pvdebbe

I don't think you're forced to file jointly?

Is "ML" some UK-ism? I sort of see the connection between machine learning and LLMs but it doesn't seem too obvious to me. And certainly it's not the common nomenclature.

A command is something I shout to an underling or my shell prompt. My shell interpreter takes my command, parses it and calls the necessary tool, be it a builtin or an executable.

I tell Tom to fetch the morning paper from the yard. Is Tom a command? No, Tom is my son.

I disabled quiet mode and I don't know what is revealed.

I have realized the same.

Only it is different 100 photos every year.

I had a linguist attend a CS class and he didn't know how to copy and paste.

Will you update the list for us, a month from now?

UTFv6

I remember a moment in an early job interview. My resumé was all proper for this entry level junior position and everything was checking out. We turned to smalltalk to round up the interview.

"What sort of music do you like?"

(What sorta ques... you don't just throw this in without a warning. ahh I'm blanking out. What was it that I listened to last night??)

"Uhhh Mike Oldfield is one..."

"Ohh... (pause) ewww. But we won't let this hinder your application heh heh"

"hee..."

I got the position and rose to senior rather quickly. I didn't have any interactions with this guy since after this one interview. Maybe for the best. He didn't mean bad, he was just a bit out of his element 2nd-seat interviewing for devs.

GP expressly said the wife does not code: she uses claude.

What was the original publication date for this strip #568?

Also contrast with the busy beaver problem and how much can be done with a small handful of instructions.

Timepieces are a classic example of this.

I really want somebody to mod this into OpenTTD because cargodist was a good step and a nice approximation for a time, it's no match to the "destination in mind" pax. I'm reduced to play commercial offerings like Transport Fever 2. But as far as I know, it would be a considerable undertaking.

Elsewhere in the thread chain:

>[OpenTTD has bad UI]

Hmm, really? It's cluttered with windows and options but I think the mechanics of windows popping and quick dismissing works out for this kind of a game really well. It scales across #n of monitors so well. I run mine on a 43" 4K television panel, no scaling, and I get all my screen estate I need. Works out so swimmingly.

Tell me what happens if you run a Windows machine for say 12 months and decide to switch the GPU from NVIDIA to AMD (or vice versa)? Yeah.

In linux it tends to be a nonissue.

Apple's hardware is their killer business.

The biggest strength of Eshell, for me, is that I can maintain some sanity in Windows environments (with or without WSL). In linuxland, it's a tougher sell. Compared to term-mode and the others, Eshell hacks better and I can fine tune my tab completions. The best thing is that my keybinds integrate better across all emacs modes when I use eshell.

Notepad going the way of Wordpad, EDIT.COM becoming the new Notepad.

What's next, in a few years we're rocking EDLIN when we need to operate on a text file safely?

It ain't extinct shit if it can't even drive the car to have it washed.

Common fonts are gigabyte downloads these days thanks to emoji support.

I love this. Shouting into the void with the distinct feel, hope that if the idea was popular enough, it'd be brute forced back to existing.

I noticed that the input is not being treated any way before hashing. I'd remove all non-letter characters, and then lowercase everything before hashing to help with some unnecessary misses.

I'm still using Firefox and loving it actually. Other browser engines don't support "zoom text only" anymore so my options are limited. And to my knowledge, there's nothing as good as uBlock Origin for those webkit/blink based browsers...

Yes, Firefox constantly introduces new degradation in UX but so far they always offer opt-out mechanisms for even the most obscure things...

>just like Scorsese intended.

>before Scorsese made The Godfather

Can you let me in on the joke?

I don't know what kind of a joke you tried here, but I think a vast majority of TV screens can be put in game or PC mode, and all the input lag and stupid picture processing goes away. I run a 43" LG 4K TV as a PC monitor and never have I had a (flat screen) monitor with a faster response rate! My cinema TV is an old FullHD 42" Philips that has laughably bad black levels. I run it also in PC mode but the real beauty of this TV is that without further picture processing it produces nice and cinemalike flat color that is true to the input material that I feed it. Flashy capeshit will be flashy and bright, and a muted period drama will stay muted.

Fantastic tools you've collected here! My rationale for building own decks for my wife is that I'm intending to start slow and easy, and I'll build new cards that are closely tied to days' lessons I'm giving her. I'm hopeful that after a while she gains confidence to start going through premade decks. With anki and similar tools it's important not to just memorize words without having some handle on how to build sentences etc. I spent a lot of time learning Japanese that way, only to find that I maybe memorized words but to build sentences with them...

Don't you just put tick "know this already" or whatever mechanism is used. It'll be asked a couple of times but it shouldn't be "relearned from scratch".

Perfect timing. I just started to teach my wife Finnish so that she'll have easier time with real language lessons when she gets her paperwork sorted out to move here. And I've feverishly been looking for a self-hosted SRS system where I can feed new content to the "decks" and she can consume it on her schedule. Making micro decks that she'd import in Anki wouldn't be very convenient. This would seem perfect to me.

I'm happy to hear other suggestions too?

My first instinct is that users abuse it like they do any other report/downvote mechanism. They see something they just don't plain like, they report it as AI slop.

Oh, you think it's undervalued?

If the target is superintelligence, then AI shouldn't be learning from humans.