In my experience, GUI was horrible because of the GIL. Python is nice because you can write and read it quickly, and packages like numpy allow you to use very natural syntax. Most of the numerical packages wrap C/C++/Fortran binaries which are very optimized. You just have to keep in mind not to do that kind of stuff directly in Python, write tight loops, etc.

IDK about Sweden. In the US it seems to pay less, and the jobs are concentrated in different regions than other SWE jobs (which kind of mitigates the lower pay). I did embedded development for a couple years right out of college, then moved to full stack web development. I’m pretty sure the job market in the US is shit for every software development field. I think embedded development is harder to offshore (unless the entire project, electrical engineering and all is offshored). I did find the work very fun; often working in the physical world, and trying to find ways to stay within the CPU/RAM/Flash/Peripheral constraints. I remember once, I had to reverse engineer a piezo-motor controller by scoping the very expensive proprietary dev board, then I reimplemented the control signals by using a microcontroller’s audio playback interface.

You can learn how to use AI coding tools in a week at most, and there’s no telling if next week they’ll be a new harness or loop or whatever that becomes the new trend, making everything you’ve learned so far obsolete. Nobody’s being left behind. AWS also isn’t hard to learn if you’ve done the similar infrastructure stuff yourself. And git isn’t hard to learn if you’ve used other version control systems.

Not a electrical engineer, so dunno the answer.

Can always just try that test circuit and investigate with a scope? Or play around with SPICE? I imagine any low capacitance mosfet would work as a driver?

Figure out what the executive’s day-to-day workflow is like, then make agents to do their job. Either they’ll get an understanding of how inadequate it is, or they’ll be able to just sit around doing nothing useful (assuming that’s not what they do already).

I already covered that in my last sentence. Those things aren’t the things that take the bulk of the time, in my experience; I guess other peoples jobs are different. Generally, everything I write needs to maintained long term, or at least until the company goes under.

I’ve tried (and still try) to find good workflows using AI, but I don’t think I’ve found anything that saves time. To get the same quality using AI, I have to go back and forth with it a lot, and review all the bloated bullshit it generates. If you don’t care about quality or the thing even working correctly (e.g. prototypes/POCs), it’s really fast.

Been thinking about trying these and crypto for online payments. For in-person, cash is fine for most things (barring cashless kiosks like EV chargers).

IIRC, there are few places in the US where it’s legal to just plop an RV on an undeveloped lot and live in it all year. I.e. you need a proper well and septic at the least (these are expensive). Many counties also have laws specifically to prevent people from doing this (I.e. only allowing camping on your land 14 days in a 30 day period). Many people break these laws, but they do exist.

It can be cheaper, but only if you give up a lot of conveniences. I.e. you can boondock on BLM if you move every 2 weeks. Would want solar, a large battery bank, and probably a backup generator. Will want to always be conserving power, fuel, and water. If you want to camp with hookups all the time, it would be more expensive than an affordable apartment.

Is it people deleting specific posts, or people deleting their accounts? I do find it kind of weird that when you delete your account, it deletes all the data you contributed. Maybe it’s the right choice though, IDK.

The trend I see are the Mac Minis with a lot of unified memory. These are typically very well off people though. Prices for even old GPUs like 3090s are ridiculous now. I don’t think connecting 2 machines over Ethernet would work well, but putting 2 GPUs in a single machine does.

Replace “sustainable,” and the bit about profit and capitalism, with “efficient” and “corruption and un-free markets,” then this is a common right-wing talking point (back when the right wing tried to engage intellectually, at least).

In my unscientific opinion, the current population is unsustainable, and there’s no known ways to make it sustainable enough to support the population in the long term (I hope there will be, of course). The most sustainable framing practices are less intensive and result in less output per acre. That’s just about survival, ignoring quality of life. I’ve heard it claimed we’d need 5 Earths for everyone on Earth to live a first-world-like lifestyle. Granted, we should drastically change our lifestyles.

Climate change will also likely lower the human population the Earth can support, and I think we will likely adopt even less sustainable practices to make up for the loss, accelerating our own demise; kicking and scratching and bringing all the ecosystems of the Earth down with us.

If it was good, the AI companies wouldn’t sell its use. They’d be using it to write all kinds of new cool software, dominate financial markets, etc.

I did most of this a long time ago when the team was 2-4 programmers. Moved to PRs when we started bringing on interns. No way the business owners would’ve allowed pair programming. Pair programming may be better and more productive, but the decision makers are often more ideological than rational.

I guess there’s a blurry line between an “application” and a web site. But, generally, for open source applications, I’d rather them be native. Preferably, the applications would actually be native (e.g. Qt, GTK, Jetpack Compose, etc).

I’ve noticed a lot of consumer embedded products have been using PWAs for flashing and configuring devices (Meshtastic, Meshcore, Betaflight, am32 configurator, etc). I have to open Chromium to use them because they don’t work with my normal browser (and I probably don’t want my normal browser being able to access arbitrary USB devices).

Yeah, I think it’s overkill unless you have 20+ nodes. When I switched from just a few VPSs with containers deployed with docker-compose, to deploying and administering a highly-available k8s cluster of 9 nodes, I’m pretty sure our cumulative uptime actually went down. There’s ironically more points of failure that can bring stuff down due to the extra complexity.

Meh, disagree with a lot of this.

AI tools can generate functional, adequate, perfectly average code

Not in my experience.

The outsourcing era taught us that the expensive part of software was never writing it. It was understanding it well enough to change it safely, to debug it under pressure, to explain to the next person why a particular decision was made at 2 a.m. on a Tuesday.

Since AI is adequate, just have AI change, debug, and explain it. You don’t even need devs running the AI. Have AI generate intent. Just have AI scrape Twitter for people complaining about applications they wish existed, and have the AI make them. Let AI do market research. It’s supposedly perfectly adequate.

Specifically, I’d love to hear your thoughts on how to better optimize the context window limits when dealing with massive raw outputs (like huge DNS dumps or nmap scans) before feeding them back into the LLM’s memory.

Launch a sub-agent that reports its findings back to the main orchestration agent. If it’s even too long to fit in the sub-agent’s context, you can chunk it up, have a sub-sub-agent per chunk report back to the sub-agent for a shorter summary, just the interesting lines per chunk, or “no relevant lines found” or whatever. Can get even fancier by allowing the sub agent to use tools like grep, head, tail, etc on the text to search it instead of reading the whole thing directly.

Surprised you’re not using LangChain/LangGraph as it makes some of the things you’re doing easier. But it looks like you’re vibecoding this anyways, so it’s just doing whatever Claude Code or whatever decides to do. My suggestion would be to code it yourself with minimal AI assistance, as this will just turn into an unmaintainable mess as time goes on, and eventually, the AI coding agent will get stuck and be unable to really progress.

Now that I think if it, you could probably get Claude Code or OpenCode to do everything this project can do by just installing all the tools needed in your environment, creating a new empty project, telling it what tools are available in your environment, allow it to download any other tools necessary, and prompting it to do the recon (may need to use an abliterate, heretic, or otherwise uncensored model to do some things).

TIL LibreWolf randomizes some fingerprinting targets.