Scheduled Tasks: A Cost Trap
Agent runs tasks on a schedule — accumulate context or rebuild? One choice, 10× cost difference
THE QUESTION THIS PAGE ANSWERS
ANSWER FIRSTWhat is the key idea behind “Scheduled Tasks: A Cost Trap”?
Agent runs tasks on a schedule — accumulate context or rebuild? One choice, 10× cost difference
Make the claim earn its place. Use this page as a decision aid, not a definition to memorize. Connect the idea to one real task, one observable result, and one failure that would change your mind.
Write one question you could answer with evidence after trying this idea.
A conclusion that sounds complete but leaves the key assumption untested.
Context Accumulation Process
How “Reuse Session vs. New Session” changes an answer
“Agent runs tasks on a schedule — accumulate context or rebuild?” shows that a model does not process the “word count” we see. It processes Token pieces. Tokenization affects input length, how much context fits, and how much computation a request consumes.
Length, information, and context are different
As “Agent runs tasks on a schedule — accumulate context or rebuild?” grows, separate three questions: how many Tokens the text becomes, which pieces can change the current decision, and whether older material has fallen outside the context window. Removing repetition is often more useful than simply making the window larger.
Keep what can change the decision
Use “Agent runs tasks on a schedule — accumulate context or rebuild?” as an A/B test: keep the same question while removing repeated background, compressing format, and trimming irrelevant history. Compare answer quality, latency, and Token count.
From “Reuse Session vs. New Session” to “Drag "Days Running" to See the Difference”
“Reuse Session vs. New Session” grounds the problem in “Reuse Old Session Run 1 Run 12 Run 24 $7.20 24h Total Cost $0.60 Last Run Cost New Session Each Time Run 1 Run 12 Run 24 $0.72 24h Total Cost $0.03 Cost Per Run”. “Drag "Days Running" to See the Difference” then moves it toward “成本计算器 Days Running: 7 Executions Per Day: 24 times $50.4 Reuse Session Total Cost $5.0 New Session Total Cost 10x Cost Multiplier”. Together, they show that the lesson is not just a conclusion to remember, but a claim with conditions.
Carry the judgment into the next situation
For long text, keep what can change the conclusion before compressing format and history. A larger context is worth its cost only when the added information is useful.
- “Reuse Session vs. New Session”: Reuse Old Session Run 1 Run 12 Run 24 $7.20 24h Total Cost $0.60 Last Run Cost New Session Each Time Run 1 Run 12 Run 24 $0.72 24h Total Cost $0.03 Cost Per Run
- “Drag "Days Running" to See the Difference”: 成本计算器 Days Running: 7 Executions Per Day: 24 times $50.4 Reuse Session Total Cost $5.0 New Session Total Cost 10x Cost Multiplier
- “Why Reusing Sessions Causes Snowballing Costs”: Context Accumulation Process 1 Run 1: context = System Prompt + current conversation ≈ 2,000 tokens 2 Run 2: context = all of Run 1 + current conversation ≈ 4,000 tokens 3 Run 10: context = all of Runs 1–9 + cu…
The final “Why Reusing Sessions Causes Snowballing Costs” brings the discussion to “Context Accumulation Process 1 Run 1: context = System Prompt + current conversation ≈ 2,000 tokens 2 Run 2: context = all of Run 1 + current conversation ≈ 4,000 tokens 3 Run 10: context = all of Runs 1–9 + cu…”. The useful thing to carry forward is knowing which judgments must be revisited when input, scale, or risk changes.
I turned one judgment from this article into a small experiment I could run today. Knowing what to observe next is more useful than simply remembering the conclusion.
After reading this, I first looked for the conditions behind the idea instead of copying the method into a project. That order made the later trade-offs much clearer.
When this judgment reaches real work, which constraint should be added first? I am curious which step matters most between reading and the first practical attempt.
No discussion on this article yet.