Meta called Muse the "centerpiece" of its vision, then spent a keynote listing options. The question that tests how you pick one.
On September 24, 2026, TechCrunch reported on the keynote that opened Meta's annual Connect event in Menlo Park, where CEO Mark Zuckerberg gave the stage to Muse, the personal AI agent the company launched in early September. "The centerpiece of our vision for what we’re building is Muse," Zuckerberg said, and the rest of the keynote read like a catalogue of places it might live: a real-time video avatar, Meta's smart glasses, and a Mac desktop where, in the words of chief AI officer Alexandr Wang, "You can walk away from your computer and it keeps working for you on all the jobs you lined up." Add an email address of the agent's own, which Wang said would arrive "soon" without offering a more precise timeline, and retail connectors to Best Buy, Gap, Sephora, Walmart, and Wayfair. The business model was stated in a single sentence: "We’re standing behind this by making Muse free for a huge number of tokens, with the expectation that over time we will profit by taking a small fee from transactions." Wang also said the company had received more than 1,500 applications from developers wanting to build connectors, inside a week. When the glasses integration ships is not clear; TechCrunch noted that Meta would only say "the coming months."
Read the keynote as a decision rather than a demo. A company with a new agent and every surface it owns is facing a problem with a number of possible solutions, and a keynote is the one venue where the permitted answer is all of them. Zuckerberg still ranked. He named a centerpiece, named a bet on where the money comes from, and let the rest sit as options with soft dates. That ordering, more than any single feature, is what a listener carries out of the room. It is also what an interviewer is listening for when they ask a question that reads, on the page, like an invitation to list.
Why they ask it
The question, as reported to L8 Loop, is: Tell me about a time when you were faced with a problem that had a number of possible solutions. It is asked at manager and senior levels because most real problems at that level have several defensible answers, and the job is not finding an answer but choosing one on purpose. The interviewer is trying to see three things: whether you generated real alternatives or one option plus straw men, what criterion you used to choose between them, and whether you knew what you were giving up. A candidate who can show all three has shown judgment. A candidate who can show only the outcome has shown luck, or at best diligence.
The trap
The trap is that the question rewards breadth in its wording and punishes breadth in its answer. Weak answers narrate every option at equal length and then land on the chosen one with a sentence that amounts to: so we went with the second. The listener has heard a catalogue and no decision. The second failure is the reverse. The candidate names the chosen solution first and mentions the alternatives as things that were obviously worse, which tells the interviewer the alternatives were never seriously weighed. Both fail for the same reason. Neither exposes the criterion. Without a criterion, the interviewer cannot tell whether the choice would have been different under different constraints, and that is precisely what they want to know.
Applying STAR-T
Situation. Two sentences of context, with the constraint that made the choice hard stated outright. Our checkout service was timing out under load during the fall promotion, and we had two weeks before the next one. Three teams each had a proposed fix, and each fix had a sponsor. The constraint is time, plus the political fact that every option had an owner.
Task. State what you owned, which is the decision, not the fix. I was the engineering manager on the platform team, and it was my call which fix shipped and which two were told to wait.
Action. This is where the answer is won. Name the options briefly, then spend most of the time on the criterion and how you applied it. The options were a cache in front of the pricing lookup, a rewrite of the query path, and a queue that shed load at the edge. I set one criterion: which fix could we prove under production traffic before the promotion, not which would be best in six months. The cache could be tested in a day with a shadow deployment. The rewrite could not be tested in time. The queue changed customer behaviour under load in ways we could not model. I chose the cache, wrote down why the other two were deferred rather than rejected, and gave the rewrite a slot the following quarter. Notice that the work is in the ranking, not the list.
Result. A measurable outcome, and a sentence on what happened to the options you did not pick. The promotion ran without a timeout, and the rewrite shipped the next quarter with the cache still in front of it. The second half matters. It shows the deferred options were real, not rhetorical.
Trade-off. This question rewards naming a cost, because every choice among several viable solutions carries one. The cache added a staleness window on prices that finance had to sign off on, and getting that sign-off took a real conversation. A trade-off stated plainly is the strongest evidence that the alternatives were weighed, because it proves you knew what the chosen option was worse at.
The follow-up that breaks weak answers
What would have changed your decision? A candidate who chose by criterion answers at once: a promotion a quarter away, and it is the rewrite; unpredictable traffic, and it is the queue. A candidate who chose by instinct and narrated the alternatives afterward has nothing to say, because the alternatives were never tied to a condition. The follow-up is not a trick. It is the interviewer asking to see the criterion a second time from the other side, and it works because a real criterion is reversible and a rationalization is not.
Score your answer against the director’s bar
Q: Tell me about a time when you were faced with a problem that had a number of possible solutions.
Bank the one decision you made with three live options, and rehearse it until the criterion comes out before the outcome. Try it free →