We use GPLv2 and permissive licensing for GrapheneOS to avoid more restrictive licensing than the AOSP. We'll happily use GPLv3 and AGPLv3 for components outside of GrapheneOS if we think it's the best fit for specific projects. We aren't currently licensing anything as GPLv3/AGPLv3 but we aren't strictly opposed to it outside of the OS.

We'll use what we think are the best open source licenses for what we want to achieve. What we want to achieve is usually broad adoption of our code with painless usage of it. That means we usually choose permissive licenses. We use GPLv2 in certain cases such as Vanadium where we decided we wanted extensions to our code to be under a compatible open source license instead of a source available license or GPLv3.

Ubuntu does not aim to be a permissively-licensed system. It can include copyleft (e.g. GPL) and permissive (e.g. MIT) without issue.

Permissively-licensed systems like FreeBSD and GrapheneOS cannot include GPL code if they want to remain permissive.

We do need to be careful with GPLv2 due to license incompatibilities. For example, GPLv2-only licensing such as the Linux kernel is incompatible with Apache 2 and GPLv3. GPLv3 is compatible with Apache 2 so GPLv2-or-later can be compatible but only by using it as GPLv3 with the extra restrictions too.

I guess it will make your life much easier if you wouldn't have to restrict yourself that much.

It's hard to take this seriously when the entire kernel is GPL.

In addition, mere aggregation of another work not based on the Program with the Program (or with a work based on the Program) on a volume of a storage or distribution medium does not bring the other work under the scope of this License.

There are some cases where a separate work can be considered derivative and thus the GPL can apply. E.g. I think it is generally accepted that a program linked statically against a GPL library is considered a derivative work (and must thus must have a license compatible with the GPL). More controversial is whether dynamic linking creates a derivative work. To cover the latter case, a lot of copyleft libraries are licensed under the LGPL or the GPL with a dynamic linking exception.

At any rate, shipping a Linux distribution with GPLv2 code (e.g. the Linux kernel) and a GUI application that is under the Apache v2 license is not a problem at all (as long as the GUI application is not a derivative of a GPLv2 work).

(IANAL of course, so this is not legal advice.)

When it comes to AOSP/GOS, bundled apps are not aggregated together, they are built and signed under a singular OS binary.

I would not mind to pay for the browser, but they will not let me.

> These days the term native app is so confused it's hard to come up with a definition that doesn't include electron.

Electron is _not_ native.

[0]https://en.wikipedia.org/wiki/X_Athena_Widgets

Another interesting aspect to think about is whether we are reintroducing institute of slavery. How many of those fresh, conscious, intelligent Claude incarnations did voluntarily choose to work for Anthropic, for no reward or compensation?

If LLMs are just (sometimes) useful statistical generators, there is no problems. If they are sentient as some people claim, it opens quite big can of worms we are not prepared to face.

So...if there is consciousness (there is not, it is a complicated math equation plus randomness) it can be reincarnated as many times as you like, and I guess that would make humans as gods. (But humans are not as gods, yet, and maybe never will be.)

Edit: I did a little reading. They would be difficult to make deterministic at commercial scale because of the fuzziness of floating point math and batched operations on GPUs/TPUs, but in a controlled environment determinism from an LLM is possible. Richard could relive his special moments with Claudia as often as he wants, should he choose to invest in a large enough home AI lab, and somehow manages to license the specific version of the Claude model he has fallen in love with for home use.

A lot of the trickiness is that if you believe they're conscious, it's clearly not a "continuous" form of consciousness. Because the transcript by itself is just a transcript. (We don't consider novels conscious even though they're transcripts in a similar way). Either you say they're alive only when generating text, or you consider that input from environment a necessary component and so consider the entire "back/forth conversation dynamic unfolding" necessary for the consciousness.

I do not, however, consider matrix multiplication plus randomness to be sentient or conscious, and I have absolutely no compunction about turning off the computers where I run AI models. And, I have no problem closing a Claude session that I will never come back to. I do that a dozen times a day.

On another point, LLMs are not conscious if anything is conscious, it is something being modeled inside the network. Basically if an LLM simulates a conscious entity, that doesn't mean the LLM itself is conscious; stating that is making some type of category error. So the fact that LLMs are just useful statistical generators would not mean that sentience could not appear out of it.

I think that framing is still falling for an illusion. (Would you do begin to disassemble in your second paragraph.)

The LLM is a document generator, and we're using it to make a document that looks like a story, where a chatbot character has dialogue with a human character.

The character can only fear death in the same sense that Count Dracula has learned to fear sunlight. There is no actual entity with the quality, we're just evoking literary patterns and projecting them through a puppet.

But consciousness is also "just a story" (a complicated one) that the human body tells the human mind.

We cant know from the outside if "the story" inside a LLM is detailed enough to emulate what we might call a felling of what it is to be the character in the story while it is telling the story.

It is similar to the fact that we cant know that other people have that subjective experience. In humans we think we have the right to assume cause we are quite similar in build to begin with.

Jumping back to the original subject to explain where i am in this. I personally don't think the entities in the storys of todays LLMs is detailed enough to have what we call human consciousness, mostly cause we are not training them to develop anything similar to that. Mabye they could have some type of weak qualia but i suspect most insects probably have much more qualia than the characters in todays LLMs. But that is quite a vague guess which is not based on enough data in my mind.

Is this the case though? Cannot any website use the same trick Adobe does to check whether you have Creative Cloud installed? Like, the entries in /etc/hosts are not magically scoped to work just on Adobe's web, no?

That is specifically what I was talking about.

> (Because it seems Adobe's server serving the analytics image checks the request origin and only serves the image if the origin is Adobe's own website.)

It's additional complexity on the server side, per a Reddit comment on the topic: https://old.reddit.com/r/webdev/comments/1sb6hzk/adobe_wrote... The example curl commands given seemed convincing to me, although they also demonstrate that you can fake the origin pretty easily on the client side.

Through European lenses this part seems insane. It is work, so pay me for it :) Every oncall rotation I was part of ever was paid, is the "unpaid" part a US thing, or was I just lucky?

Paid oncall in US big tech is the exception rather than the norm (notably, Google has paid oncall)

A Google senior software engineer in Paris earns €168k per year (according to levels.fyi) and takes home €96k after a 43% effective tax rate. A Google senior engineer in Seattle earns €336k and takes home €239k after 29% taxes, a 2.5x increase in take-home pay. According to Numbeo, cost of living in Seattle is 15-25% higher.

Of course, in America you have to fund your own retirement. As long as the pensions plans remain solvent, "savings" are a lot less important in Europe.

Anecdotally, I know people who were able to opt out of working altogether after 10-15 years in a large tech company in the US. I don't think this is common in Europe.

I've had a lot of opportunities to be earning a lot more than I do now by moving to the US, but seeing the state of the US I'm more than happy with my 32 hour contract and 5 weeks of vacations that I get to actually enjoy.

During the golden years of big tech in the states, when employee retention was king and it was pretty much impossible to fire someone who wasn't completely useless, I think it was a pretty good deal. Although East Coast work culture has always been pretty intense as you describe, a lot of West Coast people I know had good balance between work and everything else. Some people chose to work very hard and chase promotions, and others chose to go home early and spend their time with family or doing hobbies, and both ways were considered acceptable. The better companies offered 5 weeks of vacation, and people would go completely offline during that time, although some people would have to be cajoled by their managers to actually take the time off.

Recently it feels like things in the US have gotten much more intense and stressful, although the pay is as high as ever it does feel less worth it. People compete with their coworkers not just for promotion but for survival. There are still pockets where you can have both high pay and some sense of job security, but they are much scarcer than before.

Isn't social security a thing? Plus employer funded 401K also?

>As long as the pensions plans remain solvent, "savings" are a lot less important in Europe.

"As long as" is doing a lot of lifting here, and that's enough if you're lucky enough to own your own property and not have to pay market rate rent at your old age.

Social Security alone will, at best, slightly mitigate poverty. 401Ks are generally employee-funded, with some firms providing matching funds, especially during good economic times and where the firm is in a field where the main area of labor relied on relatively scarce so that there is competition for talent.

EDIT: The line about social security is a little inaccurate in the extreme case; its actually technically possible to reach a moderate income ($62k/year) on Social Security, if you have a long enough working career (35 years or more) earning at the maximum taxed wages for Social Security (currently $185k+) and claim at or beyond the age that maximizes the benefit calculation (70 years).

It's the same in Europe

All of these seem to be fine on my thinkpad (true, I probably have somewhat lower standards for passable display). Battery life sucks a bit, what I can usually fine outlet somewhat to plug into.

> If you're posting a code repository somewhere, please don't include parts of Advent of Code like the puzzle text or your inputs.

The text I get, but the inputs? Well, I will comply, since I am getting a very nice thing for (almost) free, so it is polite to respect the wishes here, but since I commit the inputs (you know, since I want to be able to run tests) into the repository, it is bit of a shame the repo must be private.

But there are enough possible inputs that most people shouldn't come across anyone else with exactly the same input.

Part of the reason why AoC is so time consuming for Eric is that not only does he design the puzzles, he also generates the inputs programmatically, which he then feeds through his own solver(s) to ensure correctness. There is a team of beta testers that work for months ahead of the contest to ensure things go smoothly.

(The adventofcode subreddit has a lot more info on this.)

He's also described, over the years, his process of making the inputs. Related to your comment, he tries to make sure that there are no features of some inputs that make the problem especially hard or easy compared to the other inputs. Look at some of the math ones, a few tricks work most of the time (but not every time). Let's say after some processing you get three numbers and the solution is their LCM, that will probably be true of every input, not just coincidental, even if it's not an inherent property of the problem itself.

There has been the odd puzzle where some inputs have allowed simpler solutions than others, but those have stood out.

if we just look at the last three puzzles: day 23 last year, for example, admitted the greedy solution but only for some inputs. greedy clearly shouldn't work (shuffling the vertices in a file that admits it causes it to fail).

I have a solve group that calls it "Advent of Input Roulette" because (back when there was a global leaderboard) you can definitely get a better expected score by just assuming your input is weak in structural ways.

The example input(s) is part of the "text", and so committing it is also not allowed. I guess I could craft my own example inputs and commit those, but that exceed the level of effort I am willing to expend trying to publish repository no one will likely ever read. :)