Non-Algorithmic Capabilities / 非算法能力

Definition

Non-algorithmic capabilities / 非算法能力 are the human capacities that cannot be reduced to searchable procedures, fluent output, or routine tool execution. They include self-definition, desire, taste, judgment, original question-setting, trust-building, resilience, social contact, and the willingness to live through a process before accepting an answer.

Current Synthesis

The concept began as EP275’s workplace answer to AI anxiety: as standard tasks become easier, the valuable layer moves toward judgment, connection, responsibility, and human-scale choice. EP278 extends the same idea into education and parenting. The Yu Hong conversation adds a practice architecture: self-knowledge grows through experiments and review, openness grows through alternative explanations and hypothesis testing, and positive orientation is trained without denying difficult emotion.

Non-algorithmic capability is not anti-AI. It decides how AI should be used. A person with stronger judgment can use AI to prototype, learn, translate, prepare, or explore; a person without enough foundation, purpose, or social reality can become more tool-like because the machine supplies answers before the person has formed a frame.

Key Claims

• AI can increase the relative value of self-definition, judgment, trust, originality, and social skill by making routine cognition cheaper.
• Process remains part of capability formation: reading, interviewing, observing, practicing, failing, and revising build the frame that later judges AI output.
• Education that over-trains obedience, correctness, and emotional suppression weakens students when AI can perform many rule-following tasks.
• Non-algorithmic capability decides AI use rather than rejecting it; it sets goals, review standards, and moments for slowing down or refusing delegation.
• In career choice, these capabilities show up as profession-person fit, resilience, real-world contact, and the ability to build relationships in scenes AI cannot fully enter.
• In media and creative work, upstream question-setting and original observation become more valuable when existing expression is easier to generate.
• Emotional literacy and reflective choice are capability infrastructure: they help a person read internal signals, delay narrowed decisions, test fit in the world, and revise a model without defending identity.

Evidence

• Workplace evidence: EP275 Token 通胀时代，谁还能“不可替代”？丨“人在中流”特别策划01 distinguishes routine tool skills from deciding who one is, what one wants, how to judge, and how to connect with people.
• Human-scale evidence: EP275 Token 通胀时代，谁还能“不可替代”？丨“人在中流”特别策划01 argues that AI use should be evaluated by whether it improves work, life, and value rather than by token consumption or adoption pressure.
• Education evidence: EP278 AI时代不卷专业，卷什么？丨“人在中流”特别策划02 argues that students need resilience, groundedness, exploration, social ability, and personhood when knowledge access becomes easier.
• Parenting evidence: EP278 AI时代不卷专业，卷什么？丨“人在中流”特别策划02 warns against raising children to be machine-like through obedience, lack of emotion, and pure textbook compliance.
• Reporting evidence: EP278 AI时代不卷专业，卷什么？丨“人在中流”特别策划02 treats interviewing, field contact, and upstream question discovery as human work that remains valuable under AI.
• SEL evidence: AI 时代，我们到底该学什么？｜对谈于红：三种不会过时的能力 grounds durable capacity in social-emotional learning, including self-management, relationship skill, critical interpretation, and responsible choice.
• Choice and openness evidence: AI 时代，我们到底该学什么？｜对谈于红：三种不会过时的能力 connects experiments, emotional signals, alternative explanations, prediction review, and hypothesis language through 反思式选择能力 and 智识诚实复盘.

Counterevidence & Qualifications

• The concept is not a claim that all relationship-heavy or creative work is immune to AI; preparation, drafting, search, simulation, and expression can still be assisted.
• Both source episodes are reflective discussions rather than measured labor-market studies, so the concept should guide synthesis rather than stand as a quantitative replacement forecast.
• Non-algorithmic capability can become vague if it is detached from actual practice; the sources ground it in work scenes such as sales, interviewing, learning, parenting, and media reporting.
• The Yu Hong episode is a reflective founder/investor interview rather than comparative evidence that one SEL product or capability framework produces superior long-run outcomes; its happiness, neuroscience, credential, and school-effect claims remain source-scoped.

What Changed

• Added a practice-level account of self-knowledge, openness, emotional literacy, and evidence-based revision.
• Clarified that positive orientation can coexist with accepting negative emotion and delaying narrowed decisions.
• Extended education evidence from professional fit and resilience into explicit social-emotional learning.

Related Concepts

• Human Agency Under AI - supplies the self-directed why and what behind tool use.
• Human Judgment Under AI - turns capability into responsibility for accepting or rejecting output.
• Human-Scale AI Use / 人作为 AI 的尺度 - evaluates AI by human work and life value rather than adoption pressure.
• Human Connection Under AI - captures the trust and relationship side of non-algorithmic capability.
• Career Cognition Education / 职业认知教育 - makes professional fit and occupational reality visible to students.
• Fieldwork As Knowledge Method / 田野作为知识方法 - shows how real contact forms judgment beyond desk knowledge.
• 社会情感学习 - develops self-management, relationship, and responsible-decision capacities.
• 反思式选择能力 - turns self-definition into experiments, value ranking, and review.
• 智识诚实复盘 - makes openness operational through prediction checks and belief revision.