Your perspective might be why I enjoy microcontroller work. I love getting to know everything about the system, reading hardware documentation, and getting the low level parts to work in a highly deterministic way.

I use ATTiny85 cores when a ESP32 costs almost the same, but the 85 only has 256 bytes of SRAM and five I/O pins so I can track it all and ensure it will do exactly what I want.

The same meme with “wiring and lights” at the top. Then you descend to motors, transformers delta-y phases, RC and RL circuits, op amps, BJT circuits, reverse bias what?, differential equations, and eventually signals and systems.

The summary that I liked from the last post was “python is the second best language for everything”. There’s always something specialized and better for every given job. But, if you want one tool that’ll do a solid job everywhere, python is your go to.

We’re entering the ‘blockchain for every need’ stage. Expect massive money to flow into scams, poor ideas, and outright dangerous uses for a few years .

Before Blockchain we had ‘the web’ itself in the dot com era. Before that? I saw it in basic computing as a solution to everything.

I had a student came into office hours asking why their program got a bad grade. I looked and it didn’t actually do anything related to the assignment.

Upon further query, they objected saying that the CI pipeline built it just fine.

So …yeah… You can write a program that builds and runs, but doesn’t do the required tasks, which makes it wrong. This was not a concept they’d figured out yet.

As per all too often, the functional programming world invented them. Haskell (and its ilk) usually has all the future cool stuff already. Then python picks it up, then it moves over to C#/Java, then C++ says “mee too”!

Yeah, those durn data size fields. At first you’re like “why would you do this? It’s specified in the spec, right?” Then you start consuming the data stream and go “oh, yeah need this”.

I was doing some driver work for a real time location tracking board. The serial stream protocol was very well documented and designed. Plenty of byte length count fields, though.

This approach is so much nicer than the threading/queuing approaches we used to have. One async showed up, a ton of the work go pulled out of protocol handing and distributed subsystem sync efforts.

Long lived the multi threaded C++ server buffer! Today, async beging to rule the roost.

If the PoE is stable, then it’s a nice and relatively unique board. Not sure about the NPU support. There’s a ton of boards and chips coming out with those claims, but I’d like to be able to get clearer info on drivers and library compatibility.

It’s wonderful how included and valued nerdiness is these days. Being interested in anything non mainstream in the conformist 80’s was hell outside of a tight friend group.

There’s all kinds of wacky taxes, regulations, and barriers to prevent the US industry from having to compete with the world. One such example is the Chicken Tax:

https://www.investopedia.com/terms/c/chicken-tax.asp

That one keeps the Toyota Hilux out of the US.

Time is possibly one of the hardest things to handle properly for a coder. There’s plenty of hard problems (network loss, 3 phase commits, etc), but time stand out as really annoying.

Another one is colors. All it takes is one library to encode colors in a weird way and then mapping them between libraries is a mess.

I like the UNIX/Unix-like category since it could include Linux, OSX, MacOS (current), and ChromeOS, but it doesn’t. It would look much different, percentage wise, if those all summed up.

This also is only desktop OSes. Add in servers, smartphones, and embedded systems and it’s a much different landscape, but I’m sure MS DOS will still hang on forever.

One of the interesting aspects of humanity is how much people like given text editing methods. There’s a handful of approaches and after learning one or two, people really figure out what works for them.

I am more than happy to say I like vim, but in the end you should use what you like best. Just done be surprised when I can write and edit a ton of text really fast while your nano session is comfortable, but slow.

If an IDE doesn’t have vi key bindings it isn’t going to be used by me. That’s what finally get me to change from terminal only dev to vscode. Until I found the vi editing extension the IDE wasn’t of interest.

Yes, vi is just that good.

I hear emacs bindings are also great, but I just know how to save and exit from emacs.

Such great Exploits of a Mom: https://xkcd.com/327/

VPNs are not the security panacea that marketers would have you think they are. Using a VPN does provide some obfuscation as to your origin, but it does change your trust model. The VPN service provider may tunnel your traffic through your ISP to hide data from the ISP, but now it’s visible to the VPN service provider instead.

There are plenty of use cases for a VPN, but just like any other technology or service, you need to know what it actually does so you know what it actually achieves or doesn’t achieve.

You can write comments for the needed design/flow of giving blocks. It’ll do its best to fill in something that’ll mirror your description. Then you get to fix what it generated.

I find it more successful as a really advanced autocomplete.

I’ve been doing a lot of using, testing, and evaluating LLMs and GPT-style models for generating code and text/prose. Some of it is just general use to see how it behaves, some has been explicit evaluation of creative writing, and a bunch of it is code generation to test out how we need to modify our CS curriculum in light of these new tools.

It’s an impressive piece of technology, but it’s not very creative. It’s meh. The results are meh. Which is to be expected since it’s a statistical model that’s using a large body of prior work to produce a reasonable approximation of what it’s seen before. It trends towards the mean, not the best.

I’m not surprised at the funding. During the Dot Com bubble in 2000 you just had to say “Web” and someone threw a brick of Benjamins at you. We’re seeing a similar thing around AI/GPT LLM tech right now.