In general, English speakers seem not to be used to pronounce words that end with a non-silent letter e. Every time they pronounce my name (Daniele) like Danielle, I tell them it's Daniele (dah-NYEH-leh), and they go: dan-YELL, right? I tell them to say Daniel/Dani or anything that gets my attention. They seem to care more about the correct pronunciation of my name then I do, and then they botch it badly. :D
Many other Italian words that end with the letter E are changed to I, or mispronounced.
My name is Daniele (I'm from Italy), and I've been living in NY for a while. My real estate attorney name was Daniele as well (she's a woman,) it was extremely confusing at closing time.
Also, fun fact, I got invited to a ton Women in Tech stuff. I even got recruited to a company that was proud of having > 50% women in their teams. I assume that was because I was working on an OSS project (Chromium) and I had no avatar there.
Also, in the two FAANG companies I worked at I was added to women groups a few times.
Talking to my friends working there, they said they were surprised when that happened and they didn't find Claude quality to be that much better than Gemini when using agy. So maybe after all the issue is not mainly the model, but harness and custom tooling.
As a SW person (working with silicon folks though), I was always interested in the same idea. So I just vibecoded a little compiler from LLM (config.json, weights) to RTL (happy to share if you want). Then synthesized it via yosys.
I picked TinyStories-1M, synthesized via yosys against SKY130 PDK. It's GPT-Neo with 8 layers. At int8 it's ~30mm^2 per layer (including kv caches, multipliers and all the attention stuff). So it's 250 mm^2 for the full toy model.
I'm kinda surprised, I spent no time, so if you want to etch a toy model, you can do that (you can probably reduce the area significantly). If you want a mask for that is probably going to cost 100k though.
Many other Italian words that end with the letter E are changed to I, or mispronounced.