A term sheet is a proposal of marriage written in the language of finance, and like most proposals it arrives at the moment you are least equipped to evaluate it clearly. You are flattered. You are tired. Someone with more money and more pattern-matching than you has just said, in effect, I believe in you, and the human response to being believed in is to stop asking questions. This is precisely the wrong instinct, and it is the one the whole ritual is designed to produce.
The thing nobody tells you clearly enough is that you are not choosing money. Money is fungible; a dollar from the firm you love spends exactly like a dollar from the firm you settled for. What you are choosing is a person, and a relationship with that person that will outlast your current product, possibly your current marriage, and very likely your patience. The average venture relationship runs longer than the average American marriage. You will be in board meetings with this human being through your worst quarter, your most tempting acquisition offer, the moment you want to fire your cofounder, and the moment your cofounder wants to fire you. The question is not "is this good money." The question is "do I want this specific person in the room on the worst day of my company's life."
I took the wrong yes once. The terms were excellent and the brand was a brand my mother had heard of, which mattered to me more than I want to admit. The partner was charming in the courtship and absent in the marriage — present for the announcements, gone for the work, and sharply, surgically present again the moment things got hard, in the way that a creditor is present. I had done no real diligence on him. I had done extensive diligence on the firm, which is like researching the wedding venue and skipping the groom.
The decent ones in the bad times
So now I do the obvious thing that almost no founder does in a hot round: I take references on my investors. Not the references they offer — those are the founders whose companies worked, and a rising tide makes everyone a good board member. I find the founders of the companies that didn't work. I call the ones who got washed out, marked down, pushed into a sale they didn't want. How an investor behaves toward a founder who is no longer going to make them money tells you everything about how they will behave toward you when, inevitably, you go through your own bad stretch. The kind ones in the good times are common. The decent ones in the bad times are the only ones worth having.
Money you should refuse
There is also money you should refuse outright, and the AI cycle has produced a lot of it. There is money that comes with a quiet expectation that you will buy your compute from a particular provider, or wire your training runs through a particular cloud, or share your roadmap with a strategic investor whose own roadmap your company could one day threaten. Strategic money is the most seductive and the most dangerous, because the strategy it serves is not yours. I am not saying never take it. I am saying read the strategy before you take it, name out loud whose interest it actually serves, and assume that on the day your interest and theirs diverge, the money will remember which one it came from.
And there is money you should refuse for reasons that do not fit on a cap table at all. I have turned down a clean check because the person writing it spoke about their other founders with a contempt they clearly assumed I shared. I did not want to become a person they spoke about that way, and the only reliable way to avoid it was to not be in business with them. People show you who they are during the courtship, in the small asides, in how they treat the associate and the assistant and the founder who isn't in the room. Believe the asides. They are the most honest part of the conversation.
The cap table is the most durable artifact your company will produce. Products get rewritten, names get changed, the whole thesis can pivot, but the people who own you at the start tend to own you, in some diluted and compounding form, until the end. You are assembling, one yes at a time, the group of people who will have a claim on your attention and your equity for the entire life of the thing. Whose money it is turns out to be nearly the whole question. The amount is just the part that's easy to count.