Yup. When cheap PC clones came around, everyone had that one Com, LPT, or VGA port with the missing screw terminal. Fortunately you need zero of those for the port to actually work.

What’s kind of amusing is that all those old ports with the screw-in fasteners were moving around anywhere from 3v to 12v at about 45mA on the high-side. Meanwhile, USB-C can move 240W and has nothing to prevent disconnecting while sending enough power to run a power tool.

And to monetize the display of video. Licensing of that port’s supporting technology is holding everyone back.

There’s gotta be a way to fingerprint the output though. Like some kind of shibboleth that gives the model away based on how it responds?

True, but IMO it makes a lot more sense to support markdown under Wordpad than Notepad since the controls and doc-style layout/rendering support is already there.

it only exists so they could easily shove copilot into it.

No idea. I want to say ‘yes’, but the last time I installed Power Toys was under Win2k.

Tabs were a welcome addition, but that’s where the good idea train swiftly leaves the rails.

What we needed was a built-in hex editor, and maybe some better tools for working with unicode that you can’t just type in on whatever keyboard you have.

Instead, they turned it into WordPad, which we already have.

“_comment”

I appreciate the workaround here, and I’ve tried this in production environments to one degree or another. This usually fails due to another problem: the number of systems that think unexpected JSON keys are an error, is is too damn high.

I’ll give it a shot, but I must be able to use my laptop to vibe-interview.

Or you’d get lucky and some other program you installed happens to have the right dependencies. Just copy them to the application install dir or to C:\windows\win32\ and off you go.

Yeah, stuff like that continues to be the best use-case for windows virtualization. Sounds a lot like trying to upgrade the BIOS or Firmware on an older PC; often the installer is some binary that only runs on Windows of the same vintage.

Backwards-compatibility with older web browsers so engineers can build websites for them, is another. I’ve also heard of industrial automation (e.g. CNC machines) being married to Win2k or WinXP, so being able to run an old OS on new hardware is crucial.

Windows, can I run this 25 year old software I just installed?

Thank you for your service.

Oooh, rocking an HP? I too like to live dangerously.

But seriously, that’s good to know. Those are probably easier to come by out in the wild. It really looks like Thinkpads go from office deployments straight to refurb companies these days. I never see them at thrift stores, and I’m not brave enough to dumpster-dive at e-waste.

Sometimes, old machines are survivors. Beware of confirmation bias when trash/thrift-picking cheap systems though. IMO, Thinkpads can be tough as a coffin nail. Including work systems, I’m on number 8 at this point with no hardware failures in sight.

That said, I have a very lightweight Acer that’s about a decade old with the worst keyboard and trackpad ever manufactured. It also performs like a slug, even with Linux on it. Still, it refuses to break so I can get rid of it.

This essay is brought to you by Raid: Shadow Legends.

I mean, that happens with CloudWatch all the time. It’s the most plausible part about this.

I use a system prompt that forces it to ask a question if there are options or if it has to make assumptions

I’m kind of amazed that even works. I’ll have to try that. Then again, I’ve asked ChatGPT to “respond to all prompts like a Magic 8-ball” and it knocked it out of the park.

so I start a new chat frequently.

I do this as well, and totally forgot to mention it. Yes, I keep the context small and fresh so that prior conversations (and hallucinations) can’t poison new dialogues.

I also will do the same prompts on two models from different providers at the same time and cross reference the idiots to see if they are lying to me.

Oooh… straight to my toolbox with that one. Cheers.

That’s been my biggest problem with the current state of affairs. It’s now easier to research newer tech through an LLM than it is to play search-result-wack-a-mole, on the off chance that what you need is on a forum that’s not Discord. At least an AI can mostly make sense of vendor docs and extrapolate a bit from there. That said, I don’t like it.

When writing code, I don’t let AI do the heavy lifting. Instead, I use it to push back the fog of war on tech I’m trying to master. At the same time, keep the dialogue to a space where I can verify what it’s giving me.

1. Never ask leading questions. Every token you add to the conversation matters, so phrase your query in a way that forces the AI to connect the dots for you
2. Don’t ask for deep reasoning and inference. It’s not built for this, and it will bullshit/hallucinate if you push it to do so.
3. Ask for live hyperlinks so it’s easier to fact-check.
4. Ask for code samples, algorithms, or snippets to do discrete tasks that you can easily follow.
5. Ask for A/B comparisons between one stack you know by heart, and the other you’re exploring.
6. It will screw this up, eventually. Report hallucinations back to the conversation.

About 20% of the time, it’ll suggest things that are entirely plausible and probably should exist, but don’t. Some platforms and APIs really do have barn-door-sized holes in them and it’s staggering how rapidly AI reports a false positive in these spaces. It’s almost as if the whole ML training stratagem assumes a kind of uniformity across the training set, on all axes, that leads to this flavor of hallucination. In any event, it’s been helpful to know this is where it’s most likely to trip up.

Edit: an example of one such API hole is when I asked ChatGPT for information about doing specific things in Datastar. This is kind of a curveball since there’s not a huge amount online about it. It first hallucinated an attribute namespace prefix of data-star- which is incorrect (it uses data- instead). It also dreamed up a JavaScript-callable API parked on a non-existent Datastar. object. Both of those concepts conform strongly to the broader world of browser-extending APIs, would be incredibly useful, and are things you might expect to be there in the first place.