// ====== CONFIGURATION ====== const LINK_PROFUNDIZAR = ""; // empty = the "Want to go deeper?" link stays hidden. Set the article URL once published. // ====== CONTENT ====== const FILTROS = [ { n:1, eyebrow:"Filter 1 of 4 · the task", titulo:"Does the purpose require someone to exist — or something to work?", help:"Don't look at the task: look at its real what for. Understanding, learning, discovering, connecting are purposes that happen to you — no one can do them on your behalf (intransitive verbs). Building, selling, organizing produce something separable from whoever does it. Watch out for the hidden intransitive verb: drafting the proposal is working, but repairing the relationship with that client is existing.", opciones:[ {t:"It requires presence", s:"it needs me in the process: I learn, I interact with others, something changes in me while doing it — or by the time it's done", goto:"v_existir"}, {t:"It produces an output", s:"a product that gets used without needing to understand how it was made — like a balance sheet: you read it and it serves you, even if the process doesn't interest you", goto:1} ] }, { n:2, eyebrow:"Filter 2 of 4 · the task", titulo:"Does going through it develop you — or is it exhausted?", help:"A task develops you if you work it at your edge (neither trivial nor impossible), it gives you signal back (did it work? did the client respond?) and it has progression. It's exhausted when doing it one more time changes nothing about how you'd do anything else. And before declaring it exhausted, do the routing: exhausted for whom? What is barren effort for you may be someone else's formative stepping stone on your team.", opciones:[ {t:"It develops me (or someone on the team)", s:"it leaves judgment, vocabulary, or a map of the territory — it's a gym", goto:"v_gimnasio"}, {t:"It's exhausted — for me and for everyone", s:"pure output: nobody learns anything new doing it, and it leaves no knowledge you'll ever use again", goto:2} ] }, { n:3, eyebrow:"Filter 3 of 4 · the market", titulo:"Is the productivity gain real — or is it a story?", help:"Four checks: can you name the specific task and how you'd measure the difference? Does the saving compound at your scale, or is it one-off and dilutes away? Would you adopt it anyway if nobody could ever find out? If the advantage vanishes when no one is watching, it was status, not productivity. Mandatoriness (\"if you don't use AI you're left behind\") is the industry's best-selling product — and the only one that always gets delivered.", opciones:[ {t:"It's real and measurable", s:"a concrete task, a difference that compounds, I'd adopt it in silence", goto:3}, {t:"It's market pressure", s:"I'm adopting it to look modern, out of fear of being left behind", goto:"v_relato"} ] }, { n:4, eyebrow:"Filter 4 of 4 · the system", titulo:"Is there anyone on the other side?", help:"The question changes level: it no longer looks at the task but at the system. What this delegation will produce faster — does it have a nameable consumer, a person whose decision changes because of that output, and a process that transfers it end to end? If you can't name that person, the output won't flow: it will pile up.", opciones:[ {t:"Yes — I can name who consumes it", s:"and which decision changes when they receive it", goto:"v_sana"}, {t:"I don't know who would consume it", s:"it gets produced because \"it has to be produced\"", goto:"v_tumor"} ] } ]; const VEREDICTOS = { v_existir:{ clase:"v-coral", kicker:"Verdict · existence zone", titulo:"In “{T}”, AI doesn't replace: it accompanies — and it's optional.", bajada:"The core of “{T}” happens in you (or in all of you): there is no product to delegate, there is a formation to go through. AI can only operate on the conditions of the act — as an interlocutor that forces you to build your own answer, as a mirror, as clearing of the peripheral weeds, as support for the practice.", naranja:"You can do this without AI. Here it competes with the book, with Google, with the walk, with the conversation with someone who knows — a historian, an engineer, an epistemologist. And it doesn't always win. Using it is only justified when it improves your acquisition of knowledge or the depth of your reasoning. If it doesn't, it's just one more option — discardable without guilt. In this zone, mandatoriness doesn't exist.", reporte:`

The risk in this zone is not that AI can't accompany: it's that it accompanies too well in appearance and simulates the act instead of supporting it. The summary that feels like having understood. The fluent answer that feels like having thought. In functional tasks, fake AI gets detected fast because the output is wrong; here the simulation is undetectable in the short term, because what's missing is invisible: the change in the subject that never happened.

The test is temporal: after real accompaniment, you changed; after a simulation, only the document changed. If in three months you can't defend without the AI what you "understood" with it, there was no understanding — there was text passing through.

The design rule: AI can lower the friction of access (finding, sorting, translating, clearing), never the constitutive friction — the difficulty that IS the act. The effort of understanding is not a cost of understanding: it is the understanding. Example: reading the summary of a book is not having read the book — what develops you is the reading, not learning the argument.

Ask yourself when you're done: did I change, or did only the document change?

` }, v_gimnasio:{ clase:"v-coral", kicker:"Verdict · formative work", titulo:"“{T}” is your gym: go through it.", bajada:"“{T}” produces two things at once: a visible output and an invisible residue — judgment, shared vocabulary, a map of the territory. The market only counts the first. Delegating it preserves the output and silently destroys the formation. AI can assist at the periphery (searching, sorting, preparing), but the core is to be gone through.", reporte:`

Why it matters: the formative residue is exactly what you'll need to use AI well later — the judgment to evaluate what it produces, the vocabulary to ask for it, the map to know what to ask. Formative work manufactures the supervisor of future AI. Automating it is eating the seed corn.

The gym paradox: when machines absorbed physical effort, we didn't stop moving our bodies — we invented the gym: a place where we pay to exert ourselves, because the effort was not a cost of work but a need of the organism. We are at that same moment with cognitive effort. The question is not how to eliminate it: it's where we put the gym.

Careful — not all effort develops: training demands load at your edge, signal back, and progression. If this task no longer meets those three conditions for anyone on the team, run it through the filter again: maybe it's already exhausted. Exhaustion doesn't belong to the task — it belongs to the relationship between the task and whoever goes through it. Example: the first budget you ever built took you a day and taught you where costs hide; the twentieth takes you an hour without thinking. The first was a gym; the twentieth may already be exhausted — for you, not for someone just starting out.

Ask yourself: what stops developing — in me, in us — if a human stops going through this?

` }, v_relato:{ clase:"v-coral", kicker:"Verdict · narrative mandatoriness", titulo:"Nobody is forcing you to put AI into “{T}”: it's a story.", bajada:"The pressure you feel doesn't come from the task — it comes from the narrative. While 40% of layoff announcements invoke AI, 90% of executives surveyed in private declare it hasn't affected employment at their companies: the gap between those two numbers is the measurement of the story. Adopting to look modern is how you end up among the 95% of implementations with no return.", reporte:`

Who the story serves: whoever sells infrastructure (the investment needs demand to justify it), the executive who needs a hard decision to look like natural law ("AI forced us"), the consultant, the media. The only one it doesn't serve is the one who adopts believing it — who pays for the adoption, harvests the failure, and becomes immunized against good future use.

The precedent: when electricity arrived, factories placed a giant electric motor where the steam engine had been and kept the entire layout of pulleys. Decades without productivity gains — until someone redesigned the factory around what electricity made possible. General-purpose technology doesn't pay off until the organization is redesigned around it. New motor, old factory, modern press release.

What to do: nothing, for now — and that's fine. If tomorrow you can name the specific task, the measurable difference, and who consumes the output, run it through the filters again. Competitive advantage is not in having AI (everyone has it almost for free): it's in the design of the system that uses it.

The silence test: would you adopt it anyway if nobody could ever find out?

` }, v_tumor:{ clase:"v-red", kicker:"Verdict · organizational tumor risk", titulo:"Stop: you're about to produce “{T}” faster — for no one.", bajada:"“{T}” passed the task filters — but your system doesn't metabolize it. AI collapsed the cost of producing to nearly zero; the cost of consuming (reading, attending to, integrating, deciding) remains human and fixed. Accelerating production without a nameable consumer doesn't generate efficiency: it deposits output in dead-end tracts — the drive, the inbox, the dashboard nobody opens. Tissue growth without function.", reporte:`

The scene that condenses it: A uses AI to expand three ideas into ten pages, because three ideas don't look like a deliverable. B uses AI to compress the ten pages back into three ideas, because they have no time to read them. Net transfer: zero. Net cost: two prompt sessions and the illusion that communication happened.

Why it's invisible: every unit of tumor tissue registers as productivity — documents produced, tickets closed, reportable initiatives. Activity metrics rise; value stays flat, because the real nodes end up strangled: the proposal that matters buried under forty that don't.

The treatment isn't surgery — it's vascularization: (1) reverse the direction: from push to pull — nothing gets produced without a named consumer; from the weekly report sent "because it's due" to answering when someone asks; (2) transparency as metabolism — if the information lives in a queryable place, most artifacts lose their reason to exist; (3) the same technology is disease or treatment depending on where it points: AI generating artifacts outward fattens the tumor; AI answering questions inward drains it.

The one-move test: name the person whose decision changes because of this output. If you can't, it's tumor tissue — no matter how polished it looks.

` }, v_sana:{ clase:"v-teal", kicker:"Verdict · healthy delegation", titulo:"You can delegate “{T}” to AI — under the five conditions.", bajada:"“{T}” passed all four filters: a functional purpose, exhausted for everyone, real productivity, a nameable consumer. This is the zone where AI replaces without formative or systemic cost. But passing the filters permits delegation; how you delegate decides whether the health lasts.", reporte:`

Execution yes, responsibility no. There is a nameable person who answers for the output as if they had made it. "The AI did it" is inward-facing AI-washing. Output with no one behind it isn't delegation: it's abandonment.
Within the radius of your judgment. Delegating what you couldn't evaluate is faith, not delegation. And delegation is healthy as long as it's reversible: when no one could do or evaluate what was delegated anymore, you've crossed into dependency — a strategic decision that deserves to be made under that name. Example: you can delegate contract drafting only if you'd know how to spot a badly placed clause.
Prior routing through the gyms. Was this task a formative stepping stone for someone on the team? Only what's exhausted for everyone goes to the machine. Skipping this step is how the entry ladder gets broken. Example: the meeting summary that bores you may be the best business school for whoever joined a month ago.
Dividend with a declared destination. Where does the freed-up time go? If you don't decide, it fills up with more production — straight to tumor. "We delegate this SO THAT…": more client, more gym, more edge. Example: if delegating the report frees two hours a week, decide right now what they're for — two more hours with clients, or two more hours of inbox?
Visible to the team. Everyone knows what's delegated, with which instructions, with which validation — not out of bureaucracy: out of metabolism. Silent delegation corrupts everyone else's judgment.

The maintenance test, periodic: could we take it back — and would we know if it's going wrong? As long as the answer is yes twice, the delegation is alive.

` } }; // ====== ENGINE ====== const app = document.getElementById('app'); let actividad = "", paso = -1, camino = []; function rail(actual, total=4){ let h = '`; if(i`; } return h+''; } function pantallaInicio(){ paso=-1; camino=[]; app.innerHTML = `

Start with a concrete activity

What's on your plate?

Write it the way you'd name it in your day: "building the monthly report", "answering client queries", "studying the new tax regime", "putting together the proposal for X".

Four questions: the first three look at the task, the last one looks at the system. No sign-up, no data sent anywhere — and no AI involved, because deciding this doesn't require any.

`; const inp=document.getElementById('inp'); document.getElementById('go').onclick=()=>{ actividad=(inp.value.trim()||"this activity"); mostrarFiltro(0); }; inp.addEventListener('keydown',e=>{ if(e.key==='Enter'){ actividad=(inp.value.trim()||"this activity"); mostrarFiltro(0);} }); inp.focus(); } function mostrarFiltro(i){ paso=i; const f=FILTROS[i]; let ops=f.opciones.map((o,k)=>``).join(''); app.innerHTML = `

${rail(i)}

${f.eyebrow}

${esc(actividad)}

${f.titulo}

${f.help}

${ops}

`; app.querySelectorAll('.btn[data-k]').forEach(b=>{ b.onclick=()=>{ const o=f.opciones[+b.dataset.k]; camino[i]=o.t; if(typeof o.goto==='number') mostrarFiltro(o.goto); else mostrarVeredicto(o.goto); }; }); document.getElementById('back').onclick=()=>{ i===0?pantallaInicio():mostrarFiltro(i-1); }; window.scrollTo({top:0}); } function mostrarVeredicto(id){ const v=VEREDICTOS[id]; const T=esc(actividad); const tit=v.titulo.split('{T}').join(T); const baj=v.bajada.split('{T}').join(T); const filtrosPasados = camino.filter(Boolean).map((c,ix)=>`Filter ${ix+1}: ${c}`).join(' · '); const naranja = v.naranja?`

Important clarification

${v.naranja}

`:''; const link = LINK_PROFUNDIZAR?`

Want to go deeper? Read the full article.

`:''; app.innerHTML = `

${esc(actividad)}

${v.kicker}

${tit}

${baj}

${naranja}

The report for this zone

${v.reporte}

${link}

Your path: ${filtrosPasados||'exit at the first filter'}.

`; document.getElementById('otra').onclick=pantallaInicio; document.getElementById('copiar').onclick=()=>{ const tmp=document.createElement('div'); tmp.innerHTML=v.reporte; const texto=`AI, now? — Analysis of: ${actividad}\n\n${v.kicker.replace('Verdict · ','').toUpperCase()}\n${tit}\n\n${baj}\n${v.naranja?'\nCLARIFICATION: '+v.naranja+'\n':''}\n${tmp.innerText.trim()}\n\n— Generated with AI, now? · Adaptant (adaptant.co)\n"The company doesn't use artificial intelligence. The company is intelligent."`; navigator.clipboard.writeText(texto).then(()=>{ const t=document.getElementById('toast'); t.classList.add('show'); setTimeout(()=>t.classList.remove('show'),1800); }); }; window.scrollTo({top:0}); } function esc(s){const d=document.createElement('div');d.textContent=s;return d.innerHTML;} pantallaInicio();