Ship fast

When people say you should ship fast, what they mean is that you should ship frequently. This advice is good. In fact, its the only advice you should care about. Maximize your sampling rate. It is so important that you ought to sacrifice everything else to increase it.

Every sample you take, the more information you gain. The ultimate sample is actually shipping your thing, and seeing how well it does in the real world. The ultimate discriminator is the physical environment. You need to reduce the friction for this. You need to bring it to the limit: make it instant.

And if you can't engineer around slow sampling rates, don't even bother doing it. Its important.

A higher sampling rate is why software engineers make so much more money than hardware engineers. Software is easy to produce and ship fast. So we get to the best version of our software, faster. We learn more, faster. We test, faster.

How fast could we possibly sample? At the limit?

Well, what if your development server was your production server? What if every time you saved while writing code; your actual development server would update? What if you had a direct line to your users? What if you were your user?

Talking fast, posting good, getting outsized results. These people have one thing in common: high sample rates

I had a friend - online handle "Marshmellow". He was a good project M player. He would wipe me every time. But eventually, I started beating him. More and more often, until he couldn't hold a candle. His play style was zoning. When I started playing against him - I didn't have a play style. I had to come up with one.

I chose style. I chose action. I chose energy.

I chose APM.

I didn't focus on techniques, I didn't focus on combos. I simply focused on the volume of actions I would take per unit time.

sample rate is everything. maximize your samples taken per unit time.

Engineering for sample time

The secret of sample time is an obvious fact to some engineers. Pay attention to the decisions they make. They'll force down the cycle down, taking seemingly ridiculous decisions to force it. Sacrificing "quality", changing the scope of the project entirely, actually aiming for a totally different thing than what they set out for in the first place.

Sampling time is so important that you should not even bother doing anything that has a long loop. It's simply not worth your time.

Designing for assembly. Dropping all dependencies and getting down to the bare bones. Pushing it, pushing it as far as they possibly can. Pushing it to the absolute limit.

Need to test motors out? You should spend an entire month building your own motor winding machine. Otherwise, you are going to have to wait at least a week of lead time to get a new configuration. *It is worth it*.

You have to set up the deployment, creation of your product such that you can get feedback instantly. This is my main gripe with ios applications - you have to wait for a human to review your changes. You have to wait for your users to update their app. You have to *wait* to get feedback.

I simply refuse to build ios apps (unless the ios app is a renderer that talks to some backend that I control through which I can immediately change my application).

The vast majority of people don't understand how important sample time is. So they'll call out my decisions irrational. They make decisions that they think are rational, which reduces their sample time.

Dependencies will kill your sample time

Dependencies on people, software, something where you need someone else to do something. Something to wait on another team. It's simply out of the question. The sampling time I'm talking about here is not in human weeks, human days even. It's not even in hours. I am talking about minutes. The absolute limit. It's getting rid of everything in the way to achieve the high rate. Testing? Development/production instances?

Simply just point your domain at your dev server. I am talking about pushing it to the limit.

I take my own advice

The higher your sample rate, the more you learn. It is a multiplier. It is worth it. Pay attention to the seemingly arbitrary decisions I make. They are not arbitrary!