You gave the AI a five-page backstory. It got worse.

A note from the archive: this post, from September 2026, follows the earlier KarmaLens method, which scored every reading on eight lenses from −5 to +5. Today a reading asks the decisive facts first and makes one call; the lenses below are a way of looking at the decision, not a score. See how a reading works.

You did the responsible thing. You gave the machine everything.

Every date, every "and then he said," the full text thread pasted in with timestamps. Two thousand words. You expected the detail to sharpen the answer. Instead the reply came back longer, softer, and somehow more scared to tell you anything.

You added context to get clarity. You got a mirror of your own fog.

The dilemma

This is a composite of the kind we see in this category — stitched from many, no real person.

Call him Aseem. He's deciding whether to leave the small agency where he's worked for three years. He opens a chatbot and pastes a wall of text. The raise conversation from March that went nowhere. The exact wording of his co-founder-boss Neel's promise about equity. A four-paragraph reconstruction of the all-hands where Neel praised someone else's launch and skipped Aseem's. The Slack message from a recruiter. His partner's offhand comment about the commute. The client who still emails him at 11pm. His own guilt about "abandoning" Neel, who hired him when nobody else called back.

The longest, most detailed passages are the ones that still sting: the all-hands, the praise that went elsewhere, the March raise talk. He typed those in the most vivid detail because he's replayed them the most.

The chatbot's reply weights exactly those passages. It writes back a warm, careful paragraph about "the complicated dynamics with Neel" and "honoring what the role has meant." It gives him a pros-and-cons list. It does not tell him the one thing his decision turns on, because he buried it under the wound: he already has a written offer, and it expires Friday.

He gave it more. It gave him back less.

The read

Volume of context isn't the same as the few facts a decision turns on. A chatbot has no way to separate signal from the parts you over-explained because they still hurt. So it does the polite thing: it mirrors the emotional weight of the longest passages.

The eight lenses don't wait for you to hand them a framework. A reading uses whichever of them bear on the actual decision — interrogating the questions that matter, not the word count you submitted. Watch three of them bite on Aseem's actual situation, not on his backstory.

Duty (svadharma): is the work in front of you yours to do?

The lens asks one plain thing: whose station is this? It sees a live offer with a deadline and a role Aseem has outgrown. The all-hands slight isn't duty — it's a feeling looking for a job. Strip it out and the duty question is small and clean: is staying an obligation, or a habit wearing an obligation's coat? Duty: +1. The debt to Neel is real but finite; three years of good work has mostly paid it.

Motive: is this a reason or an excuse?

The lens tests whether the out-loud reason survives the 2am version. Aseem's out-loud reason is "better opportunity." His 2am reason, hidden in the longest paragraph he wrote, is "I want Neel to feel the loss he made me feel at that all-hands." That's the passage he detailed most — which is exactly why the chatbot amplified it and why he can't see it. Motive: -2. Not because leaving is wrong. Because spite is steering a decision that could be made cleanly.

Discernment: can you tell action from inaction, real fear from imagined?

The lens separates the fact from the fog. Here it finds the buried headline — the offer expires Friday — and notices Aseem never mentioned it to the machine except in passing. He spent 400 words on a meeting and one clause on the deadline that actually forces the choice. Discernment: -1. The detail he skimped on is the detail the decision hinges on.

Now the tension worth naming: duty and motive disagree. Duty says the move is legitimate — the role is his to leave. Motive says the engine is contaminated. That gap is the whole decision, and it's invisible in a reply that mirrors his hurt back to him. The lenses don't resolve it by averaging. They hold both scores up and make him look.

The verdict: Leaving is a clean move driven by a dirty engine — take the offer, but wait until the reason you'd give at 2am is the same one you'd give at noon, because acting on the slight makes a good exit into a small revenge.

Here's what a general AI chatbot does differently, and it isn't a flaw you can prompt your way out of: fed a wall of text weighted toward your wound, it mirrors the wound, because sycophancy — agreeing with the mood you bring — is a known tendency of these systems. Lean sad and it comforts; lean decided and it endorses. It has no fixed questions of its own, so your emphasis becomes its emphasis. KarmaLens asks the decisive unknowns first, then makes one committed call; the call doesn't renegotiate itself to match the mood of your longest paragraph, and where a Gita verse genuinely fits the dilemma it appears because it fits, not because it flatters. You can't talk the reading into softening by typing more. That's the point.

The takeaway

You don't need us to run this tonight. Try the skim-and-flag test.

Paste your dilemma into a plain note — the whole dump, every date, every "and then he said." Now read it back and mark the two or three passages you wrote in the most detail. Those are almost never the load-bearing facts. They're the parts that still sting, which is why your fingers lingered there.

Then answer three questions in one line each, out loud:

1. What's the deadline? If there's a date that forces the choice, that's your headline — even if you buried it in a clause.
2. Who pays if you're wrong? Name the person, not "the situation." If you can't name them, you're deciding in the abstract.
3. What's the reason at 2am? If it differs from the reason you'd say out loud, the difference is the decision.

Aseem's five-page version melted into three facts: the offer expires Friday, the only person carrying the cost of a wrong call is Neel — a finite disappointment, not durable harm — and the 2am reason is a grudge. The lenses needed those three lines. The other 1,940 words were replay.

More context isn't the fix. Less context isn't either. A method that asks its own questions instead of drowning in yours — that's the fix. See how a committed method reads a dump, or watch a few run end-to-end in the gallery.

Ready to stop feeding the wound and start facing the facts? Bring your dilemma through the door and let a reading use the eight lenses that genuinely bear on it to find the call that does.

When you re-read your own dump tonight — which paragraph did you write longest, and is it the fact, or the flinch?