I'm too lazy to post every sketch and painting I create. Here's what I do post.

I've done 37 posts. At the rate of 0.14 posts per day, in 19,556,078 years I'll reach a billion posts. Yay me!

A watercolor attempt of a mid-century styled room - chairs, lamp, paintings - referenced from an old American Bungalow article.

Interiors are Fine for Practice. Why Don't I go Paint a Lamp?

One reason I picked drawing and painting as a new hobby is because it doesn't involve a screen. This makes it a challenge to find things to paint, especially in the evenings. Luckily, I found an old 'American Bungalow' magazine and it has some neat photos of 20th Century homes and interiors that, up until this point, I sucked at painting. And while I'm not awesome, I'm a little better after painting this Dirk Van Erp lamp after a few attempts. Yay me.

Three female watercolor portrait attempts with limited palettes. One primarily blue and brow, the others yellow ocre, Venetian red, and cobalt blue.

More Faces. Always with the Faces. And Accidental Shootings.

I remember in gradeschool how the movie titled "Faces of Death" made its rounds throughout random conversations yet I never saw it, nor did anyone I know. The title and schoolyard rumors were enough to spark wild and terrifying thoughts in my young head. Decades later, as I wonder the isles of the outside world with my fellow peoples, sometimes a glance at a face will surface a messed up thought: "Maybe she killed her husband and is getting that slide of chocolate cheesecake to celebrate?" What's dumb, is that thought has nothing to do with the movie, per the synopsis I've read, though it makes me think of that title. The brain works in mysterious ways.

Six watercolor attempts at a tree on a hill with a limited palette of ultramarine, yellow ocre, and sepia brown.

Trees are Cool and All, Yet a Challenge to Paint. And Fire Pokemon Perform Public Benefit.

Even with watercolor reference pictures to copy, things are a challenge to paint. This time, trees. "Slow-motion explosions toward the sky" one YouTube artist I follow describes them. The trick is to not let discouragement win and to keep practicing. Mistakes happen, it's how I learn, and I need to keep pulling up my big-boy pants and not crying about ugly results. And today's public service announcement: never get in the way of Pokemon on skateboards.

Three watercolor attempts at a scene looking down the street toward the ocean with houses on both sides and at the far end.

More Boring Boxes Sans Style. And a Supersonic Trebuchet Misfire

It's neat to see bits of progress with a hobby. Consistent work pays off. What's not easy is the patience. Great, so I can paint boxes a bit better and my brush, pigment, and water control slowly improves. I still don't have much of a style, nor are these very interesting. That's okay, because a supersonic trebuchet is interesting.

Three watercolor attempts of a bird bath from my yard.

Yard Bird Bath is not a Far Travel to Paint. Which is Nice for the Lazy.

As much as I'd like to take a day and go hike somewhere beautiful to paint, it's much easier and realistic to not do that. And that's okay. While I don't have a fancy marble fountain in my yard, the bird bath provided by a previous owner is a fine subject. And so are birthday parties.

Older woman with knitted scarf by a Scottish rocky and grassy hill, pen and dark graphite watercolor wash.

Bit by Bit the Lines and Washes Improve. And Let's Make Democracy Better!

Keep practicing the things I'm not good at because, well, how else am I supposed to suck less? It's neat to see progress. And continually humbling as I gain appreciation for the skill and technique of the artists out there who make things look effortless. Talent, sure. And thousands of hours of work. Never discount the work, effort.

A collage of three watercolor attempts of a beach-front mansion.

The Difference Between Millionaire and Billionaire Beach Properties Is Not That One Is Easier to Paint

I'm a fiend for the confident loose sketches of urban artists and watercolorists who can paint buildings and make them blend in with everything else. And it's going to take hundreds, thousands of practice attempts. Maybe, one day, I'll get smart about my process and more efficiently iterate on smaller details. No. That's silly talk. And two film students make a boom boom.

One attempt at a watercolor of an old mission style building with a tower and rounded corner end, with a back drop of poor neighborhood buildings.

Buildings Are Supposed to be Easy to Paint. They're Just Boxes.

I watched a tutorial that was supposed to transform my buildings from boring to better. It took a few tries and... slightly better? They're just boxes, how hard can this really be? It's wild how something that looks so easy requires a lot of practice when the easy doesn't come natural. Also, bulldozers are cool.

Two watercolor attempts at a Shetland countryside manor on a shoreline with a servants house extended off the main building. Giving up on this round. Slightly better than the previous ones I've tried.

Five Tries at a Scottish Seaside Scene in Watercolor, and No More Golf

Discouragement. It sucks the motivation right out of me. These five attempts at an old seaside manor were tough to do because I kept doing things that didn't produce the results I wanted. That's part of learning, right? I need to try and fail a lot. Fail repeatedly and hopefully learn from the mistakes. Also, a human in charge of capital golfs no more.

A couple of attempts at watercolor faces based on a video frame from the movie 'Better Off Dead' with Badger in his room with older, sexy women.

Faces Are Still a Challenge, Even with 'Better Off Dead' References

It gets easier to give myself room to play and not expect too much from what I try. And I sometimes use movies as sources of inspiration on what to practice, especially with faces which will remain a challenge for the foreseeable future.

More Entries:

Expand for a random shower thought with AI: Guillotine usage optimization?

Thought: What's the optimum guillotine (processing unit [PU]) usage for a summer concert season?

Google says there's about 1000 billionaires in the US. I'll label them "GU" for greedy units. For extra coverage, say each GU has 10 subordinate GUs (family units, business units). That gives us 10,000 GUs to process.
Wikipedia says there's over 2000 outdoor concert venues with 1000+ capacity. I'll go with 1,000 concert venues.
And 50 days of concert shows across those venues should be okay for a time limit.

Since I have access to an LLM and I'm lazy, I'll have it figure things out and assume it's correct. Here's the chat that cost $0.04.

Me

Please help with a calculation for optimal usage. I'm managing spectacle installations during this season's summer concert tours. I'd like to figure out how many processing units is optimal for the number of days and venues I have available. Here are my variables:

10000 inputs, labeled "GU". This is fixed. These are fed into the processing units.
1000 venues, labeled "V", where the inputs will be processed. This is fixed.
50 days, labeled "D". Fixed.
How many processing units "PU" would we install at each of the 1000 venues to maximize PU usage over the 50 days?

LLM

Processing Unit Optimization

To maximize PU usage, I need to make some assumptions about how the inputs flow through the system. Let me work through this:

Setup:

*GU* = 10,000 inputs (total)
*V* = 1,000 venues
*D* = 50 days
*PU* = processing units per venue (unknown)

Distribution

Inputs per venue: frac{GU}{V} = frac{10{,}000}{1{,}000} = 10 text{ inputs per venue}
Inputs per venue per day: frac{10}{50} = 0.2 text{ inputs per venue per day}

The Optimization Question

To maximize PU usage (utilization rate), you want each PU busy as much as possible — meaning you want the fewest PUs that can still handle the workload.
If each PU can process 1 GU input per day, then: PU = lceil 0.2 rceil = 1 text{ PU per venue}
With 1 PU per venue, utilization = 20% (each PU works 10 days out of 50).

Caveat

Your question is a bit underspecified — I'd need to know:

Processing rate per PU (inputs/day capacity?)
What "maximize usage" means — utilization % (fewer PUs) or throughput (more PUs)?
Are inputs evenly distributed, or can venues vary?
Any constraints like queueing, parallelism, or deadlines?

Could you clarify the PU processing capacity and your optimization goal? That'll let me give you a precise answer.

Me

Good questions. Here's what I can speculate:

Processing rate per PU (inputs/day capacity?)
- Each PU can process one input per show. We can have multiple processing units at each venue. We should limit this to 3 - 5 per venue to maximize the spectacle.
What "maximize usage" means — utilization % (fewer PUs) or throughput (more PUs)?
- Maximum usage: with a maximum 3 - 5 PUs at each venue, one INPUT per PU per day.
Are inputs evenly distributed, or can venues vary?
- Inputs can be distributed however we want.
Any constraints like queueing, parallelism, or deadlines?
- We can run multiple PUs at the same time, only once per show per each one.

Does that help provide enough detail?

LLM

Yes, that's much clearer! Let me work through it:

Updated Constraints

Each PU processes: 1 GU input per show (per day)
3–5 PUs per venue (spectacle constraint)
GU Inputs can be distributed freely across venues
PUs run in parallel, but each PU fires once per show

Capacity Math

Max capacity per venue over the tour: PU times D = PU times 50 text{ inputs}
Total system capacity: V times PU times D = 1{,}000 times PU times 50 = 50{,}000 times PU

Utilization at Each PU Level

| PUs/venue | Total capacity | Inputs used | Utilization |
|-----------+----------------+-------------+-------------|
| 3         | 150,000        | 10,000      | 6.67%       |
| 4         | 200,000        | 10,000      | 5.00%       |
| 5         | 250,000        | 10,000      | 4.00%       |

Recommendation: 3 PUs per venue

With only 10,000 inputs across 50,000 possible "slots" (1,000 venues × 50 days), you have massive overcapacity. Fewer PUs = higher utilization %.

However, since you mentioned spectacle value matters and 3 is the minimum, this is the sweet spot:

Maximizes utilization (6.67%)
Still delivers spectacle (3 PUs firing per show)
Leaves headroom for distribution flexibility

Alternative Framing

If you instead want to maximize spectacle (throughput), go with 5 PUs/venue and concentrate inputs at fewer shows — e.g., 2,000 shows at full 5-PU capacity, leaving the rest dark or running fewer PUs.

Want me to model a hybrid distribution strategy?