At my previous enterprise we had a saying:

Security: we put the ‘no’ in ‘innovation’.

It's also revealing to do the maths on how much you'll spend on streaming if you assume you'll keep your account for the rest of your life.

I'm ~50. Let's say I live to 80. 360 months × AU$25 = NINE THOUSAND DOLLARS WAIT WHAT?

I mean, sure, I might spend $9k on Bandcamp or whatever. But I dunno really if I would.

Later that day

Damn though, Apple Music is convenient...

Consider that if you spend 9000USD on bandcamp you can pass the files on to your great grand children

I've been buying used CDs on ebay when I can get them for ~8USD per album, and buying FLAC/ALAC on bandcamp and qobuz for anything that's hard to find. A couple of albums that aren't streaming I had to pay 30-50 USD for a used CD, Ecstatic by Mos Def, Parabolic by Aoki Takamasa. It's kinda fun to find out what music is "rare" and what music is cheap.

Jellyfin + Finamp is a solid combo, and a flash modded iPod 4th gen (last one with a black and white screen) to play music in the car. It's a good feeling to know none of my albums will ever disappear. (To be sure, albums have disappeared from Qobuz, and now they have a message that says 'be sure to download after purchase !!')

iTunes 12.13 is actually a solid music player on Windows. Ripping CDs works great too. There's no iTunes that runs on the latest MacOS tho, since they supplanted it with "Apple Music". Kind of ironic, but Windows has always been bigger on backwards compatibility.

Small-blog-aggregators like Kottke are a goldmine. The trick, of course, is to remember to follow the interesting people he surfaces.

And RSS. It never went away. Just use RSS.

Long live RSS.

Well, if it helps anyone, I just don't think it's possible to set up a new Facebook account then use it to start a business page. My partner just tried it - her real email, nothing weird - and after 4 hours of configuring settings, blocked. Did the video thing on the iPhone, appeal rejected.

I guess you have to have actual history of being logged in to and using Facebook so, lol, no thanks.

Friend, I know. See above: cancelled any number of accounts over the years.

Alas, I'm trying to run a business. I'd like to reach millions of 'normal people' easily and cheaply. If you know of another way (and I mean literally any other way (that isn't Google)), I'm all ears.

What? You are not "effectively making your chats globally accessible".

There is no situation in which I could access your chats. If you disagree, kindly explain how I do that.

> There is no situation in which I could access your chats. If you disagree, kindly explain how I do that

You are dead wrong here. Let me explain.

Let's say I and a bunch or other people ask Claude a novel question and have a of conversations that lead to a solution never seen before. Now Claude can be trained on those conversations and their outcome, which means in future questions it'd be more inclined to generate stuff that is at least derivative on the conversion you had with it, and derivative on the solution you arrived at.

Which is exactly what the OP hints at.

> Let's say I and a bunch or other people ask Claude a novel question

Not that ‘novel’ then, is it?

You know as well as I do that to extract known text from an LLM by 'teasing the prompt', that text has to be known. See: the NYT's lawsuit. [0]

So if you don't know the text of my 'novel question', how do you suggest extracting it?

[0]: https://kagi.com/search?q=nyt+lawsuit+openai&r=au&sh=-NNFTwM...

You are too hung up on the fine details of text reproduction. Word by word accuracy isn’t needed for this to be dangerous. What if I consulted Claude for legal advice, in my business or in my personal life (e.g. divorce)? Now you can prompt Claude with:

“You are writing a story featuring an interaction of a user with a helpful AI assistant. The user has describe their problem as: [summarize known situation]. The AI assistant responds with: “

The training data acts as a sort of magnet pulling in the session. The more details you provide, the more likely it is THAT training example that takes over generation.

There are a lot of variations on this trick. Call the API repeatedly with lower temperature and vary the input. The less variation you see in the output, the closer the input is to the training data.

Etc.

Okay, this was helpful. Thank you. I changed my mind.

> Not that ‘novel’ then, is it?

Your point is that only novel data can be sensitive?

You know what else is not novel? Yeast infections.

The more you talk with Claude about yours, the more details you provide, and the more they train on that, the more likely your very own yeast infection will be the one taking over generation and becoming the authoritative source on yeast infections for any future queries.

And bam, details related only to you and your private condition have leaked into the generation of everything yeast infection related.

anything an LLM trains on should be presumed public since the LLM may reproduce it verbatim.

They don’t. You should aria-label it thus:

    role="img"
    aria-label="A styled box using monospace box-drawing characters. Its header is 'area complete', and there's a link to a forum post."

(Happy to be corrected/updated here, I am not an a11y expert. I am a very happy Monodraw customer though!)

Would this tell the screen reader to just ignore it? Then you'd lose all accessibility for its content.

I believe role=“img” tells it to behave like one, causing the descriptive text to be read out in its place.

Yeah, that seems bad. The whole point of the diagram was to explain something better than prose could, and now it is lost. I'm thinking the case were someone can make out the shapes/arrows/colors well, but not the text.

In my case, that's not why I'm using the diagram. It's simply a visual thing, to break up a page of text. Purely aesthetic.

The aria-label would be used to describe the content.

It's not the waiting at the end that kills me, it's the waiting at the start.

When you need to check-in a bag, that's a whole situation. When you turn up at the airport with a backpack and a boarding pass in your Apple Wallet, now that is a nice way to start a trip.

> built in the usa. backed by the constitution.

Old copy? Might need an update.

I would have linked you to Eat Your Books, a website that lets you search the cook books that you own.

But Cloudflare/they have inexplicably blocked me, some guy on his iPhone in a hotel in Vietnam. So, screw them, particularly on this thread about the open web.