sampullman

That's exactly my point, it's a nice tool in the toolbox, but for most tasks it's not fire-and-forget. You still have to do all the same verification you'd need to do with human written code.

If natural language is used to specify work to the LLM, how can the output ever be trusted? You'll always need to make sure the program does what you want, rather than what you said.

I did not find this to be the case, except with a few low quality vendors we ended up dropping.

It was mostly the same as anywhere else, you go talk to them in person, tour their facilities/processes, and see what else they've built.

I was warned strongly about IP theft and cost cutting, but didn't find that expectation quite met reality. It may have been that our products were mostly un-copyable, and we specified everything precisely, or were just lucky.

I use a new Ryzen based mini PC instead of Mac mini, but the reasoning is the same. For the amount of compute/memory it pays for itself in less than a year, and the lower latency for ssh/dev servers is nice too.

I did this for a while, it's pretty good but I occasionally came across dependencies that were difficult to install in containers, and other minor inconveniences.

I ended up getting a mini-PC solely dedicated toward running agents in dangerous mode, it's refreshing to not have to think too much about sandboxing.

Regular WD40 should not be used as bearing lubricant!

nit: the last letter of PCB stands for "board"

Do you do this in a professional setting? I'm curious because I did some embedded/uC work about 10 years ago and considering the state of C/C++ SDKs (and IDE support) at the time, I would have expected it to take decades for Rust to get a foothold.

The mainland and Hong Kong still have significantly different visa policies. I'm not sure if it's changed at all since the handover, except for mainlanders entering HK.

If I was letting some random person rent one of my servers without oversight, I'd sure want to see some ID first.

I don't think it's naive, I often do the same in Vue. A pretty useful subset of vue-router can be implemented in less than a tenth of the bundle size.

I don't think 1, 2, and 4 apply to many high profile open source projects.

That's what the GP was saying, I think. Once they get the MAC address, they can find you. Not via software on the phone, from exfiltrating and using shady third parties that collect data from access points, etc.

Do you use AI agents on repos without version control?

No I am not, you took the opposite of the intended meaning.

I'm saying the risk aspect is similar, but I take issue with equating (product) investment and gambling, since one has a potential to create, and the other just shifts money around.

Then aren't most forms of investment based on a "gambler's argument"?

A risky investment is obviously akin to gambling, but to get things built and make a profit you have no other choice.

I remember periods of time when GitHub was down every few weeks, my impression is that it's become more stable over the years.

I don't even travel that much, and still have trouble. Tethering at the local library or coffee shops is hit or miss, everything slows down during storms, etc.

I ran headscale for a while and can confirm local devices could still chat when the internet went down.

I'd agree with you in the past, when KiCAD didn't work as nicely with git. And for a professional project, I'd still at least keep snapshots of every milestone with "compiled" output.

But for a hobby project I really don't see the need. It takes a few extra minutes to install KiCAD, clone the repo, and view the schematic at any point in time.

It's not quite the same though, when I search Google I'm generally directed to the source (though the summary box stuff might cross the line a bit).

With AI, copyrighted material is often not obvious to the end user, so I don't think it's fair to go after them.

I think it's non-trivial to make the AI company pay per use though, they'd need a way to calculate what percent of the response is from which source. Let them pay at training time with consent from the copyright holder, or just omit it.

You don't have to agree with it, but I think it's fair to parrot a take from people who have invested a lot of time and effort into considering why free software is good.

The linked page has a clear explanation for why one might consider nonfree software to be unethical.

I get pretty good results with Claude code, Codex, and to a lesser extend Jules. It can navigate a large codebase and get me started on a feature in a part of the code I'm not familiar with, and do a pretty good job of summarizing complex modules. With very specific prompts it can write simple features well.

The nice part is I can spend an hour or so writing specs, start 3 or 4 tasks, and come back later to review the result. It's hard to be totally objective about how much time it saves me, but generally feels worth the 200/month.

One thing I'm not impressed by is the ability to review code changes, that's been mostly a waste of time, regardless of how good the prompt is.

Residential proxies and other such techniques, I imagine.

I am aware, I live in Taiwan. While TSMC engineers can be poached "move between countries freely" is not true because moving from China to Taiwan is not so easy.

The key word you mention is "replicating." They'll be chasing for a while still, and it's not clear that they'll be able to leap ahead. Copying is much easier than real innovation.

That's a very pessimistic take, or optimistic I guess, depending on perspective.

Looking at the Chinese semiconductor development trajectory, and considering that TSMC won't be sitting on their hands, "within a decade" seems really unlikely.

Wait and see, then change the policy based on what actually happens.

I sort of doubt that all of a sudden there's going to be tons of people wanting to make complex AI contributions to LLVM, but if there are just ban them at that point.

It doesn't really affect your point, but biking 25 miles is going to burn more than 400 calories. Probably double or triple that, depending on their weight and the workout intensity.

Deploy to what? Staging on every merged PR (commit to stg), and prod deploy on every commit to main? That sounds reasonable to me, and I've done some variation of it on most projects for the last 10 years or so without issue.

You're right that once prices have gone up, they rarely come back down. But if the price is the same, when you pay cash you're effectively subsidizing credit card reward programs, and lining Visa/Mastercard/issuer pockets.