saurik

I think part of the problem -- answering all of why people are somehow OK with this, and even why the AI does this in the first place -- is that the majority of software developers never got to the point of understanding any of this: the lack of understanding how to carefully structure data surrounds a question they don't even know how to pose, much less answer, and so "write some tests then incrementally try to make them work without breaking any of the existing tests" is the only way they know how to develop at all. In the end, that means that, with the current state of the art (which might change, of course... potentially quickly), your strategy of treating the AI as a junior engineer who fundamentally isn't ready to do your senior-level architecture job makes a lot of sense.

Reducing access to a weapon/tool behind any variety of roadblocks because you think the weapon/tool is dangerous is not the same as shifting blame off of the person who did the crime. Saying we don't want people to be able to buy guns without limitations or even saying people should not be allowed to make or sell certain kinds of weapons/tools is not the same thing as saying the person who jumped through the hoops (or ducked underneath them) is not liable for the crime they committed using the weapon/tool. It is certainly the case that no one, at the end of the day, claims the gun is to blame for the murder a human carries out with it... some people will claim maybe the humans making or selling the gun share some blame, but it would be a fundamental attribution error to blame the gun itself.

BTW, as kind of a follow-up to this, I think the most important finding to report--for those who only use one model, as many people seem to--is just how much more often Claude seems to be extremely confidently (and even insufferably) wrong than any of the other three big models (all of which I use quite often... yet I only pull out Claude when I've given up hope in a problem and are looking for out-of-the-box brainstorming).

And like, it does this despite it speaking in extremely dense math, which both makes it sound correct and requires a lot more effort to prove when it is wrong... yet, it isn't actually correct more often, and so that time sink just isn't worth the benefit. I then think many people--including people who can speak math (as can I)--just stop bothering to check everything, as if you come across a human who speaks like this it probably does correlate with slow and careful thought that helps prevent errors.

Instead, Claude has the mistake rate of a somewhat accelerated beginner impossibly combined with the language of an expert professor; and we as humans just aren't good at that combination: it becomes very dangerous and makes it take longer to spot its egregious mistakes and trained-in biases. If you have to use Claude, I thereby claim you really need to have a team of not-Claudes to help insulate you from this, and Gemini (while being a bit senile) is a lot more collaborative and approaches problems in ways that makes it harder to get tricked.

(To translate this into more of an engineering analogy: Claude always feels to me like the engineer who put more effort into learning how to program in functional languages than into how to actually develop working code, and then confidently presents you answers in Haskell or Lisp that never quite work. To find their errors is then very costly. In contrast, Gemini feels more like a Java or Go developer who knows they are a cog... that's helpful! <- Which maybe just goes to show that AI has finally turned me into a manager, omg.)

Not in the principled sense you mean but I have in fact recently started having Gemini explain to me what Claude is talking to me about, lol.

Doesn't it feel like the exact opposite? If you care a lot about the AI being a vengeful god, you do not do what OpenAI and Anthropic are doing. These companies only pay lip service to the idea of that aspect of AI safety in the hope that they can use it to regulate open source AI out of existence.

Instead, they are focused on stuff like "can I ask the AI to help me build a nuclear bomb" or "is the AI willing to generate pornographic stories", which is neither trying to protect us from unleashing a vengeful god NOR preventing (in any honest way) the today-level problems you (very correctly) bring up.

This same level of "oops" in the backend causes users to accidentally see or delete other user's photos, which is definitely a categorically worse scenario.

No that command is "adb install". The "adb sideload" command (which I did not even realize existed) is for installing custom firmware.

If you merely deactivate your email address rather than closing your account or changing your notification settings or whatever, then the next time a legitimate service goes to send you a legitimate email that you asked for, it will bounce. In some sense this "shouldn't matter", as in a purely P2P system the only people who would notice are the sender and Apple -- and Apple knows what is going on, so should not penalize senders the way, say, Google would if they see you sending a ton of email to their domains and they all bounce -- but people tend to use services to help send email and centrally pool their reputation (such as mailchimp) and so these services themselves then watch the bounce rate of their individual customers and either charge them more or ban their access due to the bounce rate increase.

And yet why are people going out of their way to send multipart emails with useless parts? Your analogy does not fit as the Amish person is potentially refusing the HTML part: I just cited my preference, confirming to the protocol, and am being actively sent garbage information as somehow someone seriously thought it was easier for them to go out of their way to give me the option and then laugh at me for making the wrong choice than to just send an email that is HTML only.

Darling stopped using the kernel module a year ago.

But the whole point was that it wasn't a separate role: it was a re-distribution of work so the developers were doing operations. Instead, a lot of people entirely missed the point and just renamed their ops people to devops.

I honestly feel like it is the point at which you find field arithmetic intuitive that demonstrates you have finally understood computers.

Life is longer than a lot of people seem to think. I put a lot of effort into configuring vim and bash... 20 years ago maybe? And now I barely think about it, but the changes I made have made me more efficient and effective throughout those 20 years. Go ahead and put some time into figuring out how you use a tool and make scripts or aliases for using them... if you invest in good tools you will get decades of use out of, instead of chasing dumb fads, you'll just keep compounding your efficiency.

I am unsure, then, as to what your actual goal is? Like, what would you have someone come away from your comment having learned? The comment did not even say the tech IS a doomsday device, it was trying to remove the actual description entirely--so it didn't matter if it was AI or nuclear energy or what--and the placeholder was "doomsday device", as that's what a person who only hears the rest of the sentence is ever going to take away from it.

I do, in fact, think that, if someone says "my technology -- call it what you will, and I won't bother to tell you what it is -- is so dangerous that in the wrong hands or if it isn't carefully contained in a way that I am going to great lengths to try to accomplish it can end the world" -- which is what these people at these companies are saying -- that it would be hard-pressed to not be willing to call that a "doomsday device". I would even go so far as to claim that the exact opposite is more likely true: that to not want to do so requires you to have some extremely awkward bias yourself.

Sure: me using that label to ascribe something related to this product makes some people, seemingly you included, feel bad, and so maybe we can say it involves a "bias", as I clearly can avoid that label, but chose not to, and in so doing colored the discussion in a way that makes you unhappy... but it doesn't make it untrue anymore than any other purely descriptive term is now-a-days often said to be biased. As they like to point out: "reality has a liberal bias" (even if it is just as often liberals annoyed with the terms being used by people ;P).

Again: the bias against the use of that term seems much more fascinating than being upset at the idea that the parable -- which of course is trying to make a point, and so is inherently biased towards that point -- could be attempting to temporary apply it, particularly so as it seems to lead to blindness to what actually happened: people will hear you saying your tech is dangerous, and it will be banned and regulated.

And that's related to why I maintain that I think your example disproves your own attempt at a point: you try to draw a parallel to another technology, as if once we find out it is nuclear reaction research we of course know that that isn't a doomsday device... and yet that is one of the few things that is most obviously to the most people a doomsday device, it is a thing that even the people working on it have always felt might not just be a doomsday device but might be "the" doomsday device (as in, the one that actually does us in), and there has not only been countless people out there worried about what can and will happen in its usage, but it was in fact weaponized and a very very large number of people died due to it, so to try to make the argument that it would be biased to call it a "doomsday device" itself feels very awkward.

But, sure: we can decide that that's biased and throw out the "dunno it sounds like a doomsday device and should be regulated" argument as the argument itself has a bias in it, and that leaves us with... what? A world in which people make these things, talk about how dangerous they are, and then feign in shock and frantically try to backpedal when governments step in to prevent it from being released without careful analysis and regulation and study and planning?

Your example kind of proves your position wrong, no? Nuclear power--another device you are yourself admitting is widely understood to be a doomsday device in the wrong hands--is also highly regulated and I cannot just create a company that sells a nuclear power device and I extra cannot do so and sell it to foreign nationals. If you make a device that you claim is capable of ending the world, you can and should be regulated.

This is the kind of problem I expect Claude to be useless at, and while I could see Gemini Deep Think making a good showing, I'd only bother with ChatGPT Pro. FWIW, I do believe it got the correct answer as one of its first two suggestions (though I am not an electrical engineer, so maybe I am not understanding this given the vague/summarized prompt).

https://chatgpt.com/share/6a2d8c75-56f4-83e8-a61a-301e4c62b1...

This is a steel man of Anthropic's argument, though, and the premise that there could have been a different thing they claimed that would hold up more doesn't and shouldn't defend their position. To the extent to which it comes down to automating and replacing the need for humans or supporting runaway execution that might accidentally kill all the humans, Anthropic routinely measures it, warns of it, and then releases it. Instead, it is only with respect to specific functionality -- much of which is suspiciously beneficial to them, as they internally claim to use AI to improve their own products while also constantly whining about other people using their AI to improve their products -- that they will put a ton of effort into limiting the access or applicability. The day I boot up Claude and ask it to design a website or automate my paperclip factory and it refuses on ethical grounds is the day Anthropic might seem a little less hypocritical.

I do not understand why it being mandated that the vast majority of the people in the world will not allowed to use -- or pay for -- your product (and that the ones that can will have to jump through excessive hoops) could ever make your valuation go up; can you walk me through that one?

Even if this is just temporary, your #3 is more in direct conflict with #2 than you seem to be willing to admit: if you were to own stock in a company that you know has a powerful product and a market lead, but they have been required to take a time out in the market for a year, that should be devastating for their valuation.

FWIW, that's what is so dangerous about AI, though? Not that it will necessarily want to kill us, or even that it will necessarily be able to "want" to do anything, but that we will get in the way of its incessant drive to optimize the efficiency of the paperclip factory that prompted it on a whim before leaving for a long weekend.

The creation--which isn't "his" in the first place, by any standard definition--was not only itself "derived from" our creations but was always supposed to be "open".

Ok, fair, so, to try to apply this to my analogous crime: I would agree that attempted murder does need a definition of "murder"; but, the crime does not care whether your specific plan would have led to an act of murder, only whether or not the defendant was trying to murder.

We thereby do not necessarily need a way to know whether reading--either your book or any book--is addictive or not, but only the extent to which you were going out of your way to make it addictive, for which I think it might then be OK to have some specific-yet-contrived definition that is difficult to apply to any specific product but feels like a wrong thing to do (maybe my "exploit human psychology or physiology")?

I guess I worded that poorly; it isn't merely that we don't need an objective measure: we literally don't need a measure at all, as the crime would be attempting to cause it, whether or not it was even possible to do, and so we simply do not care if the activity was addictive. If you are going out of your way to exploit the psychology or physiology of other humans in an attempt to use that to sell your product, maybe that is what should be illegal.

This would then mean that "our expert witness has strong evidence that my client's product area is not 'addictive', so my client could not ever be said to be engineering addiction" would not be a defense any more than "the plan my client came up with to kill their alleged victim could not possibly have worked, so my client can not be charged with attempted murder" is (at least generally, afaik) not a defense.

Maybe trying to engineer addiction is what should be illegal, and if you want to question "how do you define whether something is addictive" you don't need an objective measure: you determine whether it seems like the people making the product seem to think that's their goal.

But like, once you start using a better window manager (as exist on Linux for X11), maybe you would no longer be able to imagine working in a window that couldn't simultaneously mix tabs from different apps.

All of those are from the same continuity. (The show was not to be a remake/reboot, but also was not to be like, another season of one of the existing shows. I do wish they would just figure out how to finish Stargate: Universe... they even likely-purposefully left the plot/characters in a state where they could actually do the next season now, even after all this time, and even if some of the actors can't or won't be involved anymore.)

> There's a way to tell if this is actually the case: can you find the members of this culture that like getting slopgrenaded? A communication culture needs both speakers and listeners. I see the speakers, I have yet to see the listeners. I could just be missing them though.

The people I know -- including myself -- who do this with each other actually don't mind it when people do this and even go so far as to read through the so-called slop... so, either I know a bunch of people who in fact are listeners, or the definition of the complaint needs to be refined?

FWIW, it isn't really that different from "here's a random Google result I found on the topic" or "here's a random comment I found on Hacker News"... if you don't paste it to the other person you can't have a shared discourse around it, as if we both ask the LLM we will get different answers.

And like, that is definitely something that is rude in many contexts: if I did not respond to your comment but just replied with a link to some other comment I clearly didn't read or bother to understand, that should make you angry. But, if I link you off to something I agree with, that's saving time.

> It seems to me that the issue arises from Apple’s “what-if-ism”... [divorce, estrangement, grief].

I don't think it is these PR issues that cause Apple such consternation, partly as -- even as someone who lives a personal life filled with such corner cases -- I just don't think those are complex issues to solve, but mostly as Apple never seems to put much thought into corner cases like this anywhere else in their business, even when it doesn't butt up against the skewed demographics of software developers (such as how Cydia had much better handling of independent developers and joint projects than Apple's App Store still does 15 years later, and the what ifs were often fascinating to handle).

In reality, the "what ifs" that Apple gets stuck on are lower level, and can even sometimes be spun in a sympathetic light: "what if a domestic abuser manages to automate so much of your software that they essentially have persistent spyware on your device" or "what if a user scripts something to the point where their phone doesn't work quite right and constantly needs tech support" or "what if people share so much of their content with someone else that they share private information without realizing"...

...but -- as is the case with their App Store restrictions that sometimes are reasonable but almost always are not -- the truth is their implied concerns are selfish: "what if a family only buys one iPad for their two or even three kids and we lose 10% of our hardware revenue" or "what if some college roommates declare themselves a family and start sharing purchases of movies and books" or "what if kids in high school (aka, 13+) can still agree to screen time limits they can't change and then don't spend as much time engaged with their phone".

It isn't just that Apple has merely not implemented some of the stuff in this article or doesn't understand what people want: instead, as their business model (like almost all big tech business models) is inherently extractive and even a bit exploitative, their need to optimize for profit is a tradeoff against what people want, so they go out of their way (in ways that are sometimes ridiculous, such as how payments work for family sharing) to make some of these use cases so broken that it forces and/or misleads their customers into spending more (and sometimes a lot more) money to work around the otherwise-arbitrary limitations.

I mean, I never put my code on GitHub, but other people put it there, as they use GitHub: you can't not use GitHub. (Hell: even closed source projects, even ones that were never distributed even as a binary, if the code leaks, end up mirrored on GitHub.)

It isn't just thin: they are quiet and fast with the best trackpads, reasonable keyboards that (except for the idiotic move when they released the touch bar and dropped the escape key) have a reasonable layout that doesn't change much, and all of the power states work correctly every single time.

I am the second most stubborn person I know in my friend group on this, and after only using a desktop for a couple years during the pandemic, I avoided having a mac laptop for the subsequent five years and it sucked. I finally caved after I realized the new M5 Macbook Air is actually likely to be faster for web browsing tasks and is somehow also (awkwardly?!) competitive at compiling code to the monster modern Xeon build I had just completed, and it doesn't even have a fan!

As far as I am concerned, it is over: Apple has won on everything except screen quality (I am sadly now addicted to OLED and I fundamentally disagree with the Apple position of not having a touch screen on a laptop, a stance that is only more emboldened now that I spend a lot of time with children).

I mean, in some sense, Cloudflare simply accepts the security posture of "already lost", right? They run workloads for multiple users within the same process separated by nothing more than V8 boundaries, which even Chrome (which always claimed to run tabs in separate processes but actually didn't due to various edge cases) finally stopped doing (now afaik they do fence origins within processes) as it was so risky... Cloudflare's best lines of defense past "we patch often" are merely that they sort of KYC at least most of their users so they can log everything they run with their identity and that they take users of similar trust levels (age of account, level of KYC, amount of usage, etc.) and group those into processes... but, at the end of the day, they rely on something that I would certainly never consider reasonable to ship in production.