once, put the device in the drawer in your head, and it rings every day at the scheduled time—including on the days you swore you’d wake up on your own. The scheduled task is your digital employee’s alarm clock: it doesn’t ask if today works; it just runs.

Two characteristics matter from the start, and both come back strongly in module 6.2. First: each run starts from scratch, as if it's the first time—the task doesn't remember what it did yesterday. Second: it has no expiration date. Once configured, it stays there, running week after week until you turn it off.

🔍 Inside

  • Frequency: how often it runs—daily, weekly, monthly.
  • Session: a “conversation” with the agent, from start to finish. Each scheduled task run opens a new, clean session with no memory of the previous one.
  • Duration: undefined. It doesn’t expire on its own.
Mon 08:00 Have 08:00 Wed 08:00 Thu 08:00 Fri 08:00 Sat PC off recovers on startup Sun 08:00 A daily scheduled task over the course of a week you configured it once, on Monday

Caption: notice that the pink circles are identical—each day is a new, independent run with no connection to the previous one. The only different point is Saturday, shown with a dashed line: the computer was off, and the cyan arrow shows the run being recovered later. That detail is covered in module 6.2.

3

🔁 What a loop is

New here? One loop (pronounced "loop", the English word for "laço") is a session — an open conversation with the agent — one that repeats from time to time while it remains active. Each time around, the agent does what you asked again: checks whether anything changed, tries again, adjusts. And unlike the scheduled task, it remembers everything it has already seen in previous rounds, because it never left the same conversation.

The analogy here is a night watchman. He doesn’t leave and come back tomorrow: he stays in the building all night, making the same rounds every half hour. By the third round, he already knows that the back door creaked on the first one— that accumulated knowledge is the advantage of the loop. When his shift ends, he goes home and takes the night’s memory with him.

And the turn really ends. A loop isn't meant to last forever: it expires in a few days—in practice, about three. If the session closes before then (you close the window or the program is shut down), the loop simply dies there, without trying to come back later. This isn't a defect: it's how the tool is designed. A loop exists to track something that's in progress now, not to load an entire lifetime of routines.

✓ When to use a scheduled task

  • ✓"Every morning, summarize what came into my inbox."
  • ✓"Every Friday at 6 p.m., generate the weekly report."
  • ✓"On the 1st of every month, back up my files."
  • ✓Anything you want to keep happening in March next year.

✓ When to use a loop

  • ✓"Keep an eye on this site and let me know as soon as the price drops."
  • ✓"Keep trying to fix this until it stops giving errors."
  • ✓"Stay with this delivery until it’s out for delivery."
  • ✓Anything that ends along with this week's topic.

💡 Practical Tip

If you catch yourself thinking “I'll leave this loop running forever,” stop. That “forever” is the clearest sign that this should be a scheduled task. A loop you don't intend to end is the wrong loop to choose.

4

📊 The table that decides

This is the central piece of the entire track. There are five comparison points — and each one, on its own, can already decide the choice for you. If you remember just one thing from this module, remember this table.

Comparison between a scheduled task and a loop across five criteria
Scheduled task Loop
Where it runs In the desktop app Anywhere: terminal, code editor, desktop app
Memory New session for each run — it doesn't remember the previous one Keeps context: it’s always the same session
Duration Undefined — daily, weekly, or monthly, with no end date Expires in about 3 days
If the PC stayed off Recovers missed runs when the machine comes back online The loop dies if the session closes
Serves to A routine for every day, every week, every month Track something that's happening now

Caption: read from top to bottom, not from left to right. Each row is a question you ask about your situation — and the column with the best answer wins.

Where it runs. New here? O desktop app is a program with windows, icons, and buttons that you install and open with two clicks — like your browser or text editor. It's different from the terminal, that black screen where you type commands instead of clicking. The scheduled task lives in the desktop app because that’s what keeps an eye on the clock. The loop doesn’t have that requirement: it runs wherever the session is open.

Memory. This is the line that causes the most frustration for people who choose wrong. The scheduled task wakes up amnesiac every day: it only knows what’s written in the instruction you left ready for it. The loop carries everything forward — on the fifth cycle, it still knows what it saw on the first. If your situation depends on “compare with what you saw yesterday,” a scheduled task alone can’t handle it (and module 6.4 shows the trick that solves this).

Duration. One is a marathon, the other a 100-meter sprint. The scheduled task has no finish line: you configure it and it runs until you tell it to stop. The loop has a short, known timeframe — about three days — and that's a safeguard, not an annoying limit: it prevents a forgotten session from consuming resources for months.

If the PC stayed off. Here the difference is dramatic. The scheduled task notices it missed its scheduled time and runs as soon as the machine comes back online—nothing is lost, it’s just delayed. The loop has no such safety net: session closed means loop ended, period. No one will resume it for you.

Serves to. This is a summary of the four lines above. A routine that spans the calendar: scheduled. Monitoring something that’s live right now: a loop.

✓ Signs it's a scheduled task

  • ✓You can specify the exact time (“at 8 a.m.”).
  • ✓The instruction is the same every time, word for word.
  • ✓If it fails one day, you can run it later without any problem.

✗ Signs that a scheduled task is the wrong choice

  • ✗You need to react within minutes, not at the scheduled time.
  • ✗What to do now depends on what happened 10 minutes ago.
  • ✗The topic ends this week and never comes up again.
5

🌳 The decision tree

Five rows in a table are a lot to remember at seven in the morning. That’s why the course condenses everything into a single question, and that question gives the module its name: "do I need this now, in this project, or every week?". Three possible answers, three different paths.

I need this now, in this project, or every week? "now" "in this project" "every week" Direct request you write, it does ends immediately (tracks 0 to 5) Loop same session, with memory expires in ~3 days (module 6.3) Scheduled task new session each time with no end date (module 6.2) If you're unsure between the two on the right, ask whether this will still exist three months from now.

Caption: the cyan branch on the left is what you already know — a direct request, with no time involved. The two pink branches are new. Notice they differ in only one way: whether the subject has an end coming soon (loop) or no end (scheduled).

🎯 Main Concept

The tree has three branches because making a direct request is still a valid option. Automation isn’t required: if you do something twice a year, making the request when you need it is less work than setting up a routine.

  • •Now → direct request. No time involved.
  • •In this project → loop. It has a start and finish close together.
  • •Every week → scheduled task. No end is planned.

⚠️ Attention

The most common mistake beginners make is setting up a loop for a permanent routine—it seems to work, then three days later it stops without warning. You only find out when you realize you haven’t received the report in a week. If it’s a routine, schedule it.

6

🗺️ The map of Track 6

With the mental model in place, the rest of the track goes into the details of each tool and then brings everything together into a routine you’ll actually use every day.

2

6.2 — Scheduled tasks in depth

Where it lives, how to create it, the three frequencies, what “a new session each time” means, and how missed runs are recovered.

3

6.3 — Loops in depth

How to start, how to intentionally stop, and what to do when the three-day deadline approaches.

4

6.4 — The Routine That Learns

The trick that restores memory to a scheduled task: a status file that today’s run reads and tomorrow’s run finds ready.

5

6.5 — Determinism

Why a good routine is a predictable routine, and how to write instructions that produce the same result every time.

6

6.6 — Notifications and Restrictive Permissions

How to be notified without being bothered, and why a routine that runs on its own deserves the shortest leash of all.

7

6.7 — Build: Morning Routine

You build, from scratch, the routine that’s ready and waiting for you every day when you sit down at your desk.

🔍 Inside

  • 6.2 e 6.3 cover each tool separately in greater depth.
  • 6.4 e 6.5 solve both weak points: lack of memory and lack of predictability.
  • 6.6 e 6.7 close securely, along with the build you take with you for life.

Quick check (optional): You want a summary of your appointments every morning, indefinitely. Which tool?

✅ Module Summary

✓
The new dimension is time: the question is no longer just “what” but also “when and how often.”
✓
Scheduled task: runs on its own at a fixed time, with a new session each time, no end date, and catches up on missed runs.
✓
Loop: same session repeating with accumulated memory, expires in ~3 days, ends if the session closes.
✓
The question that decides: now, in this project, or every week?

Next module:

6.2 — Scheduled tasks in depth: where to create them, frequencies, new sessions, and recovery