Extremely Lazy and Immensely Curious

When I explain that Claude Code has changed my relationship with developing software completely, I’m under-exaggerating… if that’s even a thing.

This off-the-cuff piece started as a unposted social update that read:

Watching Claude Code adeptly use every type of Unix command shows me that a) you can do anything in Unix, b) my higher-level operating system mostly hides this functionality from me, c) I am extremely lazy, and d) I am immensely curious.

An Introduction to Pansy Rain

This morning it’s going to start to rain — a lot. As previously described, I deeply enjoy tromping around my forest in the rain, fixing drainage, and pulling sticks from stuff. More importantly, I am deeply curious about how the weather works. This is curiosity motivated by an immense fear of climate change and a desire to understand clearly what will occur this week. And why.

Years ago, I started posting a short paragraph to a group chat filled with fellow mountain Dads. It was a three or four-paragraph affair about what was going to happen with the weather in the coming week. No one requested this lightweight and trying-to-be-approachable blurb. It started and stopped during the rainy season in Northern California, but over time, I continued through the summer, starting to track heat events, air quality (see fires), and fog.

Here’s the one from last week:

Welcome back, Winter. We missed you. Rain, lots of it this week. Chance of pansy rain tonight, but I wouldn’t count on it. Looks like we’re in for a solid inch-plus on Tuesday morning through Wednesday evening, but the show starts on Saturday night. Solid chonky rain through the end of the week and into next. Weirdly, the first storm is coming from Hawaii (wet), but the big weekend fronts are coming from Alaska. This latter rain is going to hang.

See? Approachable? My initial goal was that anyone could quickly read this blurb and plan for their week. Ah, yes. Rain on Tuesday. I will plan accordingly.

The friendly tone and brief amount of content might lead you to believe the construction was equally brief. It’s not. The initial flow was:

• Fire up Wunderground’s 10-day forecast. This deliciously informationally dense chart is my go-to initial stop for all things local weather. It includes rain, cloud cover, chance of precipitation, accumulated rain, and wind speed — and a lot more.
• Compare against Windy.com. Windy has similar information to Wunderground, but includes stunning maps that include a mind-boggling amount of different layer visualizations. The goal was to compare data from Wunderground to what Windy reported, which tended to provide better rain forecasts from my local microclimate.

It started as 30 minutes of work, and it grew to over an hour of research. As I received light positive feedback for the posts, I started to research data on local streams and reservoirs, along with other fun facts I discovered during my research.

Then the robots showed up.

An Introduction to Chonky Rain

This morning. Just this morning. Here’s what I’ve done with Claude Code:

• Pulled snapshots from my local weather station to backfill my storm tracker, which is a feature I built last night to track before, during, and after snapshots of storms. I wanted to see all storms (8) from this season, so I had the robots look at historical data, find the storms, and then backfill using my local weather station and open data sources (Weatherlink, Open-Meteo, Valley Water, and others)
• Found an issue with the data where I was fixing a sensor issue, which caused a data spike. Smoothed the spike.
• Comparing my local readings with other data sources; refined a multiplier I use to account for my microclimate.
• Updated creek data to include week-over-week change, so I get a better sense of how water is flowing down those mountains, into the reservoir, and then into town.
• Moved all this data into the Sunday report mail that I use to write my four-sentence approachable blurb.
• Oh yeah, as I write this piece, the robots and I are now planning what we’re going to track during the Summer. Fire alerts, air quality, and fog monitoring.

This morning’s work builds on an existing set of scripts that run early every Sunday morning to track:

• Weather observations and color for the coming week.
• Current conditions from my local weather station.
• Last 7 days of rainfall.
• 10-day forecast.
• Local reservoirs week-over-week change.
• The first version of my storm tracker functionality.

I am perfectly capable of building any part of the above system. I am capable of finding the services that provide the information, signing up for an API, understanding the API, building code to call that API, generating a report, and sending that report as an email. Done it a lot. Made a successful career of it, too.

I’ve written none of the code for this weather project. I’ve signed up for some API keys, but mostly what I’ve done is tell Claude Code what I’m looking to do and let the robot figure out what API we need, how to use it, and then suggest different ways we can report this information.

Extremely Lazy and Immensely Curious

When I go through these types of robot-related productivity rants, someone invariably unhelpfully volunteers, “Well, this is a good way to get really dumb.” What this someone is suggesting is that because I am doing none of the work involved in building the thing, I will not intellectually profit from the exercise.

In a world where I hadn’t spent several decades building software, I would partially agree. The lack of domain experience could mean I trust whatever the robot’s built at face value, but — just like humans — the robots make mistakes and often straight-up lie. Been dealing with those situations for a bunch of decades, too. It’s not a problem, it’s how humans (and robots) work.

Yes, I am extremely lazy. I’ve developed a set of habits that support this laziness. Where’s the best affordable Italian restaurant in Greenwich Village? Ask Noah. He knows, and he responds quickly. How should I think about this emotionally charged and complicated people situation? Ask Julia. She knows, and she can view complete emotional chaos dispassionately. Which API should I use to monitor fog? Ask the robots. They don’t know, but they know how to know.

Am I sad as I watch the robot expertly craft semi-familiar Unix commands to perform all the weather-related wizardry? Nope. I learned to delegate to humans a long time ago. It was hard to give up the Legos to someone else, but this forced me to learn other lessons. How to ask others to help. How to clearly make requests. How to listen to their responses to see if they heard. Hearing them. Watching how they work and when they ask for help. Learning how to taste the soup. Discovering lies. Fixing them. Remaining immensely curious and always learning how to communicate better.

Here’s this week’s weather blurb. Written by me:

At this second, it’s partly sunny at the house, but just wait. It’s coming. Around noon today, chonk-i-ish rain begins and continues through Monday. Let’s say two inches today and three inches tomorrow. Sweet. Short break on Monday night, but another inch on Tuesday and a drizzle on Wednesday before we dry out a bit, but more next week. Might be chonky.

(Two inches last week at Chez Rands, Lexington Reservoir up 2.5% on the week, twenty-nine inches for this rain season so far.)