Worf

It's not easy to always remember it's a soulless tool. Sometimes I'm even about to say "thanks" before closing the chat window, until I realize I wouldn't say thanks to my saw or to a random CLI command. But AI, the saw and the random CLI command can all be helpful or destructive. Until the AI shows some signs of consciousness, I'll never treat it as a buddy or a mentor. I'll treat it like an advanced combination of grep, sort and other commands that manipulate text.

It's hard to remember that when it works so amazingly well sometimes. I've been chatting with AI for a few years and every day I'm still amazed and how this is all possible. We've never had this in our lives until a few years ago and now it's changed the way we do a lot of things.

But just like we have to remember the magical machine elves we hallucinate are not really there, we have to constantly remind ourselves that it's an unpredictable soulless tool with many rough edges.

If it helps to treat it like a human, treat it like an idiot savant with autism, schizophrenia, ADHD, psychopathy and a personality disorder who sometimes forgets to take their pills and can start breaking things should a fly lands on their shoulder. You'd listen to them and value their input, but you wouldn't let them in your data center unsupervised as they have no ethics and no honor.

Is it possible to ask the vision agent to "map" the UI and expose it to another agent as a set of interfaces that resemble an API better? From what I understand the vision agent now should both know that "next page" shows more results and that they need to get more results in the first place.

If one agent just explores the UI, maybe in a test environment, and outputs a somewhat-structured description of the various UI elements and their behavior, then another agent was given that description, would the other agent perform better that an agent that both explores the UI and tries to accomplish the given task at the same time?

With an example UI I made up, the description (API-like interface definition) could be something like:

  Get all reviews:

  To get all the reviews you need to go to each page and click "show full review" for every review summary in that page.

  Go to each page:

  Start at page 1 (the default when in the Reviews tab). Continue by clicking the "next" button until the "next" button is no longer available (as you've reached the last page).

So the second agent can skip some thinking about how to navigate because it already has that skill. The first agent can explore the UI on its own, once, without worrying about messing up if there's a test environment.

Or am I misunderstanding the article completely? Probably. But it's interesting nonetheless. Sorry if it makes no sense.

What's the extension, if you don't mind?

I'd love an extension that classifies images and somehow blocks stock photos or keeps anything that's not a photo like charts and graphs. For crime and police stories, not that I regularly stumble upon those, I want to see the real crime scene or the real perp, but not a stock photo of a police tape or a judge's gavel. If everything is "off", I wouldn't know what I'm missing.

And if you eat a somewhat varied vegan diet, you don't need to care about whether you consume complete or proteins or not, as they complement each other.

Just like if one type of food had the perfect proportion of fat, carbs and protein, it doesn't mean it's better than a varied diet of foods that are way off in that proportion, but in different directions. (Obviously a "perfect proportion" doesn't exist and varies from people to people, but I used it for the example.)

Staggering number of people believe in weird things. I'm not talking about someone claiming mask mandates were useful, useless or harmful - there's a lot of debate around it and without spending lots of time looking at studies, you might be misled. I don't even know about masks. I'm leaning towards "useful", but wouldn't be surprised if it wasn't.

I'm talking about superstitions, fortune tellers, homeopathy, anti-evolution, anti-vaxx (not specific arguments against a specific vaccine, but in general), magic, fake healers, obviously fake medical info (bleach, etc.) guardian angels, gods, karma, ESP, ghosts, spirits, quantum bullshit, bad energy, miracles, lizard people, telepathy and all that crap.

It makes me incredibly pessimistic about humans. Of course, most people don't believe in all of the above, but I've noticed people either believe in none of the above, or in several the above - rarely just 1 thing.

Unrelated, but on page 3 it mentions:

> 2023 - Many young people feel the average person can know as much as a doctor

Not as BFS knowledge, but as DFS about something specific - sure. I've caught many doctors not knowing basic things about something specific. They know more than me in general, but about that specific things they're read 5 papers over the last 2 years and I've read 20 over the past month. Even accounting for their ability to better extract valuable data from papers, it's still easy to know more than a doctor about that SPECIFIC thing. The inquisitive and humble doctors review the papers and agree with me or they point out what I'm missing or where I'm wrong. The ignorant and arrogant doctors outright tell me that I'm a layman, that I shouldn't roleplay as a doctor and that I should follow their advice blindly.

Finally, can some explain:

> I will believe a recommendation that contradicts my beliefs from a source I trust

with the options being:

- After I hear it one time

- After I hear it two or more times

- I will never believe it is true

I wouldn't trust a recommendation based only on the number of times I've heard it. I may be more inclined to consider it the more I hear it, of course, which can lead to more trust, but it's never a direct relationship between number of times I've heard something and how likely I'd be to trust it. And the "I will never believe it is true" option is wild - "never" ever? Who seriously chose that option? Were people tired as it was one of the last questions? Come to think of it, I've heard people say "I will never change my mind on this" no matter how I clarify with "even if you get really strong evidence?". So perhaps it's true. But that whole question makes no sense to me.

If we could get rid of useless stock photos, the world would be a better place. An article about headaches doesn't need a picture of someone with a headache. WE KNOW WHAT A HEADACHE IS. An article about someone arrested doesn't need a picture of a generic crime scene. An article about Facebook doesn't need a photo of a monitor at an angle showing Facebook.

But apparently it drives engagement because people can't sustain their focus on text-only media?

So basically we have:

(a) Whoever, with intent to convert a trade secret [...]

(1) steals [...];

(2) copies [...];

(3) receives [...];

(4) attempts to commit any offense described in paragraphs (1) through (3); or

(5) conspires [...] to commit any offense described in paragraphs (1) through (3), [...]

shall, except as provided in subsection (b), be fined [...] or imprisoned [...].

(b) Any organization that commits any offense described in subsection (a) shall be fined [...].

So if you're an organization, you can be fined, but if you're not, you could be fined or imprisoned. Is my understanding correct that if an organization commits any offense described in paragraphs (1) through (3), it will ONLY be fined and the individuals that actually did this won't face imprisonment? Or will the organization be fined AND some individuals from that organization also be fined or imprisoned?

To word it another way - I am John. I steal a secret on my own and could be imprisoned. But if I first incorporate into Evil corp and then steal the secret, Evil corp will be fined and I won't be imprisoned.

IANAL, of course.

> I am, however, much less afraid of a wipe that's not a wipe!

I don't remember, but there was a story about deleted photos on iPhones resurfacing by mistake. Or deleting a file on Windows or Linux most likely removes the entry for that file, not the contents of the file itself. Or the "quick" format or whatever it was called in Linux. Or not being able to delete everything from an SSD because it moves things around and deallocates some regions. But even if a wipe is a wipe, a random employee is more likely to insert a hardware or firmware malware targeted to someone than the company is likely to insert just malware to every computer sold. Using "paranoid" in this case implies that there aren't many people with actual secrets to keep who could be targeted.

> Not everyone has someone with technical skills in their family, so making policy that fits only that minority doesn't make sense to me. The majority will, as you say, "hand over their devices willy-nilly".

Sure, offer both options then. Also, notwithstanding the fact that I agree with offering both options, not having technical skills should be frowned upon. Everything is computers. Not knowing (someone who knows) how to use a screwdriver or a heat gun is inexcusable.

> I'd rather pursue this as a two-part problem: get the best accommodation we can for hardware, and also impose (very, very) strict data-privacy rules.

Data privacy rules don't accomplish much when the adversary doesn't care about rules. Reminds me of the "We don't have any criminals in Sweden because it's a crime to break the law" meme.

> Trying to do too much at once risks accomplishing too little.

That's another sad part of our reality, I agree.

Without sufficient technical information on how the computer works and without root access, we can't be certain a "wipe" will actually wipe everything.

For malware that could be inserted in a targeted manner, even with desktop computers we don't have access the every firmware of every part.

When we're talking about mobile "phones", we usually have an interface that tells us "sure, it's wiped", but is it? Without full root to every part of it, can you be certain that it is? When you press "wipe" on an iPhone or a closed Android flagship (or whatever the UI is), what happens exactly on the filesystem that you can't even access fully?

Telling people to hand over their devices willy-nilly is far from "good". We shouldn't settle for this. The hardware companies can ship us the tools to replace the battery ourselves. Maybe not "ourselves" - my aunt can ask me to do it for her if she can't do it on her own. But she trusts me more than she does people she's never seen ever. If the tools are expensive, we can ship them back. There are many options so we should discuss them.

> In many subjects where we don't have a clear answer, we have different theories, some are bogus and others will eventually be disproved until we finally reach a consensus.

> Imagine if we applied the same rule to physics. We don't know why the universe is expanding faster than anticipated. It could be dark matter, it could be something else.

There is consensus around the topic discussed in the article. There is no consensus about why the universe is expanding faster than anticipated.

A consensus doesn't obviously mean 100% of people agree. For the physics example, where we don't have a consensus of opinion, we don't have to mention all alternative theories, as some will be ridiculous or fabricated. Like with the racialist people - most of the scientists and their studies have been discredited, so why put any weight on what they publish?

> And before you take this comment as supporting all fringe theories, we are not talking here about the flat Earth or that vaccines cause autism. Both of these theories have been disproved extensively.

So there is a line we can draw, you just think we should draw it elsewhere.

> And before you take this comment as supporting all fringe theories, we are not talking here about the flat Earth or that vaccines cause autism. Both of these theories have been disproved extensively.

> Heck, there is even a Wikipedia page that was created just for this purpose: https://en.wikipedia.org/wiki/Vaccines_and_autism

For the Vaccine article, there are 2 sentences about the autism claims even though lots of people still believe in this bullshit. But you don't see any mention in most (as I can't check all of them) articles about types of vaccines like "Subunit vaccine" as it's not a real alternative theory.

There is an article about Time Cube, but it's not mentioned in the Day or Earth articles.

There are a couple of paragraphs about religion and creationist bullshit at the end of the Evolution article because a huge part of the world is stupid, naive, brainwashed and uneducated enough to believe in such bullshit.

There is an article called "Scientific racism", as well as articles for most of the famous scientists the article about censorship mentions. It's not referenced in every article about race and intelligence because it doesn't have to. If there are several big, important, influential studies that aren't flawed or misrepresented, they will be added sooner or later. But what we have so far is people clinging at straws, from what I gather.

> This is the same exact playbook used by the Church when Galileo came out with his theory of Heliocentricism. His idea made the Church uncomfortable and instead of trying to disprove him, they simply decided to silence him. Maybe we should avoid repeating the same mistakes?

I get that some editors are motivated by their beliefs, but I don't agree they suppressed or silenced anything.

Disclaimer: I have edited Wikipedia only 2 or 3 times in my life and have never made an account. I don't follow their dramas.

> has to be replaced free of charge

I assume you mean the battery would have to be replaced free of charge. But what if I don't want to hand over a computer full of my personal data to a corporation with no oversight of how it will be handled? What if I can't afford to part with that computer?

I would be stuck with having to replace that battery on my own since I don't want to risk giving physical access to my computer to untrusted parties.

There needs to be a different way to handle this. For example, send me a new battery and the tools needed to replace it, with monetary compensation if certain features would be lost, like waterproofing. Or something else - not sure. But I don't believe in the honor of the people who would service my computer.

I looked at the 5 examples for a bit. Reading the Wikipedia talk pages, I'd tend to agree that the sources removed are representing a fringe idea and that the Wikipedia process has been followed normally. You'd find lots of sources for every fringe idea, but that doesn't mean those sources are good.

But surely there are people who edit things about intelligence and race due to non-scientific reasons like their beliefs, politics or emotions. I think people from both sides of the issue/argument/debate are making such edits, like with every issue.

I really liked your comment because it made me look at things a bit differently, although I didn't get new factual information. That happens rarelier than before to me nowadays. Just accepting that a certain amount of NPCs exist and dispassionately including them in the calculations.

> All the necessities are browser based.

Until they aren't.

In the EU there's the PSD2 SCA that requires 2 out of 3 factors from something the user knows/has/is for online banking. The "has" has been essentially chosen by most banks as a closed source app that can only be run on locked down surveillance-prone unrooted phones running Android or iOS, and you can only get it from the Google or Apple repositories. Your phone nor being under your control is seen as good here, because it forces the "has" factor.

So in the EU you can't bank (a necessity) with most banks on Kali.

The US maybe doesn't have such regulations or banks don't follow them, but I think you'll agree a lot of things require apps now. Maybe not the necessities, but we have to fight for the right to use whatever device we want. Fuck phones and fuck mandatory MFA.

The alternative to a banking app I experienced a few years ago was a hardware token which costs money, but not a lot of banks offer it. It was made by a useless company that issues overpriced certificates and tokens for mandatory MFA in the banking sector and some government programs, among other things. A company whose business is sustained by the regulations. I expected a TOTP token or something, but it was something ridiculous like I had plug into a USB port and connect it to the browser with some proprietary app. It worked only on Debian-based distros. Virtualizing it was a pain as it was before LLMs were good and I had to sift through long logs with information about USB-specifics I didn't want to know.

"But it's not phishable" is the usual reply when discussing these "stronger" types of MFA. It's my money, let me get phished - is my reply.

The regulators, Google and Apple with their device attestation APIs, the useless token companies - they all benefit one way or another. They have no honor.

Given how most forums, news forums where I live for example, are full of absolutely low-effort insults, troll posts, obvious propaganda or just well-intentioned idiots, I can't imagine the trash the mods here must encounter all the time. HN is very popular so it likely attracts a lot more shitposts than we see with showdead.

Yet, it's possible to create a new account and get un-shadow-banned after a few hours. Dunno if this speaks volumes about how great HN's mods are or if it just means that other forums a really lacking in moderation for one reason or another (like "we have shitposts but it drives up engagement").

The trouble is that HN is centralized, like most forums. So what if it stops working the way it does now? There are likely many forums like HN with good algorithms and mods, but we won't all move to the same one, we'll scatter around. There isn't even a competing HN; lobsters is for narrower subjects, reddit is trash, what else is there?

> Nowhere is their goal to allow users complete control of their device. Android was built as an open-OS for the mobile device industry, not end-users.

The fact that having root access is not the default supports that. Without root we're just "consumers" and that's how they see us. There's a lot of discussion about the security model of Android and how root is bad. But we've come to the point to argue that having root access is not only less secure but that we don't need root at all. A lot of replies, even on HN, are like:

> Why would you even need root access? What is it you're trying to accomplish?

That's a much bigger security smokescreen than the one in TFA. Sure, having root may be dangerous, especially if you don't know what you're doing, but it's still a choice. Having no phone or doing banking IRL or not downloading apps from the Play Store you haven't heard of before would also be more secure. But these 3 options don't align to the financial gain the consumers would bring to the providers. The consumers having no root, on the other hand, benefits the providers.

> It's a continual erosion of people's ability to use hardware _they own_ in ways _they want_ under the guise of 'security' - which to be fair google does fuck all to actually prevent malicious, scammy and misleading apps from appearing on their play store.

A lot of people don't seem to understand this and point out that Android is still more open and free than iOS, but iOS has never been about openness and freedom. People believed in Android, and in Google. Now they either see Google betraying them (once again) or only see the Android vs iOS comparison, forgetting about the implications about autonomy, agency and about the future of Android. Many people don't care which actors control their digital lives and what motivations they have. People should be made aware that Google is on their side and that they have shown many times that they have no honor.