jonathanstrange

Oh I agree, I didn't want to insinuate that war crimes on either side shouldn't be prosecuted, and the question is actually open whether using AI for targeting does not in itself constitute a war crime because it's negligent. When Israel first declared they used AI targeting for their fairly brutal (but also ineffective) initial bombing campaign, I was expecting a larger outcry.

I just wanted to challenge the one-sided focus on one incident by Ukraine in it's own occupied territory when in fact it's well-known that Russia has habitually committed war crimes in Ukraine right from the start in 2014. I mean, they shot down a civilian airliner killing 298 people and FSB officer Igor Strelkin bragged about it on VKontakte. And that was just the start of a long series of heinous crimes and mass killings. It's also worth noting that the ICC is officially investigating war crimes in Ukraine committed by either side but Russia does not allow them to investigate.

Unfortunately, it was clear with the success of drone warfare that automated killing by AI was going to be the future and it should have been banned internationally right from the start. They are already using autonomous killer robots and drones with integrated AI targeting in Ukraine on both sides.

They literally had fiber optics connections to the NSA and claimed it was without their knowledge. So No, they definitely can't keep data safe.

How the software is written isn't that important because the endpoints aren't secure. Their hardware and firmware isn't even secure enough.

This has nothing to do with mental gymnastics, both Russia and now the US have started wars of aggression, and you're seeing the results of those crimes, which are by their very definition worse than any war crime. From a moral point of view, every death that occurs as a consequence of these aggressions is the fault of the governments who started them and indirectly also on those people who enable these governments.

If you can't see that, your moral compass is completely broken.

It's definitely true. I didn't use AI at all until this January because the code it wrote was too bad. Now I'm comfortably using AI for less important programming tasks and it's okay as long as you steer it carefully. I see no reason why it wouldn't get better up to a point that AI written code is better than what any human programmer could produce. It's also constantly getting better at refactoring.

For me the only thing that matters is the content the quality of the writing. If you're primarily interested in something else than content and quality, there is no point in reading at all. So I don't understand these anti-AI posts. If all they say is that most AI written texts are too long and badly written, I agree. However, that's also the case for many texts written by humans.

I can't speak for OP but there can be many reasons. For example, I'm using a Macbook Air for development purposes, it's essentially dumb compilation/notarization machine. I don't use it for anything else because I find the user interface infuriating and can't get around it (which is weird because I used to be a Mac developer but IMHO the UI has gotten much less practical).

Is there a way to test this in a browser interface? I'm not going to run any executables from any of the frontier labs on my machine, be that Xiaomi or Google or Anthropic.

I guess you can do anything you like if you don't care when people don't take you seriously.

Your arguments aren't convincing. LLMs do not have authorship, they aren't persons. These ramblings about plagiarism are irrelevant, they're a classic case of "topic missed."

The same for human editors, about whom I have slightly more nuanced view. I do believe that some of them are offering services that amount to ghostwriting and are often unethical. I've seen people make careers in the humanities where I'd say the paid editor (who also had studied the subject area) would at least be a co-author, if not the person who in the end deserved most of the credit. I've seen that with my own eyes as someone who worked for 20 years as a researcher in academia. That's why I've always rejected private editors that many of my colleagues have employed. It's a deeply unethical practice, not at all as harmless as you describe it.

That being said, the comparison to editors also doesn't apply because, again, LLMs aren't persons and do not have authorship. They neither are persons and authors in a legal sense, nor does it makes much sense from an ethical perspective to grant them personhood and authorship. If that made sense, then I'd agree with you that they should be mentioned as co-authors or authors. I'd have no problem whatsoever with academic and other works written by LLMs and giving credit to them. However, I don't see we're there yet.

Regarding the quality issue: What I'm saying and what my stance is is exactly that the works of persons or LLMs or persons who use LLMs for writing code or texts must and should always be evaluated the same, namely based on the merits of the content alone. That would be even more important if LLMs were persons and could be attributed authorship -- but they can't.

You arguments don't work because they're based on incorrect assumptions about what LLMs are. They are tools. You might find it dishonest when people use them without stating that, and that's fine, but then you shouldn't base your views on unrelated issues like plagiarism, copyright, and editors. My hunch is that people like you simply can't cope with the change, and, again, I'm saying that as someone who barely uses AI for programming and has never and will never use it for writing. Please stop the whining. I'll be honest, to me personally, these anti-AI rants are intellectually insulting because they seem to imply that humans cannot stand against AI by the quality of their own works and require some extra protectionism.

I've written a lot in my life so far, including books and novels, and never used AI for it. I don't see the parallel. If you were right and it was the same as plagiarizing, then the use of AI should be prohibited entirely. However, to me it appears more akin to using tools like spell checkers, CAS pocket calculators, and Grammarly. It's another tool. I don't really get the fuzz people make about AI. I'm also fond of judging the content of texts regardless of the author. Has that idea come so much out of fashion?

That attitude unfortunately leads to people criticizing something without having enough evidence about it in the first place, like criticizing a book and later admitting you've only seen the cover and read the summary on the back.

Criticisms should have substance.

It's not important at all and I wish people would stop mentioning AI. If you have criticisms on the docs or code, by all means make them but they shouldn't concern "AI vs human" considerations. Humans have written plenty of bad and documentations, too, it really doesn't matter who wrote what to what extent, what matters is whether it's the right quality.

If it's not worth publishing, it's not a result at all. That being said, the problem really is that since LLMs excel at sifting through piles or papers, publishing more would aggravate the problem instead of alleviating it.

Perhaps that's good, though. The market for truly minimal phones seems to be rather small. I've had a chance to buy a new Commodore phone and they are really cool, but in the end I decided against it because I wasn't even sure my banking apps were going to run on it and I do need some social media on my phone even if I don't use them often. For example, everybody uses Whatsapp where I live. If you need to buy another phone to complement your minimal phone, it's kind of pointless.

The UK left and wasn't even able to make proper deals for exiting the EU, their own Brexit caught them by surprise and it was total chaos for a long time. They could now try to get closer integration with the EU but there doesn't seem to be much will for it in the UK.

Governments don't necessarily have the connection between your ID and what you do online, though, and some governments are known to massively buy publicly available data from data brokers to circumvent existing laws. By "some governments" I mean the US government, by the way. That's not a conspiracy either, it's well-documented.

The technology was first released into the public 4 years ago. Now it's proving major theorems in research mathematics. It doesn't take a genius to realize that this might indicate an extinction-level shift over the next few hundred years.

Academic work is based on worldwide sharing, the sharing is not the problem, it's the lack of attribution. Unsurprisingly, these companies neglect standards of academic honor and attribution. Some human researchers also used to do that but in a discipline like mathematics this used to be a small problem because people tend to be so specialized that very few people could just grab someone's research and quickly piggyback on it, and if they do, colleagues will generally understand what happened. Unfortunately, AI is changing this.

I don't trust your judgment, it's in my opinion even hilarious given that we're talking about companies worth almost a trillion dollar (4 trillion in the case of Google). Be that as it may, it was nice chatting with you!

It's trivial and a solved issue for the companies developing frontier AI models. Obviously, you don't even need AI for searching every prompt every user has ever written to find interesting topics, but you can create automated summaries and use AI on them if you want. There is no "scaling issue" here for companies who are used to processing almost everything that has ever been written anyway.

I didn't want to insinuate that it's trivial for small companies or individuals to do big data mining at that scale, sorry if I made that impression.

He was arguing against two years. A lifetime is fine.

I have the setting turned on in Gemini Pro even though I work on proprietary code because 1. the setting allows for some (very limited) "memory", and 2. I consider my source code almost public even when it's not open source because I don't work on programs that involve extremely high level of know how or proprietary algorithms. It's mostly CRUD that can be copied in a myriad of ways, whether people use my methods or other methods.

If Gemini can improve based on my code and sessions (maybe doubtful but who knows) and others can benefit from it, that would be a welcome side-effect.

It seems completely trivial to feed sessions to their own LLM and ask it to look for various things in them, from detecting problematic use cases to finding interesting mathematical work.

That sounds a lot to me like a lie.

Based on publicly available evidence, we should believe the former because AI models have made continuous advances that have been measured. The idea that the continuous advances will stop currently has no evidential support. I'm not saying it's not a possibility, just that it seems unsupported conjecture right now.

By the way, there seems to be a new form of AI skepticism emerging in the US that comes from general opposition to data centers, and in my experience the AI skeptic part of it is wholly irrational. I've met people online who suggested, without providing any evidence, that AI is useless and nobody wants it. That's a very implausible take.

Who doesn't fear that...

According to the information I have, the CIA is unlikely to be involved with SIGINT of that type. It's just not their role. I agree that most of the information the NSA might collect will come from publicly available sources like data brokers, particularly if US citizens are involved. However, what I was talking about concerns real-time capabilities and predictive power, it's very different from targeted surveillance and anything involving courts.

If I was the director of an agency of the size of the NSA and was evaluating the options purely from that perspective, I'd aim at creating a file on every living citizen on earth, including their social network topology and their activities. Basically a Google search engine that includes information not publicly accessible. I'd create much larger files for persons of interest and authorize targeted surveillance of them, of course, but with today's means to collect data a complete world database on every living and many dead persons is well within the technical capabilities. It also makes sense and is rational, if you put aside moral considerations.

That's how I evaluate these things. If it makes sense and can be useful, it's likely going to be done. Notice that there is no law against this in the US if you exclude US citizens. It's perfectly legal and within their mission parameters to do it for non-US citizens. I used to think my judgments were a bit too much on the paranoid side but when Snowden published his leaks it turned out that I was roughly right about every capability the NSA had except for their internal security.

The author might not be aware of it but there have been extensive discussions and work about such terms in the philosophy of language in the past two decades (and earlier). I've heard whole lectures about whether fruit salad really is a salad, and the sandwich examples, too.

In my opinion, geometric meaning theories capture the associated problems fairly well, for example Conceptual Spaces. In linguistics, prototype theory and related approaches have dealt with them since the 1970s. That is not to say that the problems are solved but in my opinion many problems come from an artificial focus on logical concept decompositions and the fact that language doesn't work that way.

My 2 cents as a former philosopher who worked in this area.