there are absolutely pitches that an arbitrary piano cannot be tuned to. it is also non-trivial to tune a piano.

Since an "arbitrary piano" is tuned by turning screws, like simpler string instruments, the only pitches that it cannot be tuned to are those that fall outside the range of each string. Departing from the usual inharmonically tempered 12-EDO tuning would just sound strange, possibly outside some piano tuner's professional comfort zone.

https://www.modartt.com/pianoteq_overview :) Very high quality and affordable software modeled piano. It is rather easy to tune via Scala KBD formatted files.

Not a piano, and not relevant to the parent remark.

And that isn't a nuanced opinion; a common one among rigid materialists. Its not the best piano for a stage, but I am sure it could fool your ears in a recording. Morpheus would think it was a piano too. "What is real? ..."

Θ does not usually mean average, but simultaneously upper and lower asymptotic bounds.

you might want to read that chapter of CLRS again

You're right.

It isn't the average bound, it is the upper and lower bound stated together ( as long as thats the same function )

This _IS_ the U.S. Government.

I think that's a completely natural source of frustration for someone who is used to being able to put hands to keyboard and solve problems quickly, but IMO the higher you climb the ladder, the more you have the opportunity and responsibility to take a longer view and to delegate - which often means that other people are pushing the code.

Harnesses are hands.

Why should HN pay for a questionable service when it already gets crowdsourced results for free?

If 1Password hadn't jumped the subscription model shark, I'd be all over that. Any chance you'd entertain supporting bitwarden?

Varlock has bw plugin too - and similarly you can either wire up individual items or pull a whole env style blob from a single item if you prefer.

This is an incredible resource for someone trying to prep for an audit. Thank you!

Appreciate it! Happy to answer any questions.

I don't think that is controversial.

My system CLAUDE.MD spells out my preferences in stack and architecture, and how I expect the agent to interact with me.

I have it generate a summary doc in large repos so it doesn't need to slurp in the world to review a PR or work on a corner of it.

This is interesting. I'd already had a conversation with my harness ( pi ) about incorporating continuous improvement. This is a great deal better than what I came up with.

Glad to hear it! Good luck, have fun. The agents tend to do a pretty good job incorporating these ideas. This was an unexpected thing we learned when publishing the initial harness engineering post.