wbradley

I’m guessing they adjusted the activations of certain edges within the hidden layers during forward propagation in a manner that resembles the difference in activation between two concepts, in order to make the “diff” seem to show up magically within the forward prop pass. Then the test is to see how the output responds to this forced “injected thought.”

I think the point is that ~/.claude should be dispersed among ~/.config/claude, ~/.local/state/claude, etc

I agree with this, it’s frustrating that in 2025 apps are still polluting my home dir.

> error: PyObjC requires macOS to build

Looks like this depends on macOS.

I ended up going the opposite direction. My 11th grade trigonometry teacher forced me to repeat trig despite my having gotten a B. I was angry at the time! She did me a favor because by the time I got to Calculus in University it was very intuitive and I didn’t have much of a problem with it.

So, uh… what happened?

I went looking for mentions of Semantic Networks, RDF, triplets, taxonomy, or other common knowledge-related terms but I found nothing. Thoughts about pre-existing knowledge graph systems and how they may or may not apply to System? Thanks!

The idea of static meaning “fully evaluate this at compile time” in D is a good one.

Please do not promote the use of wildcard imports.

Yeah, that's true. I could have chosen a better example there. I was trying to pick a tech that Apple tended to prevent competition in back then. Perhaps Safari/Webkit would have been a better example.

My brain kept leaping to this analogy...

Chef : Software Engineer

2017 : 2010

Donaldson : Apple

Alcohol : iMessage

Food Hall : App Store

30% : 30%

Looks interesting! However, there is some confusion over whether it's truly REMOTE or not. The job description page says

> Live or be willing to move to Mexico City - this is not a remote position

For about three years I maintained the official internal coding conventions documentation for the Microsoft Word development team. This was when we were working on Office 2007. I remember loving the spirit of Hungarian notation. Using it in practice was also a major lesson and shaped my valuation of naming variables and types.

That same timeframe we ported Word's codebase from C to C++. That meant rewriting a bunch of manually ordered vtables for COM as regular virtual member functions on classes, and sadly, the elimination of anonymous unions from the codebase. But, type systems have really taken over and I think that's a good thing.

Because Word was written in C it followed a beautiful pattern of functions calling other functions all the way down. It treated data as data, and did not utilize classes or inheritance. I later came to understand the value of that way of thinking but at the time I thought it was so old fashioned. Of course, lacking the OO mindset meant that Word also lacked a great deal of encapsulation, and had horrible coupling across certain modules. Most notably, Undo, and Display.

I suppose that's enough reminiscing for now.