In a lot of work contexts, forming a strong operational working relationship with people is sufficient to get on with things. You know what they're up to, they know what you're up to, but you're mostly individuals working on separate parts of the puzzle, solving assigned problems individually.
But for solving the big, existential difficult problems, like what product should we make, it can help to bind creative intelligence more tightly. Multiple minds with different perspectives have a greater depth of insight, and exchanging ideas more synchronously reduces latency - latency of feedback and validation, latency of spotting blocking issues, latency of reaching total clarity.
For this reason, it's not unusual for startup founders to go all in and live together, in the style of an artistic commune, or, say, a band.
Bands are an interesting comparison to working for a company because they bring a group of potentially highly incompatible creative people together as a result of necessity. Most song-writing musicians can only write well for one or two instruments, so they're forced to work tightly with other creatives to get a result, otherwise they won't succeed in writing hit songs. The early Rolling Stones, for example, lived together in a series of shifting arrangements in various locations in London (e.g. here), even after becoming successful enough to afford not to do so, which lead to lots of organic jam sessions, which lead to successful songs being written.
You don't necessarily need to live in a communal arrangement with your work colleagues, but in situations where you need to really nail an idea, you should work closely together just as a band does. Touching base asynchronously via emails, or having a Zoom call once a week, won't get you there - the latency between your minds is just too high and the bandwidth is too low. You need to jam, and you need to progress the jamming to a recording, and the recording needs to be something that people want to listen to.
With all this in mind, here's some notes based on what's worked for me in the past.
2 people is good. 3 is fine and can be better than 2 for breaking deadlocks. 4 people with equal creative stake is too many people for a strong creative coupling. It's possible this can be made to work; I've not seen it happen. If the fourth person is a drummer, and cool with that ("hey, can you come up with something rhythmic for this song we've just come up with?"), things might be okay.
Overt power imbalances in the mix can ruin things right from the start. If one person sets another person's salary, or one person is a founder with a massive upside and the other isn't, the salary recipient will inevitably feel like a junior partner and most likely yield too easily to the ideas of the other. Ideally, creative partners are organizational siblings or cousins. Don't forget this when you reach the point in your career when you're the person setting the salaries - your reports might not be enjoying those ideation sessions quite as much as you are.
Meet as frequently as you feel comfortable. If you are truly focused on sorting out this one idea, you're probably thinking about it all the time, especially in down time - when you're going to sleep, in the shower, while getting on the bus. You should have something new to discuss once every few days, rising to every day when things are actually in progress and you're figuratively in the studio making the recording. Informal settings are best - coffees, lunches, walks - and in as many different locations and contexts as possible, which will help memory and creativity. (see Ed Cooke's suggestions for how to pull this off during a pandemic).
Understand at the outset whose skin is in the game when it comes to the ideas and the success of the project. Who is going to get fired if things go bad? Who is going to make the most money if you succeed? Whose career prospects might be dented if you don't? If you are an engineer, even at senior management levels your ultimate ups and down sides here are often pretty light - your performance is usually judged on how well you and your team builds the thing, your benefits are mostly driven by salary, and you're usually not on the hook when it comes to ideation, even though you can have potentially a huge influence over whether things go well or not. But your counterparts might be in a position to have quite good or quite bad things happen to them depending on the outcomes. Respect this asymmetry and bear it in mind when pressure arrives.
Hidden or not-so-hidden stakeholders with creative power outside the group can unsettle things. Yoko Ono was only one factor in the break up of the Beatles, but her presence exactly didn't help matters. It can help to be explicit with outside people about your creative model, and ask if you can draw some boundaries. Supportive, powerful, reliable managers, who are happy to let you pursue your own sound, are best.
You need a lot of trust for creative relationships to succeed. Have each other's backs. Don't bitch or gossip about each other. Don't go off starting solo projects. Actively listen to each other - seek first to understand, then to be understood. Be vocally self critical of your own ideas. Seek to actively disprove your own assumptions.
Try not compete to with each other in each others domains, and try to have clarity on what each person's bringing to the table (and implicitly, what they're not). Each person should know their role - if you're a singer who doesn't play an instrument, nobody wants you to strap on a guitar!
Do your best to understand other people's craft well before weighing in on their decisions. If you're not an engineer, don't second guess which stack the team is using. If you're not a product manager, before questioning the direction, read some product management books and understand how what you're doing ties in to the core principles. If you're not a designer.. you should probably avoid aesthetic feedback altogether.
When you're succeeding, it can feel at times like you are co-conspirators. You're in on a glorious plan to take over the world together, with a great depth of understanding and a shared vision.
This is good - but you may find that while you and your creative partner(s) understand each other perfectly, but there's a gulf between that understanding and the rest of the world. Being misunderstood is the worst thing that can happen to a creative group. Dedicate time and effort into communicating what the plan is. Spending time releasing an unpolished mix tape to test the waters might be a good idea.
Don't forget that the point is to write some hit songs - things that move numbers, that your users truly adore - not to have a jam session.
Creative partnerships can be unstable and sometimes have a shelf life. A certain amount of creative tension can be highly productive: Fleetwood Mac produced arguably their greatest album, Rumors, in the midst of not one, but two intra-band relationship breakdowns. But assuming you don't want to flame out, there are some basic things you can do to promote healthy collaboration.
Work out what you don't agree on, and talk about it. Be methodical about doing this - if the relationship is working well, there will probably be a queue of things you want to argue about; in some ways this is kind of the point. Take notes and work through disagreements methodically with your partner(s), closing them out.
Sometimes there will be an asymmetry in clarity; one person will know more than the others about something, and not always obviously so. It could be to do with their prior experience or their training, or it might be something a bit edgy like their understanding of the company's politics. In some ways this is inevitable because the partnership is supposed to contain people with different perspectives; that's its strength. The key to dealing with this is active transparency: each partner must work to explain their thinking to others, not just present the final result of their thoughts.
Sometimes the constraints aren't clear, or you might be ignoring them because they don't suit you. Sometimes this creeps in unconsciously, or it might be because you're deliberately challenging the constraints. It is always a mistake to go against the constraints without engaging with them directly; don't drift into writing songs based on eastern music when you're in a grime outfit unless you all agreed to do that and it's a conscious decision to push boundaries and expand people's minded. It can help to document the constraints somewhere, especially if they're complex and you're worried about unspoken things like company politics etc.
Lastly, above all else: assume positive intent. Everyone involved is almost certainly doing their very best to generate awesome creative ideas and wants to build something truly great.