You typed it calm. You feel it at a 9.

A note from the archive: this post, from October 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 open the box and type one tidy line. "Considering whether to raise the missed deadlines with my business partner." Reasonable. Measured. You even deleted the word "again." Meanwhile your actual pulse is somewhere near a nine, and part of you is rehearsing the version where you stand up in the Thursday meeting and let the whole room watch you walk.

The sentence you submitted is calm. The decision underneath it is not.

The dilemma

This is a composite of the kind we see in this category — the confrontation that arrives dressed as a scheduling question.

Nomvula co-runs a four-person studio with Teo. On paper the issue is small: Teo keeps missing the client-handoff deadlines they both agreed to, and she keeps absorbing the slip. She types her dilemma into the box as "whether to bring up the missed commitments with my partner." Flat. Procedural. The kind of sentence you could read aloud in a budget meeting.

What the sentence leaves out: the last slip happened in front of their biggest client, and Teo blamed "the team" — meaning her — to cover it. She smiled through it. She has smiled through it three times. What she actually wants, if she's honest, is not a calmer handoff process. It's for Teo to feel, for one minute, the specific heat of being the problem in a room. She has a whole speech. She has an exit. She has pictured his face.

None of that is in the prompt. She sanded it off, partly out of professionalism, partly because the polished version is easier to send and easier to be told she's right about. The question she typed is "should I raise this." The question she's living is "do I blow this up on purpose, and call it feedback." Those get very different answers.

The read

Here's the trap a general AI chatbot walks straight into: it reads the register of your prose. Calm sentence, calm situation, calm advice — "it's healthy to communicate openly with your partner!" The real temperature never gets read because you never typed it. A reading doesn't mirror your tone. It works the decision itself. Four of the eight lenses bite hard here, and two of them openly disagree.

Duty — is this yours to raise?

Duty asks whether the action is actually required of your role. Nomvula co-owns the studio; the handoffs are hers to protect as much as Teo's. A partner who lets a repeated, client-facing failure go unnamed isn't being gracious — she's quietly abandoning a post. On the plain facts, raising it is squarely her job. Duty: +4. The confrontation, as a thing that needs doing, is clean.

Motive — what's the engine?

Motive asks what's really driving the move, at 1am, not at noon. And here the calm prose falls apart. The thing powering Nomvula isn't process repair. It's the picture of Teo's face. It's the walk-out. The speech she's rehearsed is built to land a wound, not fix a workflow. That's not a neutral fuel. Motive: -3. The same confrontation, aimed at humiliation rather than correction, curdles from the inside.

Notice the collision. Duty says do it. Motive says not like this. A chatbot reading the tidy sentence sees only the duty half and cheers her on — straight into the version she'll regret. The tension IS the dilemma. The act is right; the charge behind it is wrong; and the charge is exactly what she edited out of the box.

Restraint — who's holding the senses?

Restraint asks whether you command the impulse or it commands you. A nine is not a platform you negotiate from. 2.63 traces the short ladder — anger clouds judgment, judgment loses memory, and the person you meant to be is gone before the meeting ends. The speech she's rehearsed is the ladder, pre-built. Restraint: -2. Not because the anger is wrong — the blame-shift earned it — but because an action fired at a nine lands as the nine, whatever words it wears.

Non-harm — what does this cost Teo, and the other two?

Non-harm weighs the damage the move does to people who didn't apply for it. A private, specific "you blamed me in front of the client and I need that to stop" costs Teo an uncomfortable afternoon. A staged public reckoning costs him standing he can't recover, in front of two employees who then run a studio built on a grudge. The reading uses non-harm as one of its instruments here, applied to the move actually on the table — and the move she's rehearsed is the public one. Held against that intention, non-harm points sharply against it. Reroute to the private correction and the same instrument reads the situation differently; the entire swing between those futures is the intensity she didn't type.

The verdict: Raise it — today, privately, as the specific correction it genuinely is — and do not raise it at a nine dressed as a two, because the confrontation is yours to make but the speech you've rehearsed belongs to the anger, not to you.

A general AI chatbot would have answered the sentence Nomvula typed, and leaning toward whatever your words carry is a known tendency of these systems — lean toward confronting and it reasons you toward confronting, lean away and it finds you an exit, mirroring whichever register your prose arrived in. KarmaLens runs the other way: before making the call, it asks the decisive unknowns first, using the Gita's eight lenses as instruments on the decision where they genuinely apply — not to compute a verdict, but to show the weighing honestly — and a Sarathi you can talk the reading through with, one that can stop and ask "what number is this really at for you?" because your emotional register isn't noise around the dilemma, it's a fact the read needs. That weighing doesn't renegotiate itself to match your mood; where a lens finds the motive weak, it says so even if you'd prefer applause, and when a dimension can read two ways it reads the move you're actually poised to make, not the gentler one you wish you were. Any Gita verse the reading cites is there because it genuinely fits the case, not because it flatters it.

The takeaway: the two-number gap

You can run this tonight without us. Take the dilemma you've already written out somewhere calm and reasonable.

1. Write the tone number: on a 1–10, how upset does the sentence you typed sound? (Nomvula's reads like a 2.)
2. Write the true number: how upset are you actually, in your body, right now, about this exact thing? (Hers is a 9.)

Now read the gap. If the two numbers match, your prose is telling the truth and you can trust the answer it earns. If there's a four-point gap or more, you've handed yourself — and anything you ask for help — a decision with the engine hidden. The fix isn't to type angrier. It's one sentence: finish "the version of this I've rehearsed is built to make them feel ___." If the blank is "informed" or "accountable," raise it. If the blank is "small," "ashamed," or "the way I felt," the act might still be right — but you're about to fire it at a nine, and the nine will be the thing that lands. Name the number before you move. See the gallery for how the same confrontation scores at a 2 versus a 9, and browse more deciding-with-ai reads if you keep catching yourself writing the calm version.

Ready to score the decision instead of your tone? Write your dilemma — and this time, put the real number in. What would your dilemma look like if you typed it at the temperature you actually feel?