Guanella Pass

Heading North into Georgetown. The valley north of this vantage drops significantly, and if the ground wasn’t in the way, Georgetown would be dead ahead and slightly to the right.

The Morning After

So I got the numbers.

The morning after in terms of leaving my computer up and running all night, btw.

The numbers exist. They’re even worse than I expected, and I half-heartedly expected fire. But that’s actually not so bad, because I can start poking at them now.

I’m using somewhere in the vicinity of 12,000 samples, each one 20,000 data points raw. I downsample by factors of 2, 3, 4 or whatever so they don’t throw memory errors, and up-and-downsample batch sizes for accuracy. I was really surprised that didn’t make a whole lot of difference.

My validation dataset may be the issue, so that’s the next setting to change. After that, I’ve got a dozen ways to manipulate GradientThreshold, batch size, weights, and momentums, etc. I can downsample by factors of 10, 100, or more and throw huge minibatches around. If nothing changes the results, I have bad data. The data doesn’t clearly indicate the stuff I’m looking to measure.

However, I know the signal is in there. That makes this somewhat easier, because I know the problem is solvable. It just takes longer than one might think.

Favorite speeches

Probably my favorite bit of the Return of the King movie.

Arise, arise, Riders of Théoden!
spear shall be shaken, shield shall be splintered,
a sword-day, a red day, ere the sun rises!
Ride now, ride now, ride! Ride for ruin and the world’s ending!
Death! Death! Death!
Forth Eorlingas!

It just gets you fired up.


Basically there are three steps in a loop for algorithm development.

1) Make it work at all.

2) Make it work better.

3) Make it work faster.

Repeat while exist(time/money/energy) == TRUE

Right now the landmine detector has just worked at all. At all. Numbers come out. Those numbers are terrible, and they take forever, but they EXIST!

‘I live!’ screamed the dust speck at uncaring heavens.

So I moved on to step 2, making things work better. Notice how fast isn’t even a consideration yet. To make something work better, I need to get more numbers, so I left my machine up and running overnight. I’ll hit it with 12 hours of processing and see what happens. Maybe fire. Probably data. Maybe good data, maybe bad data, but the data will probably exist in the morning. That’s a pretty big hurdle.

The thing I like about engineering that’s different from physics or bomb disposal is I can make progress just by trying. Just by working at it for a while, I can make progress. It’s not only dumb luck. It’s not only rolls of the dice. Some of it is just grinding. Entrepreneurship and seeking funding is rolling dice, but actually solving the stupid problem, making the blinky lights on the computer blink right, that’s mostly grunts, effort, and relentlessness. The feeling of being able to accomplish something by working, that my fate is in my own hands, is immensely empowering.

Fantasy Dollars

If you’re going to get some money in the future and you’re spending it in fantasy, you can buy lots of things. All of the choices exist in fantasy.


I’m going to receive a thousand bucks. Yay! I want a thousand bucks. I can get a couch and a fairly nice one. I need a couch, and it would be nice to have a place to sit and read other than my work desk. I can also save it, invest it, or pay off debts. I could buy new shoes, do the brakes on my car, or replace my ancient snowboarding boots. In fantasy, all of these things are somewhat done. In my head, now, while I don’t have the money, after I get the money, all of these things are accomplished because they could be.

I can’t do all of them. It’s only a thousand dollars. But emotionally, all of them feel like they’re done because in the fantasy, they’re all possible.

When I actually get the money, I’ll do one of them. And then the others won’t be done. They won’t be viable possibilities anymore. I won’t think of getting a couch if I pay off the credit card; I won’t imagine getting comfortable boots if I get the couch, and the car’s brakes can go a little bit longer. It just shimmies, just a little, under heavy braking.

All of those undone things go from being possible to impossible, because after that check drops and I spend it, they’re no longer possible. Getting the money means losing possibilities. Perhaps it would be more accurate to say spending the money means losing possibilities, but either way, those possibilities get lost.

This is how writing feels to a lot of people, myself somewhat included. Before the book is written, when it lives in my head, it’s a romance, an adventure, a message, and a saga. After writing, it’s…just this. Maybe it’s good, maybe it has a romance, maybe it’s many things to many people, but it isn’t all the possibilities it could have been before being written.

This is how money works, and is one reason why rich people don’t think they’re as rich as other people think they are. Because if you’re looking at a rich person or imagining a bank account, you’re in the realm of possibility where could buy a jet, a mansion, pay off the debts, or build the stock portfolio. In fantasy, you could do all of those things. But when the person with the money does one of those things, they spent the money. It’s gone. The jet OR the mansion OR the debts OR the portfolio, or, or, or exclusive, are now not possible.

What do you do?

Write more books. Earn more money. Because as long as more is coming in, the possibilities are real.

Accept humility? Enough may never be enough?

Give up fantasy?

That’s one of the great questions.

My thousand bucks exist in fantasy anyway.