Scanning the immediate area — high-demand initiatives are one or two hops away, and asking is the bridge

The most useful initiatives — the rising teams, the unfunded-but-real projects, the operator who has capacity for exactly one more apprentice this year — are not, by selection, in your immediate environment. If they were, you'd already be on them. The thing you're scanning for is, by definition, what your current vantage doesn't show.

That sounds tautological. It isn't. It is a specific structural claim about where your option set comes from, and where the next one has to come from.

The selection effect on what you can see

Your visible environment is not a random sample of the world. It is the set of initiatives your current network has already routed to itself. The colleagues who are around you are the ones who passed the same filter you did. The opportunities you hear about are the ones your existing channels carry. The ceiling you can picture is the tallest chair in the visible room — the Possibility Models doc calls this the option-set problem, and the stable-job entry walks through what it looks like in retail.

The high-demand initiatives sit just outside this radius. Not because they are exotic — most of them are mundane to the people inside them — but because the channel that would carry them to you doesn't exist yet. The selection effect cuts both ways. Your environment is reliable about everything except the thing you most want it to surface.

The corollary is uncomfortable: scanning harder inside your current radius is unlikely to bring results. One of two things is true, and usually both. Either the high-demand initiative is genuinely not in your radius — it's in someone else's, two hops away, and no amount of intensified search inside your radius will surface it. Or it is in your radius and your mindset is filtering it out before it registers as an option — the constraining memes ("that's not for people like me," "I'm not ready for that," "they wouldn't take me anyway") remove the move before the search even runs over it. More LinkedIn time, better job alerts, a more thorough read of the trade press do nothing to either case. They intensify a search where the answer is either absent or invisible by construction.

The fix is therefore double-sided. Extend the radius by one or two hops and test the constraining memes by acting on whatever the search did surface. Both moves require the same thing: speech aimed outward.

One or two hops, not more

Granovetter's weak-ties point, decompressed: most consequential opportunity moves through acquaintances, not strangers and not your inner circle. The inner circle shares your environment — they're scanning the same room. Strangers can't vouch — the request from a stranger arrives without trust, which is the substrate everything else rides on.

The hop count matters. Two hops is roughly the limit at which a warm introduction still carries signal. "My friend Anna, who you trust, says this person is worth ten minutes" compresses years of trust accumulation into a single sentence and the recipient parses it accordingly. Three or four hops in, the chain attenuates: the introducer doesn't know you well enough to vouch, the recipient doesn't know the introducer well enough to act on the vouch, and the request degrades to a cold pitch dressed in cold-pitch's clothes.

So the operational target is small and specific. Not "expand my network." Two hops, well-chosen, named.

Networking is exposure-engineering, not a soft skill

The word networking carries a smell that the project should reject. It pattern-matches to event attendance, business cards, performative LinkedIn endorsements — the surface behaviours of an attention-economy operator pretending to be useful. That version is real, and it is largely Path A in the sense the emotional-vs-structural entry means it: anxiety management dressed as career investment, with no endpoint and no observable structural output.

The structural version is different. Networking-as-information-pump asks specific questions of specific operators and observes what is actually moving in adjacent rooms. The output is concrete: one named initiative you didn't know existed, one operator who would take a real meeting, one apprenticeship slot that opens because you knew to ask before it was advertised. This is the same exposure mechanism the possibility-models doc names — the difference is that here you are engineering the exposure deliberately, retail, one operator at a time, instead of waiting for it to arrive.

The Path A/B filter applies cleanly. If your networking activity does not resolve to a named initiative, a named operator, or a named next step within a few iterations, you are running the emotional version of the move and the energy is leaking into a market you didn't mean to enter.

Speech as the protocol

This is the part the project has not addressed directly until now. The bridge across the one-to-two-hop gap is speech. Specifically, the ask. And the ask has a shape — not a script, but a structure that lowers the cost of granting it. Most people who don't ask, don't ask because they have not internalised this shape and so default to a version that is expensive to grant and easy to refuse.

The components, named:

A specific target. "Do you know anyone working on X?" is a question that can be answered yes, no, or let me think. "Do you have any advice for someone in my situation?" is a question with no graceful answer — the person asked has to either invent advice on the spot or decline and feel rude. The first is cheap to grant. The second is expensive. The yes rate moves accordingly.

Honest cost-disclosure. Name the favour size before asking. "This is a five-minute question, no follow-up needed" lets the recipient price the request and grant it without dread. "Could we get coffee sometime" hides the cost — they don't know if you mean ten minutes of advice or six months of mentorship — and the cautious default is to decline. People decline because they cannot calibrate the ask, not because they don't want to help.

A visible return path. Make the reciprocity readable. Sometimes that's an offer — "happy to send you my notes on Y in exchange" — and sometimes it's an honest disclaim — "I don't expect to repay this one directly; I'll pay it forward." Both are clean. What is not clean is leaving the recipient to wonder what implicit obligation they're entering.

A graceful exit. Build the no-cost refusal into the ask. "If you're slammed, no worries at all — happy for a one-line decline." This raises the yes rate, counter-intuitively, because the recipient stops bracing for a confrontation. People say yes more often when saying no is free.

An introducer when you can. Two hops works because hop one carries trust into hop two. Whenever the path is Anna → operator, route the ask through Anna. The cost of asking Anna for the introduction is small; the cost to the operator of receiving Anna's vouch instead of your cold message is enormously smaller than it appears.

The protocol is not a manipulation. It is the dual of the cost it imposes. A well-shaped ask is one the recipient can grant in under a minute, refuse without guilt, or route to the right person without effort. That is what makes it ethical and what makes it work — those are the same thing.

Why most people don't ask

Two failure modes, and both are diagnosable.

Status protection. The ask exposes you as not-already-in. The ego reads not-already-in as a loss, even though every operator who is currently in was, at some point, not-already-in. The status cost is felt acutely; the silent cost of not asking compounds invisibly across years. The asymmetry is the trap.

Inherited constraining memes. "Don't impose." "Make it on your own." "People like us don't ask people like them." The starter-loop entry names these as the constraining class of memes — they don't drain energy, they prevent you from spending it. They survive in stillness because the move they prevent is never tested. The first ask, granted, kills the meme. Until then it stands.

The cost of the ask is almost always overestimated by the asker and almost always trivial to the asked. Operators who are themselves a hop away from where you want to be have, in most cases, made the same ask of someone else within the last year. They remember being on your side of the request. The empathy is closer than it appears.

A small protocol, runnable this week

Run the starter loop over the network surface, with the same five steps and a tighter scope:

1. Reflect. What initiative would I want to be one hop away from in six months? Name it. Vague answers indicate the possibility-models work is incomplete and should run first.
2. Design. Three asks. Three operators. Two hops maximum. For each, who is the most plausible introducer.
3. Filter. For each ask: specific target, disclosed cost, return path, graceful exit, introducer named. If any are missing, fix the ask before sending it.
4. Execute. Send them. Not perfectly. The protocol is robust to imperfect wording — it is not robust to not being run.
5. Observe. In two weeks: which got a yes, which got a no, which got a referral, which produced a named next step. Update the ask shape against the data.

Three asks is small enough to run, large enough to generate signal. Two hops is the maximum where the speech protocol still carries trust. Two weeks is short enough that the loop closes before motivation decays.

Frontier and caveats

The framework is not universal.

• Internal-mediator networks import rivalry. Girard's distinction and the coaches-teach-you-to-coach entry apply. Asking inside the same status market you are competing in routes you toward the people who treat you as competitor before mentor. Prefer hops outward, into adjacent domains where your gain is not their loss.
• Low-trust environments degrade the protocol. In contexts where any ask is read as encroachment — certain corporate political climates, regimes with weak property rights, communities under siege — the speech protocol still operates but the cost of every ask is multiplied. The Kresy carriers asked inside narrow trusted circles for exactly this reason. Calibrate.
• The protocol presupposes a target. Asking before you can name what you want produces noise — the recipient cannot route a request that isn't pointed somewhere. Run possibility-models work first if the named target is missing. The ask is a transmission tool; it is not a substitute for knowing what you would do with the answer.

Closing

Your environment is, by selection, the set of moves your current network has already made. The next move sits one or two hops outside it, in someone else's environment, and the bridge is a sentence said out loud to a specific person about a specific thing. The ask is not a personality trait. It is a protocol with a learnable shape, and the cost of running it badly is almost always smaller than the cost of not running it at all.

The first step out of any environment is making a request that names what you want, of someone close enough to route it, in a form they can grant in under a minute. That is the speech move. It is small, and it compounds.