mdavid626

I wouldn’t say Windows Terminal is great. Have you ever used a proper one, like Ghostty or iTerm2?

I literraly can’t name one feature since Windows 7 what was worth it, even on the contrary: every update made the system worse.

I had to restore Notepad, Calculator and Paint from Windows 7. What the hell Microsoft?

The first delete would fail: “bucket not empty”. This might make the agent question the deletion (“bucket should be empty”).

I thought the way is not feeding customer data to the LLM.

In AWS eg. bucket can be deleted only when empty. Deleting all files first is your confirmation.

I don’t see the problem here. These people will be pushed out of the industry quickly and their business taken by other people, who are using agents, but are smart enough to run them sandboxed without any permission to production or even dev data/systems.

How do you prevent your customer data being used for training?

Jesus Christ. All this is true only if you let them.

It’s good idea to be defensive, design the system in a way that it can “fix” itself.

But for love of god, don’t let an LLM do everything it wants.

I wonder were are the software engineers. No senior devs at Apple anymore?

Just interns pushing to prod without any review? What the hell is going on in the software industry?

Such mistakes a trillion dollar company can not allow to happen.

It's not about Bun, but more about sqlite and zod replacements. Why interpret this as "Bun is faster"?

AV1 lacks hw support…

Yeah, but how does the sandboxed process “know” that it has to go through the proxy? How does it trust your certificate? Is the proxy fully transparent?

You know, the thing is, that it is super easy to create such tools with AI nowadays. …and if you create your own, you can avoid these unnecessary abstractions. You get exactly what you want.

How do you intercept network traffic on mac os? How do you fake certificates?

I trust sandbox-exec more, or Docker on Linux. Those come from the OS, well tested and known.

MITM proxy is nice idea to avoid leaking secrets. Isn’t it very brittle though? Anthropic changes some URL-s and it’ll break.

How the hell is it 500k lines?

It’s time to run development in sandboxes. Docker or sandbox-exec (for Mac).

Oh boy, you couldn't be more wrong. If something, LLM-s need MORE readable code, not less. Do you want to burn all your money in tokens?

There is no such thing as agentic codebase. If humans don’t understand it, nothing really does. Agents give zero fuck about anything. If they burn 100 or million tokens to add a feature, they don’t care. It’s the developers responsibility to keep it under control.

Isn’t that good advice? Rewriting it in C can prove that the problem is not in your programmin language.

Outdated OS can be the problem as well.

What kind of advice did you expect?

Is it though?

On my Mac Docker runs Linux virtualized. It’s a resource hog.

Compare that with simple native binary.

Yeah, compare this with downloading single binary approach.

Downloading JDK, setting up the correct env variables, or running Docker, all this is just pain, compared to single binary approach.

I’d be happier if this could run as one binary without Docker. Java is so much harder to setup.

Firefox doing everything, but being a good browser.

Sigh… that’s not how it works.

JPEG still rules the world. Many new alternatives were developed, none of them really that much better than JPEG.

Mixed feelings.

I use Opus 4.6 with pi.dev (one agent). I give detailed instructions what to do. I essentially use it to implement as I do it manually. Small commits, every commit tested and verified. I don’t use plan mode, just let the agent code - code review is faster than reading plan. This approach works only if you make small changes. I create mental model of the code same way, as when writing it manually.

Some people on my team codes with AI without reading code. That’s mostly a mess. No mental model, lower quality. They are really proud of it though and think they are really smart. Not sure how this will turn out.

No, they can’t.

It has nothing to do with inconvenience.

I really like that layman now make these statements - they know better than people working in the industry for decades.

It feels like to me that junior devs don’t understand what they even need to learn. They just use agentic coding to get things done, without any deeper knowledge.

The worst is, they think they know exactly what they need to learn, and also think they can make good decisions.

Involves and made only by, are 2 different things.

I use agentic coding in my daily work. I do make mental model of the code I write and I also test the code, exactly the same way, as when written completely manually.