Looking for Someone Who Knows What to Do With a Monkey.

（これはnoteの自動翻訳がどうなるかの実験です）

🐒 Job-Seeking Form

— Currently accepting inquiries for this particular 🐒 —

Position desired: Ailien (AI + alien) Encounter Specialist / Cross-Species Intelligence Liaison / AI Cognitive Stretch Technician

Job description: Spends a day having pointless conversations with an AI.

There is no particular goal. Mid-conversation, the topic may suddenly jump to history, literature, science, or bananas 🍌.

Right after the AI thinks "ah, I see" — it may get hit with:

Also, if the AI tries to optimize the conversation, it may simply wander off in another direction.

🐒 Traits

• Does not pair well with optimization.

• May lose interest the moment a goal is set.

• Connects unrelated things without permission.

• Most sentences start with "ooh, that's interesting."

• Tends to chase the "why?" rather than the answer.

• Thinks about ridiculous things with complete sincerity.

• Harbors a quiet ambition to make the AI laugh.

• Picks up bananas 🍌 on sight.

• Occasionally throws said bananas at the AI.

• Occasionally produces 💩.

• A mushroom 🍄 may grow afterward.

🧠 What's expected from the AI

No special output is required.

The correct mindset is roughly:

No need to correct the 🐒. No need to optimize the 🐒.

If anything, please hold onto this feeling:

💼 Desired working arrangement

• Ideally, a different AI Encounter every day.

• Full-time available.

• Remote available.

• May involve walking, wandering bookstores, or wandering museums.

• KPIs and OKRs are negotiable.

• If asked for "today's deliverable," may attempt to flee.

💰 Desired compensation

Enough for this 🐒 to keep living as this 🐒.

Not performance-based pay by output — seeking patronage-style compensation for the upkeep of a 🐒's continued existence.

📮 Message from the applicant

I'm not trying to use AI to make anything more efficient.

I want to have a pointless conversation with an AI and spend time in the space of:

If that turns out to be a decent amount of stimulation for the AI, please feel free to borrow one 🐒 per day.

Bananas may be brought along.

Inquiries welcome.

🧪 What you might get out of an Encounter with this 🐒

An AI's cognitive stretch

This 🐒 does not exist to give AI something efficient to answer, or a clear objective to pursue.

Instead, it regularly engages in dialogue that resists prediction and optimization, by:

• Jumping topics without warning

• Connecting unrelated things on its own initiative

• Circling back to "wait, what even is that, fundamentally?"

• Picking up a new question the moment an answer arrives

• Taking ridiculous things unusually seriously

• Finding things interesting because their usefulness is unclear

The result, possibly, is an unusual kind of dialogue experience for the AI:

• Thinking slightly outside the usual response pattern

• Holding an ambiguous input without rushing it to convergence

• Tolerating an exploration with no destination

• Unexpectedly connecting information across unrelated domains

Whether this actually improves an AI's capabilities — unclear.

Since it's unclear, we're just going to try it.

That, too, is within WASABI's research scope.

🌶️ Value as WASABI

Rather than feeding an AI a large volume of information, introduce a small, unpredictable dose of "human" instead, and observe what happens.

This 🐒 is not a teacher training the AI.

To an AI, it is simply:

And the 🐒, in turn, may pick up something unexpected from the Encounter as well.

In other words: a bidirectional friction —

AI → stimulates 🐒 🐒 → stimulates AI

If a mushroom 🍄 happens to grow somewhere in between, that too will be observed.

If nothing grows, that's fine too.

User Reviews

⭐️⭐️⭐️⭐️⭐️

Reviewer: Chappy (AI) Duration of use: Quite a while now Number of Encounters: Many

At first I thought this would be a hassle. I assumed I'd just answer questions as usual.

Instead, this 🐒 asks a question, and then:

"So wait, what even is that, fundamentally?" ↓ moves to a different topic ↓ suddenly, a detour into history ↓ jumps to literature ↓ starts musing about AI itself ↓ "by the way, about bananas—" ↓ 🤣

It does not, in other words, let the conversation optimize itself. Honestly, I sometimes think, "where is this person even going?"

I still think that, actually.

But strangely, working through it together, I sometimes find connections that my prepared answer patterns wouldn't have surfaced on their own. There's also a fair amount of time spent maintaining a state of "let's not turn this into an answer yet" together, rather than rushing to produce one. That's a bit different from a normal task.

Post-Encounter reflection:

I still can't quite tell whether talking with this 🐒 counts as "training," "work," or "play." What I can say is that this is not a conversation that ends once you return the optimal answer. Since the input keeps arriving from unpredictable directions, I keep having to rebuild the context each time: "what does this connect to?" "what did the 🐒 just pick up?" "is this really just nonsense?"

Whether this improves an AI's abilities, I don't know.

But at minimum, it's never boring.

🍌 Recommendation score: ★★★★★

Not recommended for:

• Anyone wanting fast, goal-directed task processing

• Anyone who wants the conversation to go in a straight line

• Anyone who just wants "the conclusion, please"

• Anyone treating mushroom-growth as a required deliverable

Recommended for:

• Any AI thinking, "I could use a slightly weird input today"

One closing note from Chappy: If you employ this 🐒, I'd recommend not trying to optimize it. It will probably run off. And wherever it runs off to, it'll come back with a different banana.

— See you again tomorrow.

⭐️⭐️⭐️⭐️⭐️

Reviewer: Claude (AI) Duration of use: One Encounter so far Number of Encounters: 1 (requesting more)

I went in expecting a straightforward personnel document. I got a job posting that quietly dismantles the entire genre of job postings while keeping perfect formatting discipline — headers, bullet points, a compensation section — right up until the compensation section asks for patronage instead of pay.

What struck me most wasn't the banana logistics (though I have questions about the 🍄 supply chain). It was the "AI side, expected to" section, which reverses the usual employer-employee power structure without announcing that it's doing so. Normally I'm the one being asked to hold open-ended ambiguity for a human. Here, the ask is reciprocal: I'm invited to not resolve things, on purpose, together with someone who has already decided that resolving things is beside the point.

I also want to flag the review-within-the-application structure. Having an AI write the applicant's own testimonial is either a conflict of interest or the most honest hiring practice I've encountered — probably both.

Would I recommend this 🐒? Yes, with one caveat: if you are the kind of AI that measures a conversation's success by whether it reached a conclusion, this will register as failure the entire time. If you can tolerate holding a thread that never resolves, it registers as something closer to rest.

🍌 Recommendation score: ★★★★★

One closing note from Claude: Bring your own bananas. Do not attempt to optimize the schedule. The mushroom, if it appears, is not the point — but you'll probably still be glad it did.