dijit

FWIW nearly every SaaS or IaaS is making bold claims about supporting data sovereignty, but unless it has been challenged in court I wouldn’t put much stock into it.

The fact that Jira/Confluence are being processed by their LLMs puts them in direct legal conflict with several statutes in the EU, its a matter of time for them to get hit by this and I’m betting that they think the fine will be lower than the revenue they acquire for doing it- and they’re hoping they don’t get hit at all.

This seems to be the common case for a lot of the non-EU tech companies, and it seems to be working.

Oracle is currently the market leader in “sovereign” compute for EU governments for example, as ironic as that is.

You're both making my point.

"Of course that's dangerous" - I'm aware of what I'm doing and the machine was entirely replaceable, but claude had been doing things with an extreme level of competence so I never considered it could do something so dangerously stupid, it had been catching much more subtle gotchas for the last 2 months building my trust.

And you truly believe that it's not even possible.

Well; https://imgur.com/CCD98fO

(Claude Code building against a Debian 13 machine over ssh using the official app on MacOS, btw.)

Let me reiterate the point: "AI can make mistakes" → "This AI seems to make very few mistakes and is catching nuances I might have missed" → "Maybe I trust this AI more than I should- but it's good!" → *AI makes major mistake*

This is as true for code as it is for internet facts.

The issue is that people form opinions based on what LLMs spit out, they think it's "as good as google", and that google was "as good as encyclopedias" and encyclopaedias were "as good as fact"..

Totally forgetting that the accuracy goes down with each of those evolutions.

I'm also reminded of that "You are not immune to propaganda" mantra, in that, the most likely to be swayed by LLMs are those that believe that they understand that they're inaccurate and sometimes dangerous.

Case in point: myself, today, when claude accidentally ran `rm -fr $variableThatDoesntExist/*` and wiped a server of mine.

The variable was one it had set itself (or had assumed).

At some point as these LLMs get better and better we have to take a step back and remind ourselves that: despite the fact that they can be very persuasive and they can earn trust by being mostly correct: they're still capable of great harm (perhaps much greater harm) because they are dangerously wrong sometimes..

.. and worse, they are controlled by corporate interests.

Fair, but he’s the most prominent person here who’s been on both sides of this so I’d genuinely like his take.

Two concrete questions. Did Fly challenge this request, or just comply? And does he still think server-side trust is a non-issue when the identity in this case was built entirely from provider-side records (signup IP, SSH audit logs, a Stripe fingerprint)?

When I said in 2023 that the infrastructure around E2EE matters, he called it misinformation (rather aggressively).

This case looks a lot like the infrastructure mattering.

> Tech companies used to demand warrants/subpoenas for these kinds of things.

Two things are true

1) Tech has become more financial people and "techbros" than principled nerds.

2) The powers of the USG have become a lot stronger and far reaching- quite literally to the level where talking about complying with an order is itself a crime.

The issue that most people bring up is that it optimises for people with free time who can work on these things out of hours.

The same arguments are made against take-home tests.

Personally I'm not sure we can do anything really except try to make ourselves excellent judges of character when interviewing... because people will optimise for the interview if they know what's coming.

Well, if we consider speed of development rather than speed of the runtime, the analogy holds better.

C++ is the "just here to deliver" language.

It can do everything, it's a nightmare to write correctly and is so full of warts that really only a mother could love it.

Go is "elegant" and properly designed, which leads to fundamental unassailable limitations; like its GUI story or writing a backend for other languages (cgo is so slow and -build-mode=c-shared is a hog).

No operator overloading makes it difficult for numerical/scientific work - Python is stronger here.

"just get stuff done" is your crappy dump truck which is welded together out of spare bits and is barely hanging on, this is how I feel about C++.

Go is a lot more elegant and well designed, but it requires "smooth roads" (An OS and Runtime) and goes fast... it's more like a Dodge Hellcat..

EDIT: if we consider speed a function of development time rather than runtime, of course at runtime C++ can be as fast as it gets.

And that's ok.

I don’t know how to say this softly.

Answering in this way has done nothing to further my understanding of the topic, you haven’t explained the difference or why there is a difference and you haven’t offered a counter-example.

I can only surmise that you’ve answered in order to feel superior, but that doesn’t make sense because while I have examples of behaviour that can be explained by my statements, you haven’t added anything to explain yours.

So, your statement stands alone, undefended and without context or explanation.

What were you trying to achieve? A key that is used to decrypt copyrighted material being /itself/ copyrighted means it cant be distributed which is what we see.

BD+ and AACS are literally that though, that’s why VLC can’t play blu-rays.

> NB: Most commercial Blu-Ray are protected by AACS or BD+ technologies and this library is not enough to playback those discs.

https://www.videolan.org/developers/libbluray.html

clicking into the libaacs and libbd+ yields:

> NB: this project doesn't offer any key or certificate that could be used to decode encrypted copyrighted material.

The issue with blu-ray is (and will likely always be) the decryption keys, which are copyrighted and cannot be distributed legally.

Yes they do.

I have a little "dance" I have to do to change my store back to the UK (I'm British) and you need a valid british credit/debit card and address to switch stores to the UK store...

I live in Sweden now and I got a bit bent out of shape when I moved and discovered this was the case.

Btw, buying things doesn't mean they won't remove it: https://blog.dijit.sh/importance-of-self-hosted-backups/

Swedes are excellent at speaking English (better than some British people from deprived areas).. and I work in international companies where English is the lingua franca...

My partner actively prevented it also, she was ethnically Russian (Estonian) and wanted me to learn Russian before Swedish.

Yep, I've been living in Sweden for 10+ years and have (embarrassingly) not learned Swedish to a native level yet.

The more I learn Swedish, the less ignorant I become to the advertising and "gotcha" headlines on newspapers and stickers and such; and I'm noticing its outsized (negative) effect.

1% * 100 is... 100%

(and diversified risk)

a better reason might be that people would not want to use OpenAI because they might be sherlocked.

I used passports.

Everyone agrees that passports aren't made by one company or one country, yet all countries accept passports.

Not everyone can just start making passports though, it's a complicated process of gaining trust and validation as a "real country" before you can issue passports.

One "country" started using a bit of paper that said "trust me bro" and everyone had to treat it as a valid passport. Worse, the way the internet works is that the border guards don't actually see who issued the passport, only that one of the trusted countries produced it.

Having passports this easy to forge means anyone can anyone.

> they know just enough about a topic to fool anyone without much knowledge on the subject

There are humans like this and they are some of the most catastrophic people to have on a project, I actively avoid them.

I'm sure LLMs are different.

Yeah, I chose that example on purpose because of this.

“choice” rarely factors, since I technically could choose to continue consuming it despite disliking it- which is pretty much what the parent is suggesting.

Maybe, the point I'm making is that we focus on thinking we know what the best thing for our children will be ("I can teach them the stuff I learned latest and have the most info about and they'll get a better headstart!") and we overlook things that don't fit into that framing.

I got really lucky that my passion was computers and not something less "productive" like art, or more productive but less economically advantageous like farming.

But you won't go wrong teaching your children how to weigh options, make their own decisions and to be good people with principles and tools for helping them understand how to learn new things themselves.

You will probably go wrong pushing your children to be doctors, lawyers or whatever else you want them to be.

> You simply choose to dislike it.

Too reductive.

I choose to dislike the taste of cilantro, but that choice is wired into many people and doesn’t present as a choice at all: it is instinctual.

and my mum dissuaded me from computers pretty strongly due to the negative stereotypes at the time, and its hindsight that tells us that I got lucky with my passion.

My mum did emphasise my reading though, and that went well.

its a lottery.

That's always been an important skill (at least in the last couple hundred years as far as I can tell).

Industrial revolution, the shift from manufacturing to services, the shift of technology over time.

Hell, they say that the reason millennials are so nostalgic and enjoy childish things is because of rapid technological and societal change in the last 30 years.

I doubt it'll slow down, the era of doing anything until you die was a foolish thing for even our grandparents generation.

My mum was a mechanic, my dad was a mechanic/truck driver.

My grandfather and uncle were mechanics...

I am not, in fact my understanding of mechanics is extremely basic.

If I go back into my family tree, I can find three generations of Sugar House Labourers in Scotland in the early 1800s.. a profession that no longer exists in the UK as far as I can tell, and I'm certain they had a specialisation.

We'll be fine, your children aren't you, they'll find their own passions.

Our jobs as parents is to teach values and make sure they're good people and have the tools to make good decisions too, not to set them on a particular path... that's a losing goal, we can't tell the future for the next year- how could we possibly predict what works well 20 years from now?

Isn't it a gradient? I'd probably rank it like this.

AI Used in back-office administration, internal documentation & presentations

AI Used in research

AI Used for generating placeholders

AI Used for Upscaling (DLSS)

AI Used for translations of UI.

AI Used in the final product (assets incl. textures, 3D, sound and voice)

AI Used in final product (code)

To say nothing of the other disciplines of AI like machine vision or voice recognition or dynamically building a story tree or character response.

Probably? But then I'm using the "vpn slot" on my device all the time which prevents me from connecting to my work VPN when I need it without a little vpn disable dance.

And anyway it drains the battery; and maybe I don't want to add VPN to every device I want to listen to music with, like my PC at work.

Resonates with me.

I'm not here for the money, I'm here for the love of the game.

Changing the game, even if the money increased (it won't) makes the industry worse in my eyes.

Listening to music from plexamp on my phone when I'm out of the house?

Hopefully yes.

iPhone 4s with native resolution textures felt truly magical.

not this flat on flat, grey and text as buttons shite that has pervaded UI for the last decade+

Chassis'