Consider a case study of a small group building quantum software. The people are invented; what happens to them is a pattern that anyone in fast-moving technical research will recognise: a handful of sensible goals, written in good faith, slowly turning into the wrong ones. The team in our story is four people building tools that let researchers test quantum algorithms on simulators before touching real hardware. Their habit of rereading their goals every month, and what it caught, makes a useful lesson for any group working in quantum computing or another quickly changing field.
The team and its original goals
At the start of the year the group wrote down four goals on a shared page.
- Build a benchmark suite that compares how several simulators handle circuits of growing size.
- Port one well-known algorithm for factoring into a clean, documented example.
- Write a short tutorial series for newcomers who know classical programming but not quantum.
- Investigate whether a hybrid approach, where a classical optimiser steers a quantum routine, could suit a scheduling problem a partner had mentioned.
Nothing about them was foolish. Each had a reason, and everyone agreed. They agreed, too, on a small ritual: the first Thursday of every month, thirty minutes, goals only. No status updates, no demos. Just read the list and ask whether each goal still fits.
Month two: the first small catch
At the second reread, someone pointed out that the benchmark suite goal said “several simulators,” but one of the simulators the team planned to include had stopped receiving updates. The goal’s wording quietly assumed something that was no longer true. The team reworded it to “two actively maintained simulators,” which was a smaller promise and a more honest one. It took ten minutes.
That was an unremarkable adjustment, but it set the tone. Reading the goal aloud had made the assumption visible. Nobody felt blamed for the mistake, because the whole point of the meeting was to find such things.
Month four: a goal that outlived its reason
By the fourth reread, the tutorial series had become an awkward item. The team had written two instalments, but an external project had just published a very thorough beginner course. The lead asked a plain question: if we were starting today, would we write this? After a pause, the group admitted they would not. They would write something narrower, aimed at the particular simulator interface they had built.
They did not scrap the goal. They changed it to a short guide covering their own tool. The two existing instalments were folded into an appendix. The shift saved weeks of effort that would have gone into duplicating work already available elsewhere.
Month six: the goal nobody wanted to say was dead
The hybrid optimisation investigation looked the most exciting of the four yet had the least to show for it. Each month someone said, “We should get to that.” At the sixth reread, the lead asked for evidence: what had the team actually done toward it? The answer was a single afternoon of reading. The partner’s scheduling problem had also changed shape since the first conversation, which nobody had mentioned until that moment.
The group decided to retire the goal and write a note: “Revisit if the partner returns with a firmer problem.” It was not a defeat. It was a recognition that the team’s energy was finite and that the goal no longer matched the problem. For anyone curious about the kind of problem such a hybrid approach targets, a primer on quantum algorithms for hard optimisation problems covers the basics.
What the page looked like by then
It helps to picture the shared page at month six. Each goal had a short line of text, a date of the last decision and, in a narrow column, a single next action. The retired goal sat at the bottom under a grey heading. Nothing about it was fancy: a plain document that anyone could open in a few seconds. The simplicity mattered. Because the page was easy to read, the team did not need a special meeting room or a long agenda to use it, and a visiting colleague could take in the whole plan in under a minute.
One of the group’s habits deserves a mention. Whenever a goal changed, the person who proposed the change wrote the old wording in square brackets next to the fresh version, and left it there until the following session. When the next reread came, the team could see at a glance whether the change had helped or whether they had simply reworded a problem without touching it.
What the team noticed about the process
Looking back at six months of notes, the team saw several things.
- The most useful questions were short: “Is this still true?” “Would we choose this today?” “What did we do toward it?”
- Goals about external facts, such as which tools exist, drifted fastest. Goals about internal skills drifted slowest.
- Retiring a goal felt much easier when the decision was recorded with a reason and a revisit condition.
- The meeting never ran over thirty minutes, because it never became a status review.
They also noticed that the monthly cadence matched their release rhythm. Every few weeks something changed in the simulator landscape, and the reread was the one place where those changes were consistently checked against the goals.
A side conversation about learning
Along the way, one team member raised the question of personal learning goals: which new topic should each person spend evenings on? The lead suggested adding them to the same page, with the same monthly check. As a small aside, someone mentioned having browsed the AirCrewU site, where aviation professionals take courses online and which calls itself independent, just out of curiosity about how another trade organises its study. The relevance to quantum work was slight, but it prompted a useful conversation about how to describe a learning aim so it could be checked later.
One person’s goal had been “understand quantum machine learning.” At the reread, the team agreed it was too broad to check. It became “explain, in a short note, how a quantum circuit can serve as a feature map.” The new version was concrete enough that, a month later, the person could say plainly whether it had been done. For background on that general area, a general read on how quantum methods may change machine learning work is a reasonable starting point.
Month nine: the quiet payoff
By the ninth month, the list looked different from the original. The benchmark suite was complete and in use. The tutorial had become a compact guide. The factoring example was nearly done. The hybrid investigation sat in a “revisit if” list. Most importantly, no goal was a surprise. Everyone could say what the team was working on and why, and a new colleague who joined that month understood the plan within a morning.
When an outside collaborator asked what the team had learned, the lead did not need to prepare slides. They opened the page with its dated decisions and walked through it. The record spoke for itself.
Lessons you can borrow
- Schedule it and protect it. A fixed monthly slot makes the reread routine rather than optional.
- Keep it about goals only. Status talk will swallow the time.
- Ask for evidence. “What did we do toward it?” cuts through good intentions.
- Record revisit conditions. A retired goal with a clear trigger is easier to let go of.
- Reword rather than rescue. Many goals need a smaller, more accurate sentence, not more effort.
The story is invented, but the mechanism is real. In a field where tools, libraries and expectations shift every few weeks, a goal written in January is a guess about a future that has since changed. Reading it again, honestly and regularly, is the cheapest way to stay in step with what is actually happening.