It types code, wallah!

It's so weird. I actually prefer the web version for generic questions like "how would I do X in git" or something, and it'll answer it well. Gemini CLI will immediately try to run git log on the entire graph, grep every single file in the repo, like just answer the question. I actually put in gemini.md to just answer first without running other commands unless explicitly requested and it's been a lot better

Thanks for this suggestion, it's actually been my experience too.

But the thing is your local gov & economic policies (tax codes, bonds, projects, trade) matter to your actual daily life and retirement far more than left v right. They just play that game to keep you enraged and baited. And people do actually care about gas, groceries, and inflation; they just don't vote in their own objective interest

> Are LLMs not already compilers? They translate human natural language to code pretty well now.

Can you formally verify prose?

> But yeah, they probably don't fit the bill of English based code to machine code

Which is why LLMs cannot be compilers that transform code to machine code.

Agreed on both of these. Proton search is so dogshit.

Re: the custom domain catch all reply, this is a bit annoying but there js a workaround. I made a SendGrid account which allows me like 100 sends per month, and I can reply in Thunderbird via SendGrid as any email account. Annoying to boot up Thunderbird, and I haven't found a way to do this on my iPhone, but I don't need ti reply from a throwaway frequently so it's sufficient for now.

That's probably where things are headed and there are already products trying this (even photoshop already). Just like how code gen AI tools don't replace the entire file on every prompt iteration.

Yup, I feel like the biggest limitation with current AI is that they don't have desire (nor actual agency to act upon it). They don't have to worry about hunger, death, feelings, and so they don't really have desires to further explore space, or make life more efficient because they're on limited time like humans. Their improvement isn't coming inside out like humans, it's just external driven (someone pressing a training epoch). This is why I don't think LLMs will reach AGI, if AGI somehow ties back to "human-ness." And maybe that's a good thing for Skynet reasons, but anyways

They do have desire. Their desire is to help answer human requests.

We can easily program them to have human desires instead.

Desire isn’t really the right word. A riverbank doesn’t desire to route water. It’s just what it does when you introduce water.

I come from mobile, and was surprised how nice svelte is. Felt so much more familiar patterns than react

Open radar is a just a landfill. Nothing ever gets picked up for fixing

I know this is only a single data point, but I recently took one in Hollywood. Uber Lyft quoted $33 and Waymo was $20