Developers are overly optimistic, undervalue their time (they don't know the loaded-cost of their time), and overvalue their solutions (they rarely account for maintenance costs or the long-tail of time to perfect the solution). So developers think they are cheap. For example, I had an engineer express disgust that we might have to pay $4/month-per-user for gitlab just to have mirroring enabled. He said "we can do our own mirroring." When asked how long it would take he said, "between 2 hours to two-weeks if we have to iron out bugs". So when accounting for his time, his price to deliver ranged between $200-ish in the best-case scenario, to $8000-ish in the worst-case scenario (which far exceeds the cost of paying for the licenses). Needless to say, we paid for the licenses. He also wasn't considering the opportunity-cost, i.e., what will he NOT do that's more valuable while he works on this proposed solution that we could just pay for?