Meta cast "a very wide net" at Connect. The question that tests whether you can narrow one.
On September 26, 2026, TechCrunch reported from Meta Connect that the company is "all-in on its burgeoning line of smart glasses," and that the glasses were "pretty much everywhere" at the event. The reporter tried a still-unreleased audio-only pair with six microphones and no camera, an integration with Meta's Muse agent that "isn't yet available to the public," and a second audio-only pair built for the hearing impaired. A member of Meta's research team said that pair had been "in development for about five years," and pointed out a price difference: "Whereas a lot of hearing aids can run as high as $1,600, the glasses will sell for $150." TechCrunch's verdict on the category was cooler. "Smart glasses remain a niche that has yet to truly find its footing," the piece said, and what was most obvious at Connect was that Meta had "cast a very wide net, trying to make its glasses stylish, functional, and, most of all, useful."
Look at which product the reporter said "serves a clear and practical purpose." It was the narrowest one: a device for a specific population, with a single job, years of development behind it, and a price that answers a specific objection. The wide net is a way of searching for a market. The hearing glasses are what a program looks like after someone has decided what the problem is. That gap, between a mandate and a plan, is what the roads question measures. The mandate is enormous, the budget is not, and the panel wants to watch a candidate decide what to do first and what to leave undone.
Why they ask it
Nobody on the panel expects you to know asphalt. The question is unfamiliar on purpose, so that domain knowledge can't substitute for judgment. What they want to see is a vague, politically loaded mandate turned into a program with a definition of done, a first move, and a way of knowing whether it worked. "Fix city roads" hides several distinct problems: potholes that generate complaints, structural failures that generate lawsuits, arterials whose closure paralyzes a district, and a backlog that grows faster than crews can cut it. A strong candidate notices that these are different programs with different measures and says so before proposing anything. Roads are a budget problem and a disruption problem before they are an engineering problem, and the panel is listening for whether you find the real constraint.
The trap
The trap is the wide net. The weak answer lists everything a road program could contain: survey every street, rank by condition, hire contractors, fix the worst first, build a dashboard, engage the community. Each item is reasonable. Together they are a table of contents, not a plan, and they reveal a candidate with no theory of what matters most. The opposite trap is leaping to one clever mechanism, a citizen pothole app, say, and treating the program as solved. The panel is waiting for the moment you decide, out loud, what "fixed" means and what you are willing not to do. That decision is the answer. Everything after it is execution detail, and the panel will grant you competence there if the scoping is sharp.
Applying STAR-T
This is a hypothetical, so the frame works a little differently. You are not recounting a story; you are building one in real time, and the panel is watching whether you build it in the right order.
Situation. Establish the facts you'd need before committing, and be open that you're assuming them. What data exists on road condition, how the budget cycle works, who owns roads versus utilities versus transit, and what residents actually complain about. "I'd assume a mid-size city with a condition survey that's a few years stale and a budget that covers a fraction of the identified backlog" is a Situation: specific enough to plan against, honest about being invented.
Task. Define done. This is the step most candidates skip. "Fix roads" becomes something like: raise the share of arterial lane-miles in acceptable condition to a stated threshold, bring pothole response time under a stated number of days, and stop the backlog growing. A small number of measures, each tied to a distinct failure mode, each something a resident could feel. Naming what's out of scope belongs here too. Cosmetic resurfacing on quiet residential streets might wait, and you should say you'd defend that choice in public.
Action. Now the sequence, with a reason attached to each step. A rapid condition assessment of the arterial network comes first, because that's where failure costs the most. A pothole crew funded separately from resurfacing comes next, because the two compete for the same money and the small, visible one always loses without a fence around it. Then a resurfacing schedule that follows the assessment rather than the loudest council district, with the ranking published so the trade-offs are legible. Finally, coordination with utilities, so a freshly paved street isn't trenched the following spring. "Because that is where failure is most expensive" is the sentence that separates a plan from a list.
Result. Say how you'd know. Regular reporting on the measures defined in the Task, a stated point at which you'd revisit the plan, and a leading indicator, such as the trend in new complaints per mile, that would warn you early. Strong candidates also name the result they would not claim: this program doesn't make roads perfect. It makes the backlog stop growing and the worst failures rare.
Trade-off. Say what the plan costs. Prioritizing arterials by condition means neighborhoods with bad but low-traffic streets wait longer, and those residents will notice. Fencing the pothole budget means less resurfacing early on. A candidate who names those costs unprompted, and says how they'd explain them to the people paying, has shown the judgment the question exists to find.
The follow-up that breaks weak answers
"The council just cut your budget. What goes?" A candidate who gave a list has to start over, because a list has no ordering. A candidate who defined done and sequenced by cost of failure answers in a sentence: the assessment and the pothole crew stay, resurfacing slows, and the published ranking shows everyone exactly which streets moved. The follow-up is not about money. It checks whether the priorities you stated were real, or whether you had only decorated a table of contents with confident verbs.
Score your answer against the director’s bar
Q: How would you prepare a program to fix city roads?
Bank this as your scoping story, with the definition of done written out, and rehearse it until the trade-off comes unprompted. Try it free →