To be honest, without the sponsorship from PCBWay I would probably have stopped making videos on my channel.

It’s not a lot of money - but there is an informal commitment that I will try and produce a video a month. It’s also very on brand for my content - hobby electronics with a focus on embedded (ESP32 range of microcontrollers).

I think the videos are entertaining and educational. Actual viewer numbers fluctuate wildly and despite over 50K subscribers - a “successful” video for my channel is around 3000 views (channel is in my profile).

I still find it amazing that I can get PCBs manufactured at such an affordable price. Even SMD assembly is reasonably priced. Short production runs are more than doable at the amateur level.

I learned circuit design and went from not knowing what a ground plane was to having an 8 layer board working in three years.

https://GitHub.com/oro-os/link

It probably wouldn't hold up super well to professional scrutiny but everything on it works.

I would not have been able to do it without companies like JLC. It made an entire industry approachable, which is old fashioned 'good business' IMO.

I absolutely love your project and I hope it will become a breakout success. It has all the right components for a computing environment that is not controlled.

Have you thought about RISC-V implementations of the kernel as well (iirc you're on ARM and on x64)?

Thank you! I'm working on a refactor of the kernel now that has all three, yes :)

Oh, it only gets better. Thank you so much. If you ever get to the point where you have something ready to order please drop me a line (mail in profile) and I'll buy one set to evaluate and if it works well I will get some more people on it.

Can you talk more about the design of the board?

Also, what's that large five pin connector in the bottom left for?

The board is part of the CI pipeline for the OS. The kernel is built in the normal CI pipeline, unit tested, etc. then platform-specific images are built.

Those are picked up by GitHub CI runners (could be anything but I'm using GH for now) that pull those image artifacts and send them over the internet to the board, which stores them on the microSD slot.

Then the board will boot the device-under-test (either by enabling a USB VBUS line, asserting PS_ON and pressing the power button, whatever the device needs) and will serve the image either a via USB mass device or by switching on access to the microSD card directly via a ribbon connector/custom microSD PCB and ribbon cable.

The kernel then communicates over serial back to the Link, which proxies that back up to the CI runner for evaluating test runs, etc.

Everything is configured using MQTT and mDNS. Using async Rust via Embassy for the firmware.

5-pin on the bottom left is for power - 5V 2A 'always on' supply (on the ATX24 adapters that's the 5vsb line), 5V 3A aux line (for VBUS, optional and not otherwise used to power the board itself), a sense line for the aux power (board will shut down and display an error on over-current of the main line if not sensed), active-low aux line enable signal (PS_ON for ATX24 sources), and ground.

This means that it's used to cut power on x86 machines, or to use a stock desktop PSU even for arm/riscv dev boards. In the future I want to make this all rack mounted and have a dedicated power supply for multiples of these.

Interesting thought. PCBWay and JLCPCB sponsoring channels to show use-cases of their capability, thus growing their market.

Would be similar to the distributer/producer of a food item sponsoring channels to use their ingredient in recipes.

Makes a lot of sense.

Long time fan of atomic14, thanks for all your videos!

It's interesting perspective and I'm happy to hear it works well for you.

Are you willing to say how much they pay you?

I watch all your videos by the way. By the PCB Way!

I would add the capability to be able to seamlessly rotate keys.

But otherwise, yes, for love of everything holy - keep it simple.

What’s not surprising is how despite the upvotes this never appeared on the home page.

Very much like humans when they drown in technical debt. I think the idea that a messy codebase can be magically fixed is laughable.

What I might believe though is that agents might make rewrites a lot more easy.

“Now we know what we were trying to build - let’s do it properly this time!”

Potentially, yes, but as with other software, you need to know AND have (automated) verifications on what it does, exactly.

And of course, make the case that it actually needs a rewrite, instead of maintenance. See also second-system effect.

> Potentially, yes, but as with other software, you need to know AND have (automated) verifications on what it does, exactly.

Yes, but even here one needs some oversight.

My experiments with Codex (on Extra High, even) was that a non-zero percentage of the "tests" involved opening the source code (not running it, opening it) and regexing for a bunch of substrings.

>And of course, make the case that it actually needs a rewrite, instead of maintenance.

"The AI said so ..."

I'm wondering how much value there is in a rewrite once you factor in that no one understands the new implementation as well as the old one.

Not only is it difficult to verify, but also the knowledge your team had of your messy codebase is now mostly gone. I would argue there is value in knowing your codebase and that you can't have the same level of understanding with AI generated code vs yours.

The point of a rewrite is to safely delete most of that arcane knowledge required to operate the old system, by reducing the operational complexity of it.

> “Now we know what we were trying to build - let’s do it properly this time!”

I wonder if AI will avoid the inevitable pitfalls their human predecessors make in thinking "if I could just rewrite from scratch I'd make a much better version" (only to make a new set of poorly understood trade offs until the real world highlights them aggressively)

It will make rewrite quicker, not "easier".

When the management recognize a tech debt, often it is too late that nobody understand the full requirement or know how things are supposed to work.

The AI agent will just make the same mistake human would make -- writing some half ass code that almost work but missing all sorts of edge case.

I was involved in a big re-write years ago. The boss finally put the old product on his desk with a sign "[boss's name]'s product owner" - that is when people asked how should this work the most common answer was exactly like the old version. 10 years latter the rewrite is a success, but it cost over a billion dollars. I have long suspected that billion dollars could have been better spend by just fixing technical debt.

That's correct, the more I work with AI the more it's obvious that all the good practice for humans is also beneficial for AI.

More modular code, strong typing, good documentation... Humans are bad at keeping too much in the short-term memory, and AI is even worse with their limited context window.

Would be handy to actually see what these companies do…

That’s not necessarily true. A lot of companies are very risk averse and will sit there creaming off profit and not making any investment.

If someone came to you and said - you have two choices:

Work incredibly hard, raise a lot of money, build a bunch of infrastructure. And at some point in the future you will make some more money.

Or - keep taking your very nice high guaranteed salary for the foreseeable future.

What would you pick?

I hear vending machine ppl often have 1-3 very profitable locations and 10-40 locations that only barely make sense often only because they already are in the business.

I imagine hiring someone to fix or restock them makes a lot more sense if you have 100 machine rather than just one.

It really depends what the goal is. Profit with fiber or fiber with profit?

Here public transport is required to cover all routes. Postal service is the same. Fiber doesn't seem that different?

Depends on how much margin there was. Sounds like there's not enough.

Emotional state, tiredness, drunkenness, a goods nights sleep… the number of factors that drive our responses is ridiculous.

A lot of the railway network uses a “third rail” to carry power. You don’t necessarily need overhead lines.

https://www.networkrail.co.uk/our-work/looking-after-the-rai...

There'll be no new third rail electrification, though (apart from some minor infill, or reorganisation around depots).

The conversion of remaining mainlines to 25 kV overhead AC is going slower than anyone wants, but already over 70% of passenger rail journeys use electric traction (and actually more like 80% by passenger kilometers).

There are an awful lot of low-traffic rural lines that it won't be economic to electrify using current technology, so we'll need to rely on battery electric for those.

Either way, it's largely orthogonal to the problem of electrifying road transport.

Oh, right, I completely forgot about those! To be fair, there are reasons everyone else usually sticks to overhead lines outside of subway systems.

I find it all quite entertaining. A lot of developers are discovering what it’s like to manage “people”. And they are realising that it’s not fun.

If you’ve run a team or managed people it’s quite a familiar feeling of “I’m pretty sure we were very clear on what needed to be done. But somehow, what’s been produced is just not quite what I wanted”

Yeah, but humans can learn, remember, and usually don’t hallucinate. They get better over time on a specific codebase, framework, and system.

That is the hope...

I've laid the groundwork to do this locally, but the current crop of LLM agents just can't close the loop. Trying to get them to finish the job inevitably sends me right back into the spiral. Thinking i should do it manually.

Read my other posts too.

Moving to electric is not some optional thing that you can choose not to do. Fossil fuels are a finite resource. They will run out at some point.

We can either make the shift now or we'll have to do it later. Much better to do change early and invest in it early.

Strictly speaking fossil fuels being finite does not mean we have to move to electric vehicles. We could switch to 'green' hydrogen, biofuels, or synthetic fuels. We'd still switch to renewables (or maybe nuclear), just not necessarily to electric.

When I was 12 my school books said we would run out of oil in 20 years

I’m older than 32 now

When you were 12 it probably said the life expectancy was X years. If you reached X+1 years would you conclude that you would live forever?

Do you believe there is an unlimited amount of oil in the world?

At least until the end of our respective lifetimes