shdon

Cool video, though the AI voiceover is jarring.

Somebody obviously coined the word as a self-referential joke. And somehow it stuck. That makes it memorable.

Speaking of things that stick... arachibutyrophobia is the feat of getting peanut putter stuck to the roof of your mouth. (I admit I had to look that one up, as it's not nearly as memorable, though I knew the word existed).

I thought that one was pretty well known. But then, I can also rattle off Llanfairpwllgwyngyllgogerychwyrndrobwllllantysiliogogogoch at will.

Bill Gates famously said something along the lines of "Measuring programming progress by lines of code is like measuring aircraft building progress by weight". Now we're measuring AI effectiveness by the number of tokens generated? Plus ça change...

And yeah, what is it with C level executives and AI? I'm pretty much getting it forced on me by the higher ups. Management has insisted all employees come up with ideas on how to "leverage AI most effectively for the business". And now we're hiring an AI expert... For the small internal IT team of a publishing company...

At one point, I spent so much time fixing the disastrous results of people using ChatGPT to generate oodles of CSS that they would put in the WordPress Customizer that I had to resort to disabling the custom CSS functionality entirely. All because they wanted a link to be a different colour.

I have been eaten by many of those.

I've created my own adventure game engine starting in the late 1990s. Only learned about the existence of AGS many years later. Although my own engine allows much more flexibility than AGS, there is no userfriendly IDE and besides the runtime, it's mostly just a bunch of separate tools. I have to applaud Chris Jones for going all the way, it's really quite impressive.

Thank you for reading it.

https://www.shdon.com/ I post infrequently, but have been maintaining a site for nearly 30 years now, 25 of which at this address. It contains random musings, some tech content, my game development efforts, and showcases some of my pixel art.

Because it would make scrolling more frequent? For multicolumn text to reduce scrolling, the column height would have to match the available viewport height. And if your text exceeds what can fit in multiple columns on a single screenful, scrolling also becomes awkward, because you'd have to scroll exactly to the next screenful to have any consistency. Multiple text columns only make sense on extremely restricted layouts, or where the volume are entirely independent instead of a single flowing piece of text, or where there is still a direct horizontal relationship (like annotations or translations beside the main text).

The third common CD-ROM interface that wasn't IDE was Mitsumi.

I run several job boards, and though there hasn't yet been any postings requiring the candidate's using AI (also not all that likely in the particular field), I am noticing a definite increase in the number of lists with emoji preceding each bullet point, and the use of em dashes. Personally, if I were a job seeker, I'd find that off-putting just as much as if I were presented with the requirement to use AI in my job.

Post on his actual channel: https://www.youtube.com/post/UgkxlxGXbAp493rLsgFSTZkLyuz_dEw...

How long before spam filtering is also done by an LLM and spammers or black hat hackers embed instructions into their spam mails to exploit flaws in the AI?

And for cows

Technically there still is, for a few more months. There is a 32-bit version of Windows 10, and that won't end mainline support until October (and then a while of extended support, for those who wish to pay for such).

These were existing MS-DOS programmes that had already shipped. They wouldn't have shipped with a Windows icon as they were made before that Windows version existed (or at least shipped) and weren't even intended to run on that platform. Once Windows had shipped, and software vendors started making software for it, they will of course have included their own icons. The "why" is simply Microsoft wanting to make Windows play nice with users' existing software, and thus enhancing the user experience.

Surely that's Doom8088 rather than the original version if this thing truly emulates an XT level machine (or rather an 80186 CPU)?

I do indeed check against both DNT and Sec-GPC (and navigator.doNotTrack and navigator.globalPrivacyControl in JS) basically treating them identically. GPC is ostensibly not about tracking itself, but about sharing data, though I just figured that data that isn't recorded can't be shared either.

I wonder how many web developers actually honour Do Not Track. I do, in all the websites I've made for my employer too, but I think I'm only getting away with it because my employer doesn't know. I've even made it so that browsing with Do-Not-Track enabled also skips the cookie consent banner and just assume the user wants no cookies other than the strictly necessary ones (like their session/login cookie), and doesn't include Google Analytics, instead just upping a single view counter on the page, with no PII in there.

Then it would make a lot more sense to check event.pointerType == 'mouse' vs event.pointerType == '' (for keyboard)

The CD-ROM version of SQ4 was especially insidious. It has many timing issues that the floppy version just plain didn't have (including the one you mention). Also, the scan quality of the VGA backgrounds was just better in the floppy version, with a much better gamma curve, the CD version looking rather more washed out.

It's assume that the CD version was developed alongside the floppy version and that the code forked at some point. The floppy version got the fixes, whereas the CD version didn't.

Somebody in the retro adventure gaming community with the handle NewRisingSun made a patch that used the floppy disk background images and has patched script resources to solve the myriad timing issues, whilst integrating seamlessly with the speech of the CD version.

Mullenweg calls it a fork. I could see that being somewhat okay if it's indeed for security fixes, but removing the upsells is petty at the very least. But a fork doesn't take control of the original, so I wonder what they did there? Perhaps a redirect from the ACF entry to SCF?

To be honest, none of this makes WordPress look good... It just seems like a douche move.

Same as what we do. Only one additional thing: when people log in to their accounts, it switches to the preferred language for that account (and any changes in language selection whilst logged in are also saved into the account preferences).

The article just says those 30% of people weren't convinced the version with the calculator "would do better", not that it "is a bad idea". Granted, they might have thought that, but it seems quite a leap. They could just have easily thought it would make no difference, or assumed the version with calculator was underperforming because of the cases where it gave the wrong answer.

Or write an entire book without them, then publish a second edition with an appendix of just punctuation marks, for the reader to place wherever they feel is appropriate. Oh wait... Timothy Dexter already did that... https://en.wikipedia.org/wiki/A_Pickle_for_the_Knowing_Ones

To reach the exact number, I would think that:

- if the number n is prime, it would be an SC + (n-1)*P, making the total number of actions n + 1 (as SC counts as 2 actions)

- if the number n is composite, you'd need to factor it, find the smallest and largest prime factors (pf1 and pf2), which would necessarily multiply into n again, and the total number of actions is (pf1 + 1) + (pf2 + 1)

If my reasoning is correct, reaching 100007, which is prime, would require 100008 actions. And reaching 100001, which factors to 11x9091, would require 9104 actions.

To reach at least n, I reason:

- 1 or 2 action sequences do nothing, as an SC is necessary followed by at least 1 P to have any effect on the output

- 3 actions (SCP) allow you to at most double the size

- 4 actions (SCPP) is more efficient, allowing you to triple the size

- 5 actions (SCPPP) is yet more efficient, allowing you to quadruple the size

- 6 actions of (SCPPPP) would allow you to quintuple the size (still more efficient than SCPSCP)

- 7 actions gets interesting as SCPSCPP, SCPPSCP and SCPPPPP would all sextuple the size

- 8 actions is the tipping point, where SCPPPPPP would hextuple the size, but you can achieve a better result by combining the previous steps - only SCP+SCPPP [identical to SCPPP+SCP, as multiplication is commutative] would make 8 actions, leading to octuple, or SCPP+SCPP leading to nonuple

That means that SCPPPPPP is no longer a viable choice and that more than 7 actions always means it's more effective to combine shorter sequences to muliply things out. The order of them does not matter. As x^y gets bigger faster with increases in y than with increases in x, and we can see from the list above that, whenever we starting combining sequences, multiplying the size by 9 is the most efficient we can do, for any significant large number n, we'd just wants as many repetitions of SCPP as possible in there, only switching to something else down the list when reaching the end goal can be done in fewer steps than multiplying by 3 repeatedly (and there's only a finite number of combinations there, when the number of actions is not evenly divisible by 4, not allowing you to use the SCPP sequence).

Disregarding any further occurrences of SCPP, you can do a combination of 1 sequence:

SCP (x2), SCPPP (x4), SCPPPP (x5), SCPPPPP (x6)

or a combination of 2 sequences, once again disregarding SCPP and more efficient shorter combinations: SCPPPSCPPP (x16), SCPPPSCPPPP (x20)

Even SCPPPPSCPPPP (x25) is already less efficient than the SCPPSCPPSCPP (x27) sequence of the same length.

So I figure that to reach at least n, you would always end up with as many repetitions of SCPP as possible, followed by one of the 6 above combinations (of one or two sequences, at most 11 actions). Any sequence of 12 actions or more could be more efficiently reduced to ones containing SCPP again.

Fortunately it's now been approved, so it looks like they cleared that backlog quickly or it was overstated.

Eep, I just learned the AMO approval process is currently backlogged for weeks or even months. The GreaseMonkey user script is identical, though, and you can peruse the source to make sure it does nothing untoward.

Having the HN comments side by side with the article is an awesome idea! I took the liberty of creating a simple Firefox extension for that. It is still pending approval by AMO, but once that is done, it will be available @ <https://addons.mozilla.org/en-GB/firefox/addon/hn-side-by-si...>

There is a divider that you can drag sideways to change the split ratio. Note that it creates an iframe inside the HN comments page, which means some websites (especially high-profile ones, unfortunately) can't load, as their security settings forbid embedding - there is nothing I can do about that, as it's a browser security feature.

In case you prefer a GreaseMonkey user script, I've got you covered too @ <https://pastebin.com/iVVYUUQ6>

Screenshots @ <https://imgur.com/a/2y41J0Z>

That was indeed my guess.