Many of them reference or reproduce a Twitter thread at https://twitter.com/jugurthahadjar/status/131066829330549965... and add relevant input for the question.
Your experience might be different. You can read the threads where these replies were posted and get more insights.
One way to speed up finding clients is to go meta: do consulting for a consulting company that does something different but complementary/related. This company has clients, and you'll partner to solve problems for their clients. Instead of going after 40 clients, you go after one that has 40 clients.
The link https://news.ycombinator.com/item?id=25760622 addresses the case of NDAs between you (A) and your client (B), and between your client (A) and their client (C), so that you could work on C's problem.
Doing that many projects has lead us to figure out patterns and ramp up on the experience curve, and we then have abstracted several aspects into our "MLOps" platform, https://iko.ai, to be able to deliver faster. This corresponds to the "Abstract" phase in the Twitter thread. You can really do that beforehand (you can see it especially in the machine learning field where some startups with no experience and no prior projects want to improve the tooling and come up with pipelines and what they consider a good way to do things, working on the wrong problems or problems that don't exist. This is dangerous).
- https://news.ycombinator.com/item?id=25402572
- https://news.ycombinator.com/item?id=25583145
- https://news.ycombinator.com/item?id=25587058
- https://news.ycombinator.com/item?id=25614824
- https://news.ycombinator.com/item?id=25621850
- https://news.ycombinator.com/item?id=25719052
- https://news.ycombinator.com/item?id=25760622
- https://news.ycombinator.com/item?id=25881430