I think a better approach might be to require that any algorithm used to suggest content to users must be made open source so that people whose world views are being shaped by the content you're feeding them can analyse how you're deciding what to show them.

I feel like there's definitely a problem here with social media and its effect on society, but our first approach should be to increase transparency and accountability, rather than to start banning things by force of law.

Knockout js worked something like that. There are complexities with the observable approach though, especially around batched updates and avoiding update loops.

Not to mention debugging Knockout.js applications was as complex as debugging jQuery applications, mutable state that can be changed from anywhere and from any direction makes it really hard to see what's going on. React was a breadth of fresh air when it appeared.

I assume that this app is not a mailing app, its an app to control your fastmail service, which includes email, file storage, webhosting, various dns controls, notes, contacts, billing etc. They already contributed JMAP. It would be impossible for them to get all of those features into k-mail - k-mail wouldn't want them.

And packaging the app in electron is relatively easy and a nice convenience for some users.

Honestly I just want an API endpoint to upload a new set of Sieve rules.

My aphantasia mainly feels like a super power, that I process at the conceptual level rather than the visual level. I feel like it helps me focus on the important things and ignore the extraneous. It's strange to read it described as a 'deficit' and as something needing to be corrected.

I came to say the same. For the longest time I never realized that I was missing anything. Most of my dreams don't feature color unless its relevant to the story. This is the superpower, no distractions. It allows keeping a large amount of context of related things, even without physical form, abstract things, while working through them. It also explains why I discovered early on that I can't stand descriptive prose that paints with words for pages to set the scene before getting on with why any of it is relevant.

Reminds me of the quote "Mathematics is the art of giving the same name to different things." - Henri Poincaré. [As opposed to Poetry is the art of giving different names to the same thing]. If different things can be the same, then their physical form wasn't relevant and you can consider the abstract thing singular.

Although I don't feel like I'm missing out, I read that Nikola Tesla was able to work out how to construct the first A/C motor by constructing it in his mind's eye. That's pretty cool, I'd never have to draw any system diagram again except to transcribe for communication.

I think that often theres a difference in ability to put capital to work between the US and europe. Lots of innovative work has happened in Europe, but it often gets bought by US or Asian capital, or gets outcompeted by vastly more resourced US competitors, or just withers on the vine from a lack of investment.

Sabre seemed a really interesting space approach that failed from lack of funding.

> I think that often theres a difference in ability to put capital to work between the US and europe.

It’s because the EU is 27 different countries with different regulations while the US is 1. Some work is being done to fix this but remains to be seen if we can reach a point where we have unified capital markets instead of national ones.

That does not prevent something like this project happening which includes 23 different European countries, and not even through the EU.

The point of the EU is to provide unified markets, and capital flows pretty freely between major European economies (including those outside the EU) so I do not think that is it.

Meanwhile three of those EU member states can't be manage to cooperate on their 6th-generation fighter programs.

https://www.politico.eu/article/germany-france-europe-fighte...

That is something that remains outside the EU's remit because it is military and decision lie in the hands of member states.

I was rebutting a claim that differences in regulations preventing capital flows. That is not true in general in our globalised world (capital clearly flows freely between many countries) and certainly not true in the EU.

It is possible that the EU loses somewhat from not having the central budget that the US federal government has which which to back big projects, but, in most cases, the bigger EU states are big enough to back most things (as is the UK) but will not do so. The lack of fiscal centralisation in the EU also affects its financial stability but I do not think that is directly related to this problem.

As for sixth general fighters, a number of smaller economies than Germany alone have or plan their own sixth generation fighters, or have done so (some programmes have merged). Again, a difference in attitude and priorities.

I think it is a difference of attitude. There is a lack of belief in big projects, particularly from the government - especially in the UK and Sabre is British.

SpaceX was started by someone with a few million dollars and self-funded at the start.

Sabre was always a dead end.

It took 100 million to get to falcon 1, and the company was probably not viable until NASA gave space X the crs1 contract for >1billion...

And where did that initial money come from? Musks sale of PayPal shares, which he got when PayPal acquired his payments company... Would that have happened if he'd started his payments company in Europe? Would Compaq or an equivalent have been interested in paying for zip2, which provided the initial money to found x.com if he'd built zip2 in Europe?

Musk deliberately and intentionally moved to silicon valley, and I think a big part of the reason was almost certainly that the US is more prepared to invest money in these kinds of ventures.

Are you saying this with an awareness that both DeepMind and ARM were founded in the UK?

I am convinced that a lot of people have developed a completely distorted view of the UK by spending too much time on x.com where Elon Musk pushes his “come on, have a civil war already” bullshit and paints a picture of a country in its death throes.

Do we know if this violence is politically motivated yet? (Other common motivations are mental health issues, paranoia, revenge, desire for fame etc). Of course it seems likely, but it also seems premature to jump to trying to use this as proof of a particular personal position.

I definitely believe that people should be more understanding of each other, and less quick to jump to insults and othering, but we know so little about this situation, to be so confident that it was caused by speech seems extreme.

I am also aware that a lot of the political violence of the last few years ended up not being motivated by the reasons one might naturally expect.

> Do we know if this violence is politically motivated yet?

How many long-range rifle shot assassinations do you know of that were not politically motivated? Jilted lovers and such don't do that. In context it's hard to take this assassination as anything other than politically motivated.

I guess that largely depends on how one qualifies "politically motivated". By some definitions it's easy to include any of what you listed as also part of a politically motivated attack, by a narrower definition one could just as easily choose to exclude them. E.g. whether an attacker is paranoid is orthogonal to whether the attack involved the victim's political views/activity in some way.

At the root I agree in principal though. It's, for example, still possible he picked a bad fight with an unstable individual in a bar last night (over something not politically related) and they followed him to the event he was speaking at to shoot him. I'm not as convinced I've seen that kind of thing happen "a lot", but it's true we don't have post validation yet.

I mean... it could have been a jilted ex-lover.

my basic guess would be that its epstein related, which is still politically motivated in some sense, but "killed him for protecting pedophiles" is quite different from "killed him for being right wing"

Isn't it more likely that this is a false flag operation designed to distract from the Epstein birthday card signed by Trump? The timing is suspicious and there's certainly a lot of bandwidth given over to a single shooting, compared to the school shooting on the same day (three shot).

Kirk would seem like an ideal target as he has a high media profile and is not involved in running the government. I would guess that the aim is to promote civil war and thus provide an excuse for martial law.

You start your comment saying we should avoid making apocalyptic statements and end it by saying "the cycle is going to destroy our society".

My conclusion is that you don't mind making apocalyptic statements about actions you think are dangerous to society, which sits uncomfortably with your asking other people not to.

I'd say the appropriate read there is to slip the word "unjustified" into a few key slots. The view is nearly impossible to avoid in context. How do you see society surviving if the prevailing view is that anyone with a different belief is trying to bring on the end times? To the point where assassinating political opponents is justified?

It would bring on the end of a society. It might well happen in the US case, they've been heading in a pretty dangerous direction rhetorically. If we take the Soviet Union as a benchmark they probably have a long way to go but that sort of journey seems unnecessary and stupid.

> I'd say the appropriate read there is to slip the word "unjustified" into a few key slots.

"You shouldn't do anything unjustified" is an uncontroversial and useless prescription.

I think they're politely asking for the far left to stop with the language inflation. Use words with appropriate and proportionate meanings. Do not try to gradually be more and more dramatic and impactful.

> My conclusion is that you don't mind making apocalyptic statements about actions you think are dangerous to society, which sits uncomfortably with your asking other people not to.

This is a nonsense argument. It is possible that constantly making apocalyptic statements can result in an apocalypse, and saying that people should stop doing that is not contradictory.

The words you use matter. If trump is an existential threat to democracy, he should be assassinated. If you're not advocating for murderous escalation, then stop using those words (for example).

> If trump is an existential threat to democracy, he should be assassinated.

Who/what is defining assassination as a reasonable response to that threat, who/what maintains the list of words which can replace "democracy" in that section, and what happens when someone disagrees with the maintainer of that list?

Those are all great questions, and why the point under discussion is whether or not we should choose our words more carefully and stop making apocalyptic predictions.

I wholeheartedly disagree - we need to be less concerned with who might say something and more concerned with how we teach society to react to it. Whether or not someone is making apocalyptic predictions should not define our ability to hold back from assassinating.

How we say things is how we teach society to react to it. We're all teaching each other, every day.

I'd agree there are aspects of how we say things which can reinforce how to react about it, but I don't think that's a good primary way to teach how to engage with polarizing content and certainly not via the way of avoidance of the types of statements bigstrat2003 laid out. I.e. there are very reasonable, particularly historical, examples of belief of potential threats to democracy which turned out to be true, so I don't inherently have a problem with that kind of discussion. I actually think calling that kind of statement as the problem would actually drive more extremism.

At the same time, I do believe there are ways to share such statements while also reinforcing healthy ways to react at the same time. kryogen1c's example ending in "he should be assassinated" crosses the line from bigstrat2003's talk of apocalyptic claims to direct calls to violence about them - the latter of which I agree is bad teaching (but I'd still rather people be encouraged to openly talk about those kinds of statements too, rather than be directly pressured to internalize or echo chamber them).

This is why the first question posed about the statement from kryogen1c was "Who/what is defining assassination as a reasonable response to that threat". The follow on questions were only added to help highlight there is no reasonable answer to that question because it's the call to assassination which is inherently problematic, not the claim someone is a threat to the democracy here. The latter (talking about perceived threats) is good, if not best, to talk about directly and openly. It's the former (calling for assassination about it) which is inherently incompatible with a stable society.

Humans are not rational machines.

You can "educate" someone all you want, they will still suffer from all the normal biases and those biases will still affect their choices.

This is why we have double blind trials even though doctors are "experts"

I agree with this, and, as a result, I don't believe there is any possible approach which results in 0 people assassinating political figures for what other people say. I think the same conclusion can even be reached if people were supposed to be expected to be perfectly rational beings.

I do believe education on how to effectively engage against an idea which feels threatening is better equipped to handle this apparent fact than bigstrat2003's approach of teaching people to not say certain beliefs because they'd be worth killing about. That doesn't mean it results in a perfect world though. Some may perhaps even agree with both approaches at the same time, but I think the implication from teaching the silencing of certain beliefs from being said for fear they are worth assassinating over if believed true ends up driving the very problem it sets out against. Especially once you add in malicious actors (internal or external).

> stop making apocalyptic predictions.

You’re one to talk.

For fun over the last few days, I've built a compressor / decompressor that uses the logits from an LLM, for each token in the input, then takes the ranks and exponential goolomb encodes them. Then you work in reverse to regenerate the original

It took me ages to get the prediction for the second token after "hello" to match the same as the prediction for the second token when running the model on the string "hello world", despite the fact that I was using a causal model. I tried all kinds of things before discovering that `quantized: false` was the important setting.

What's the Weissman score? Or more seriously :) did it perform well. Sounds like it should. If more and more text is AI slop it should do well.

I dont fully understand what you said but I guess higher probability logits are encoded with fewer bits. If your text is the LLM output then you may need a bit or two per token?

I used exponential golomb coding, so the rank 0 logit is encoded with a single bit, ranks 1 and 2 are encoded with three bits, ranks 3-6 are encoded with 5 bits, etc.

In terms of performance, I've not done any serious testing, but e.g. the wikipedia article on volcanos compresses to about 20% using GPT2. I've seen other strings compress even further.

The big issue is that while encoding is not unreasonable, decoding any significant amount of data is incredibly slow, since I'm doing a model run for every token in the output. It's bad enough that the scheme is probably unworkable as it is. I'm thinking about changing my code so that it streams out the tokens as it decodes them, so you're not just left there waiting for ages.

I don't know about golomb coding, but with Arithmetic coding you can do stream decoding(AC), if I remember correctly.

I supervised a student's project whose goal was exactly that : implement compression with LLMs using AC.

Since AC is optimal, if your LLM has an average cross entropy x on some dataset, you can expect that the compression will compress data using x nats per token on average!

Arithmetic coding looks like an extremely interesting approach, given that you can use the model at each step to give you the probabilities of each token.

I've become convinced that the real problem is probably impossible to get away from.

Ultimately we want a nice set of reusable UI components that can be used in many different situations. We also want a nice set of business logic components that don't have any kind of coupling with the way they get represented.

In order to make this work, we're going to need some code to connect the two, whether it's a 'controller' or a 'view model' or some other piece of code or architecture.

However we choose to achieve this task, it's going to feel ugly. It's necessarily the interface between two entirely different worlds, and it's going to have to delve into the specifics of the different sides. It's not going to be the nice clean reusable code that developers like to write. It's going to be ugly plumbing, coupled code that we are trying to sweep into one part of the codebase so that we can keep the rest of it beautiful.

These seemingly inescapable tradeoffs are almost always actually quite escapable if you look at it from a different perspective. You have to stop thinking about pages and button-clicks, and you have to stop using frameworks that try to do everything, box you in, and force you to architect your state-flow logic based on your visual hierarchy. This is the biggest problem with almost all UI frameworks: all your logic has to be partitioned along the lines that are set up by how your screen looks, or you're swimming upstream to prevent it. Instead, have your domain models declare how they work and interact using intermediate services, and consume those services to generate the UI as a consequence of those declarations. It's very hard, and I don't have all the answers yet, but I've tasted enough to know that it genuinely avoids this otherwise seemingly inescapable tradeoff. I'm not planning to ever build another UI (above some complexity) differently again.

Speaking of a different perspective, you and GP are describing it with different viewpoints and I'm not sure you're fully aware it is different:

GP is describing the Model and View as two co-equal things with something in between that links then.

You're describing the Model more as a foundation that the View builds on top of (and this is also how I look at it).

I like the stacked view more because it more closely matches the flow of data, making it almost explicit that what's presented to the user is a projection/transformation of the stored data. Views can't really exist independent of that data, they rely on it, while models can exist independent of the view.

Yeah exactly, you can come up with all sorts of abstraction layers but your ugly code has to go somewhere eventually

I'm not saying it's not simple to write, I'm saying it's ugly and contingent and you can't really avoid that. It's exactly this reason that has led to a proliferation of MV* patterns, including the one you describe.

But to try to explain myself more clearly - in the architecture you describe, who is that it is implementing TableData and TableCallback? Is it your beautiful clean business logic classes that have no coupling to their representation - in which case that is weirdly coupled in an ugly way, or is it some other class that acts as the bridge between the two worlds, in which case, that's where your ugly code is living.

Ideally a bridge (e.g. a typeclass) with nice language support like Scala has. I'm not seeing what's so ugly. In math, you have an abstract interface for e.g. monoids (your interface like a Table). Then you have e.g. complex numbers as a set (your business model). The you can identify how C is a monoid via addition. And how it's also separately a monoid via multiplication. Same idea. There's nothing "ugly" about having to write down what you mean.