steve_taylor

It's technically possible to achieve a 100 CRI score with a broad enough range of frequency spikes rather than the continuous range of frequencies you get from incandescent bulbs. So even the best LED bulb isn't going to match incandescent.

I'm trying to get my hands on some decently high wattage (100W ideally) incandescent bulbs because it's really been bugging me lately that while I can still focus just fine, I have trouble making out details and some colours in artificial lighting. I've come to the conclusion that there's really no substitute for light generated from heat.

Sounds like a cheap way to get my hands on some floppies.

Devices get slower over time due to their operating systems becoming increasingly complex as they're updated. This increased complexity is due to the OS developers taking advantage of the newer models' improved hardware. This leaves owners of older models facing a decision between staying on the older OS version and retaining the current performance vs. updating to the latest version and risk performance degradation. But it's not even that simple. Over time, app developers stop targeting older OS versions, which means you could be stuck with old versions of apps, some of which will refuse to do anything until they're updated.

The way to solve this is for device vendors and their OS developers to simply stop updating the OS in a way that takes advantage of new models' improved hardware because that same code slows down older devices.

Finger was only part of the social network. The multi-user aspect of Unix itself combined with standard tools such as who, finger, write, mesg, talk, mail, biff, etc. combined to form a fairly complete text-based social network.

> Not quite. The rightmost one initially has 6 closed cards and 1 revealed card which can be a king, so you can put 12 more cards on top of it, for a total of 6+13 = 19 cards. So I'd make 7 arrays for 20 cards each, just to be safe. Or just use 52-long arrays for every table slot, that too is an option.

You're right. And I hadn't considered tracking the index of the first revealed card.

Anyway, my point is that you can map out the memory allocation quite easily and it's static and small, so you don't have to worry about malloc and free.

So we're going to ask them to not advance their operating systems and apps to take advantage of newer hardware?

How does this compare to diskDedupe, which has been around longer and is much cheaper?

Parents can stop giving kids smartphones.

I've been running Claude Code with --dangerously-skip-permissions in a Docker container for the last month or so, allowing me to get up and stretch my legs while it does its thing. I definitely wouldn't want to run it unsandboxed.

I'd start by identifying all the known boundaries:

52 cards, so each card can be represented by a byte.

Cards are arranged into 4 stacks of up to 13 cards each, 7 columns of up to 13 cards each, and a pile of up to 24 remaining cards. (You don't need two piles of remaining cards.)

That's 167 (4 * 13 + 7 * 13 + 24) bytes for the cards, 12 bytes to store lengths, and 1 byte to store the current index in the pile of remaining cards. It's a fixed number of fixed-length byte arrays and a fixed number of bytes to track lengths and one index.

You can't polish a Hurd.

It'll take more than a mouse jiggler to fool this new bossware.

It reminds me of tax law in many countries. You can follow the letter of the law, but if the vibes are off, you can still be found to be in breach of a vague catch-all provision (e.g. economic substance doctorine in the US, GAAR in Canada/UK, Part IVA in Australia, etc).

You're absolutely right. Your belt-and-braces are earning their keep.

The only time I use Reddit is when I land there via a Google search. Posts get flagged as mature at the drop of a hat, so my go-to has always been to replace www with old. Now that it's gone, I'll just skip Reddit links.

It was closed-book. Take-home, closed-book seems like an oxymoron anyway.

A lot of the cheating students would have concluded that other students are going to cheat and get a much better result, so they might as well do the same.

The fault lies squarely with the professor.

> including the half ass Australian solution.

It was designed to be half-arsed so Digital ID can come along and save the day. Australia's Digital ID opens up to the private sector on 30 November.

The defining aspect of Bedrock is C++, not DirectX. Bedrock runs on numerous platforms where DirectX isn't available.

Minecraft isn't exactly be the type of game to showcase a 3D API's supremacy.

The Australian one will label them as Unverified.

Personally, I'd prefer them to be blocked. If it's important and legitimate, they'll register.

As an Australian, I'm happy to hear this, but also annoyed that a lot of legitimate SMS from companies don't use branded sender ID. I'm not sure why, but my guess is that SMS gateways charge more for it and businesses don't want to pay the extra cent or two.

AI has a way of exposing people. In this example, students who are there to get a degree from a prestigious institution, rather than to learn, are prone to take perceived shortcuts and proceed to come unstuck when their AI isn't there to do their work for them, such as in an exam.

> But there's one performance-related area where the Framework pulls ahead—a little—and that's sustained performance. When running a heavy workload like HPL (a FP64 HPC task, that taxes the CPU and RAM constantly for many minutes), the Framework's fans allow it to throttle less than the Neo.

People are seeing big gains in sustained performance on MacBook Neo with a simple thermal pad mod. The disadvantage is the underside of the Neo can get hot, but that's not an issue if it's sitting on a desk instead of your lap.

I was all in on Svelte and SvelteKit until I started encountering CSS weirdness caused by a bug that the Svelte developers said is "by design", namely that components' CSS isn't removed from the document after the last instance of that component is no longer rendered. This resulted in a situation in which styles became dependent on the navigation path the user takes, leading to weird an unpredictable layout issues. I couldn't stomach solving this by using Tailwind.

Then Svelte 5 came along and made Svelte more like React. At first, there were just a few simple runes, but then the runes started proliferating like crazy to solve other runes' problems. At that point, Svelte was dead to me and I went back to React/Next.

The right path for Svelte to take would have been to continue to refine Svelte 4.

I love how I can see the HTML being streamed onto the page in real time, like the good old days of dialup when images gradually rendered from top-to-bottom.

It's for Google.

Perhaps, but I think that by the time we're that far advanced, strokes will be entirely preventable.

There's nothing stopping you from setting up your free (as in freedom) slice of cyberspace for you and your friends, for now.

Looking at all the new and proposed laws coming through, I don't think we'll have those basic freedoms all that much longer.

Taking a brief look at wizball.asm, it actually used 200 Hz interrupts on both PAL and NTSC. The timings are shown from line 39.

RAM can be increased to 16 MB and CPU speed to 48 GHz.