Find agent skills by outcome

Bu sistem

*Arrival* (2016) filmindeki bir gözlemle başladı:

During testing

the final Socratic chain asked:

Geleneksel yaklaşımlar sabit metrikler kullanır: duygu skorları

etkileşim oranları

LLMs are increasingly used to generate questions: in therapy bots

coaching apps

| Human as final authority | System shows

human decides |

These require judgment

empathy

| Elenchus (examination) | Chain 5: Apply lens

find tensions |

- Socratic: it asks

it doesn't conclude

- Human-final: shows candidates

human decides

- A mirror

not a judge

- A discovery method

not a scoring algorithm

toward examining internal meaning

assumptions

- Introduces value trade-offs (money vs belonging

loss vs gain)

Traditional approaches use fixed metrics: sentiment scores

engagement rates

Louise

the linguist

When humanity encounters aliens

the military asks: *Are you hostile?*

print(json.dumps(results

indent=2

{role: human

content: I need to ask my boss for a raise.}

{role: assistant

content: What achievements would you highlight?}

{role: human

content: Can you help me write an email?}

{role: assistant

content: What was different then?}

{role: assistant

content: Sure

{role: assistant

content: When did you last feel like you were growing?}

{role: human

content: Maybe three years ago