Took this in London the other day.

It's too common a symbol for X.org to have much of a shot because they're not competitors. This "X Social Media" might have a stronger claim if they've actually used and defended a trademark specifically in the social space.

The reality of course, is that the reason they want to go after the IRS is because they don't wan't them to be able to afford to go after the big fish.

A victim of bullying will eventually lash out whether or not they think they have a chance because they become desperate.

Treating all muslims as if they're all the same and interchangeable is pure racism.

ANC bombed civilians and their attacks were celebrated by many. The IRA did, and were celebrated by many. ETA did, and were celebrated by many. It is common, and suggesting it's unique to Palestinians is pure racism.

EDIT: Ah, looked at one of your other comments that were equally awful. Block incoming.

Trying to detect poisoned images is the wrong approach. Include them in the training set and the training process itself will eventually correct for it.

I think if you build more robust features

Diffusion approaches etc. do not involve any conscious "building" of features in the first place. The features are trained by training the net to match images with text features correctly, and then "just" repeatedly predict how to denoise an image to get closer to a match with the text features. If the input includes poisoned images, so what? It's no different than e.g. compression artifacts, or noise.

These tools all try to counter models trained without images using them in the training set with at most fine-tuning, but all they show is that models trained without having seen many images using that particular tool will struggle.

But in reality, the massive problem with this is that we'd expect any such tool that becomes widespread to be self-defeating, in that they become a source for images that will work their way into the models at a sufficient volume that the model will learn them. In doing so they will make the models more robust against noise and artifacts, and so make the job harder for the next generation of these tools.

In other words, these tools basically act like a manual adversarial training source, and in the long run the main benefit coming out of them will be that they'll prod and probe at failure modes of the models and help remove them.

My first "paid" programming project (I was paid in a used 20MB harddrive, which was equivalent to quite a bit of money for me at the time):

Automate a horse-race betting "system" that it was blatantly obvious to me even at the time, at 14 or so, was total bullshit and would just lose him money. I told the guy who hired me as much. He still wanted it, and I figured since I'd warned him it was utter bunk it was his problem.

It's wildly unrealistic but also pointless, because nothing stops us from building new services on top of the existing net. See also: Lemmy, Mastodon etc.

Convincing "regular people" to move is the hard part.

Three things: Scale, recency and contrition or perceived lack thereof.

The British Empire is the largest empire there has ever been. At its greatest extent, in 1920, it covered about 1/4 of the entire world, long after having lost many holdings like the US. The second largest, the Mongol Empire, reached almost the same size, but hundreds of years earlier.

In the same time period as the British, the Russian empire covered <20% in 1895, but its proportion of colonial lands to their own was much smaller than for the British Empire and the proportion of the current world population living in those areas is also much smaller. The French colonial empire covered less than 1/10th of the world at its peak in 1920, and was by far the other largest recent holding of colonies geographically and culturally outside of the immediate sphere of the holding country.

Spain is rarely brought up, I think, in large part because the Spanish empire reached its peak in the early 1800's and so is "history". Belgium doesn't get discussed at much because 98% of their colonial holdings was Leopold II's personal ownership of the Congo Free State. And then we get to the last bit: Contritition.

Nobody goes around saying the massive scale of gross abuse that happened under Leopold II's rule of the Congo Free State was a good thing. Few people I've met ever defend France's atrocities in Vietnam. Even the defence of their ownership of Algeria, which was special enough to trigger an attempted coup against Charles de Gaulle when he wanted to let it have independence because many saw it as part of France itself, is relatively muted.

But there's still mainstream support for the British Empire in the UK. There are still people who insist the British Empire was awesome for the colonies that were exploited because they got English and rails and British legal systems and that somehow outweighs the mass murder and brutal exploitation and erasure of local cultures.

E.g. this survey from 2019, where 32% were proud of the British Empire, 37% were neutral, and only 19% considered it "more something to be ashamed of". 32% were proud of their country's history of colonialism and oppression. Critically this was significantly higher than for other colonial powers other than the Dutch. At the same time 33% thought it left the colonies better off vs. only 17% who thought they were worse off.

I'm not British, but I've lived in the UK for 23 years, and I've experienced this attitude firsthand from even relatively young British people (ok, so all of them have been Tories) - a refusal to accept that the fact that a substantial number of these former colonies had to take up arms to get rid of British rule might perhaps be a little bit of a hint that the colonial rule was resented and wrong.

No other modern empire has left behind such a substantial proportion of the world population living in countries that have either a historical identity tied up to rebelling against British rule, and/or have relatively recently rebelled against British rule, and/or still have substantial reminders, such as Commonwealth membership or the British monarch as their monarch. When a proportion of the British population then keeps insisting this was great, actually, there you have a big part of it.

My own custom text-editor, because it's written to fit into my environment exactly how I want it.

They don't even need to detect them - once they are common enough in training datasets the training process will "just" learn that the noise they introduce are not features relevant to the desired output. If there are enough images like that it might eventually generate images with the same features.

Conflating religion and national belonging like this is pure and vile xenophobia. Thinking all Muslim countries supports Palestine is also staggeringly ignorant.

Having worked at, and co-founded, multiple startups over a period of 28 years: Sure. But why are you choosing that?

The reality is that the moment I started standing up to employers or investors and expecting decent standards, they folded and I was able to have a good work-life balance and get paid market rates and still get to work on cool startups and get shares.

These companies prey on most people never thinking to negotiate (and having been on the other side of the table, and tried to be decent: most people never negotiate, even though we almost always have space to do so)

This is basically the concept of a Webring, and used to be big. Some were fixed (as in the path through the ring was always the same), but some were more flexible or random or semi-random.

A decentralised approach would be new, and not necessarily too hard since the dataset for each ring would be small, so each member could just store all or a subset of the entries in their ring and submit updates to their "neighbours" in the ring that'd eventually spread out to everyone. The challenge is moderation - you'll still end up with some entities that have a privileged position to weed out bad entries, because the appeal was always to a large extent to make discovery "someone else's problem" and the moment you let someone put links on your site someone will try to abuse it.

I actually ditched Emacs because I realised I could write a text-editor that suited me better in fewer lines than my Emacs config took up...

You can see the difference in the process in the results, for example in how some generated pictures will contain something like a signature in the corner

If you were to train human children on an endless series of pictures with signatures in the corner, do you seriously think they'd not emulate signatures in the corner?

If you think that, you haven't seen many children's drawings, because children also often pick up that it's normal to put something in the corner, despite the fact that to children pictures with signatures is a tiny proportion of visual input.

Or how it is at least possible to get the model to output something extremely close to the training data

People also mimic. We often explicitly learn to mimic - e.g. I have my sons art folder right here, full of examples of him being explicitly taught to make direct copies as a means to learn technique.

We just don't have very good memory. This is an argument for a difference in ability to retain and reproduce inputs, not an argument for a difference in methods.

And again, this is a strawman. It doesn't even begin to try to answer the questions I asked, or the one raised by the person you first responded to.

That at least proves that the process is quite different to the process of human learning.

Neither of those really suggests that all (that diffusion is different to humans learn to generalize images is likely true, what you've described does not provide even the start of any evidence of that), but again that is a strawman.

There was no claim they work the same. The question raised was how the way they're trained is different from how a human learns styles.

That was my starting point, and I changed because it wasn't easier.

I switched because my Emacs config was thousands of lines of code to try to wrangle it to do what I wanted. My editor is ~3.5k lines of code and is closer to things how I want them. It's spartan, and you and most other people would hate it. That's fine - I have no interest in writing a general-purpose editor.

Writing a good general-purpose editor is immensely hard, but writing a small editor for yourself is not.

I could absolutely manage to squeeze everything I want into any open-source editor and many proprietary ones via extensions, but there's no value in that to me when I can write less code and get something that's exactly adapted to my workflow.

For starters, I use a tiling window manager, and there are no editors that are designed with that in mind. That doesn't mean they work badly with them, but that e.g. they spent a lot of code on window and tab/frame management that my window manager is already doing the way I want it, and so just by making my editor client-server (a few dozen lines of code with Ruby via DrB), I got that "for free": When I split a view in two, I use the API of my window manager to halve the size of the actual top level window and insert a new editor instance that observes the same buffer. I could retrofit that on other editors too, but doing it from scratch means the "split a view in two" code in my editor is about a dozen lines of code.

Another example is that for my novels, the syntax highlighting dynamically adapts to highlight things I've taken notes about (e.g. characters, locations). I could do that with another editor too, but having full control over the way the rendering layer works meant it was trivial to have my custom workflow control the lexing.

In large part because at least some portion of California lawmakers knows their history well enough to be aware that all of Silicon Valley is a thing in the first place because people were able to leave and take their ideas with them and start something new.

A huge portion of the value of Silicon Valley today can still be traced back to when the "Traitorous Eight" left Shockley Semiconductor to form Fairchild in 1957, and build tech based on what they'd learnt at Shockley, with many of them then going on to leave Fairchild and found further new companies. The outcome of that among many others resulted in both Intel and AMD, and the same pattern has repeated many times.

Human brains clearly work differently than AI, how is this even a question?

It's not all that clear that those differences are qualitatively meaningful, but that is irrelevant to the question they asked, so this is entirely a strawman.

Why does the way AI vs. the brain learn make training AI with art make it different to a person studying art styles? Both learn to generalise features that allows them to reproduce them. Both can do so without copying specific source material.

The term “learning” in machine learning is mainly a metaphor.

How do the way they learn differ from how humans learn? They generalise. They form "world models" of how information relates. They extrapolate.

Also, laws are written with a practical purpose in mind - they are not some universal, purely philosophical construct and never have been.

This is the only uncontroversial part of your answer. The main reason why courts will treat human and AI actions different is simply that they are not human. It will for the foreseeable future have little to do whether the processes are similar enough to how humans do it.

They are a victim of bullying when they've been under decades of illegal occupation. Hamas is an awful organization, but it was only formed as a result of ongoing brutal oppression. When you keep punching someone in the face, sooner or later they'll start punching back, and sometimes they'll fight dirty. That doesn't make them good, but the bully is still the one who kicked things off in the first place and the one who should be first and foremost held responsible for the situation they created.

Hamas individual victims get my full sympathy; they're victims of both Hamas and Israel. Israel as a state does not - without their brutal oppression, extensive war crimes, and apartheid regime, there wouldn't be any Hamas in the first place.

We would not.

The extra amount you need as life expectancy increases diminishes with each extra year. E.g. let's assume (for each of calculation only; you can just scale it up linearly) that you need 10k/year on top of social security to live off in retirement. If your savings is 100k, and you only get a 5% return every year, you'll run out after about 15 years. Hence a typical lifetime annuity bought at age 65 will be around that in the US because it matches up with current US life expectancy (it won't deviate much elsewhere).

So that's for living to roughly 80. Here's how it'll play out as you approach 120:

85: ~20% more 90: ~38% more 95: ~52% more 100: ~62% more 105: ~70% more 110: ~77% more 115: ~82% more 120: ~86% more

As you can see, the curve flattens out. It flattens out because you're getting closer and closer to have sufficient money that the returns can sustain you perpetually (at a 5% return, which is pretty conservative, at $200k, you can perpetually take out $10k, and no further increase in life expectancy will change that).

Now, that of course is not in any way an insignificant increase, but if we assume 40 working years, $100k is about $850/year additional investment + compounding investment return at 5%. $186k is around $1550/year compounding.

But here's the thing, if you work 10 years longer, you grow it disproportionately much, because you delay starting to take money out, and you need less, while you get the compounding investment return of ten more years, and that drives down the yearly savings you need to make back down to around $850/year.

So an increase of 40 years of life expectancy "just" requires 10 more years of work to fully fund it assuming the same payment in during the later years. But here's the thing: Most people have far higher salaries towards the end of their careers, even inflation-adjusted, so most people would be able to fund 40 more years with far less than 10 extra years of work.

(Note that if you already were on track for your pensions to last you to 90, if you were pre-retirement now, you'd "only" need about 35% extra savings to have enough until 120, because you'd get returns from a higher base, so the extra savings or extra years of work needed over what you managed would be even lower)

These all work on averages btw. - due to differences in health, this is where we really want insurance/state pensions rather than relying on individual contributions.

This doesn't mean there aren't problems to deal with. Especially if the life expectancy grows fast enough that it "outpaces" peoples ability to adjust. But it's thankfully not quite as bad as having to add another 30 years of work.

Ignoring the intentionally esoteric languages, of languages in actual use: J, K. Any descendant of APL, basically, and APL itself, though arguably APL is less obtuse than many of its descendants.

E.g, quicksort in J (EDIT: Note Lemmy seems to still garble it despite the code block; see the Wikipedia page on J for the original and other horrifying examples)

    quicksort=: (($:@(<#[), (=#[), $:@(>#[)) ({~ ?@#)) ^: (1<#)

(No, I can not explain it to you)

That's hilarious, given that if these tools become remotely popular the users of the tools will provide enough adversarial data for the training to overcome them all by itself, so there's little reason to anyone with access to A100's to bother trying - they'll either be a minor nuisance used a by a tiny number of people, or be self-defeating.

When they say "can't be blocked" I presume they mean "can't be blocked with the block function in X/Twitter". They also say it can't be liked or retweeted.

So far ads have been treated as sort-of regular posts that are just shown according to the ad rules rather than because they belong in the timeline under normal criteria, and you could like, retweet and block them just like any other post.

So this is basically them treating ads as a fully separate thing rather than just a different post type.

Though the article suggests they'll still try to make them look mostly like posts, except without showing a handle etc. though, which is extra scummy

I remember growing up in the 1980's and 90's when there were still a horrifying amount of people who refused to believe CIA did things like that at all, even in a relatively left-wing country like Norway.

This idea that copyright and IP shouldn’t exist at all is kinda absurd.

For the majority of human existence, that was the default.

Copyright exists as an explicit tradeoff between the rights of the public to be able to do as they please with stuff introduced into the public sphere, and a legal limitation infringing on the publics liberty for a limited time for the purpose of encouraging the creation of more works for the public benefit. It was not introduced as some sort of inherent right, but as a trade between the public and creators to incentivise them.

Stripping it away from existing artists who has come to depend on it without some alternative would be grossly unfair, but there's nothing absurd about wanting to change the bargain over time. After all, that has been done many times, and the copyright we have now is vastly different and far more expansive and lengthy than early copyright protection.

Personally, I'd be in favour of finding alternative means of supporting creators and stripping back copyright as a tradeoff. The vast majority of creators earn next to nothing from their works; only a very tiny minority makes a livable wage of art of any form at all, and of the rest the vast majority of profits take place in a very short period of initial exploitation of a work, so we could allow the vast majority to earn more from their art relatively cheaply, and affect the rest to a relatively limited degree, while benefiting from the reduced restrictions.

I interviewed with them once, and they swore up and down that they were cleaning up and divesting of all the harmful stuff, and wanted me to trust they were all about health and a smoke-free future.

Thankfully they were so staggeringly full of bullshit during the interviews that I quickly realized it'd be an absolutely horrifically toxic (groan, yes, sorry) place to work irrespective of my other doubts, and I ended up telling them I didn't want to continue the process and that I was so unhappy with the assorted bullshit during the process that I didn't want to ever be approached by them again.

That's the very long way of saying I'm not the slightest bit surprised it turns out they are in fact still massive asshats, and I'm very happy I caught on early enough.

Do it. It's fun.

My advice is to start small, and look at some simple examples. E.g. I knew I wanted mine to run in a terminal, and I love Ruby, so I started with Femto which is a really tiny Ruby editor. By itself, it's pretty useless (but beautifully written), but it was remarkably quick to get to something that was "tolerable" for light editing, and then I iterated from there.

There are many options for small ones for all kinds of different values of "small" that can serve as inspiration. E.g. Linus Torvalds has his own branch of MicroEmacs (as do many others, it's a popular starting point, and the basis for e.g. Pico, mg, Vile). Antirez (of Redis fame) has Kilo, so named because it was written to be <1k lines excluding comments, and there's an "instruction booklet" on how to write one that's using Kilo to demonstrate approaches to writing one.

The first starting point, I think is deciding how general you want it to be. E.g. I early on decided I don't care at all about being able to use it as my only editor ever, and that meant I could pick and choose use-cases that were out of scope. For example, I just want to edit "human-scale" files, not multi-GB datasets or log files - I'm happy to open that in Emacs if I ever need it - and so that gave me far more flexibility in terms of data structures because I don't need it to scale beyond a few thousand lines, and that saved me a lot of effort.

To me, that's not an argument for regulating AI, though, because most regulation we can come up with will benefit those with deep enough pockets to buy themselves out of the problem, while solving nothing.

E.g. as I've pointed out in other debates like this, Getty Images has a market cap of <$2bn. OpenAI may have had a valuation in the $90bn range. Google, MS, Adobe all also have shares prices that would trivially allow them to purchase someone like Getty to get ownership of a large training set of photos. Adobe already has rights to a huge selection via their own stock service.

Bertelsmann owns Penguin Random-House and a range ofter publishing subsidiaries. It's market cap is around 15 billion Euro. Also well within price for a large AI contender to buy to be able to insert clauses about AI rights. (You think authors will refuse to accept that? All but the top sellers will generally be unable to afford to turn down a publishing deal, especially if it's sugar-coated enough, but they also sit on a shit-ton of works where the source text is out-of-copyright but they own the right to the translations outright as works-for-hire)

That's before considering simply hiring a bunch of writers and artists to produce data for hire.

So any regulation you put in place to limit the use of copyrighted works only creates a "tax" effectively.

E.g. OpenAI might not be able to copy artist X's images, but they'll be able to hire artist Y on the cheap to churn out art in artist X's style for hire, and then train on that. They might not be able to use author Z's work, but they can hire a bunch of hungry writers (published books sells ca 200 copies on average; the average full time author in the UK earns below minimum wage from their writing) as a content farm.

The net result for most creators will be the same.

Even wonder why Sam Altmann of OpenAI has been lobbying about the dangers of AI? This is why. And its just the start. As soon as these companies have enough capital to buy themselves access for data, regulations preventing training on copyrighted data will be them pulling up the drawbridge and making it cost-prohibitive for people to build open, publicly accessible models in ways that can be legally used.

And in doing so they'll effectively get to charge an "AI tax" on everyone else.

If we're going to protect artists, we'd be far better off finding other ways of compensating them for the effects, not least because it will actually provide them some protection.

I'm mostly shocked they've not lost more and faster. I get more engagement on my ~500 follower Mastodon account than my ~50k follower twitter account these days (sure, I could paid the manchild for more exposure, and if he wasn't so insufferable and had actually made the service better, I might've considered that)

Society is built to distribute wealth, so that everyone can live a decent life.

As a goal, I admire it, but if you intend this as a description of how things are it'd be boundlessly naive.

When you are the only one here who is publicly stating your support for a coup that brought a mass-murdering lunatic into power you do not have the moral high ground on anything at all.

My own. My Emacs config grew over years to several thousand lines, and it got to a point where I decided I could write an editor in fewer lines that it took to configure Emacs how I liked it. It's ... not for everyone. I'm happy with it, because it does exactly only the things I want it to, and nothing else, but it does also mean getting used to quirks you can't be bothered to fix, and not getting to blame someone else when you run into a bug.

That said, writing your own editor is easier than people think, as long as you leverage libraries for whichever things you don't have a pressing need to customize (e.g. mine is written in Ruby, and I use Rouge for syntax highlighting, and I believe Rouge is more lines of code than the editor itself thanks to all the lexers)

You realise just how racist it sounds when you're blaming the Palestinian people as a whole for the actions of a subset?

You see similar issues with French ex-colonies, but since they weren’t as many they don’t appear as much in the news.

Or people aren't as aware of them. E.g. notably their mandates in Syria and Lebanon after World War 1 where they intentionally stirred divisions on the basis of a theory of wanting to keep it so France as a mediator was needed in order to keep them stable. And then they fucked off and left chaos behind.

So you're supporting an actual, proven far-right extremist mass murderer over a democratically elected leader who didn't kill anyone because in your head you imagines crimes he never carried out. That is the typical mindset of a fascist.

Nobody will take the authority of someone who is on record in this very thread supporting a mass murderer.

Thanks. It was a ~100 pound shitty low-end Android phone, so I don't blame you for disliking the camera quality.

The word "nice" in every setting is almost the full extent of Nice's tourism branding.

Labeling media with their significant owners and affiliations of board members would be a great thing. As long as it'd be uniformly applied... And as you've implied, that would certainly be unlikely to happen..