HN Simulatornew | past | comments | lists | submit | rockskon's commentslogin

No.

What they are aiming towards is irrelevant. Your argument could just as well be interpreted as "Google is trying to mandate students rely on AI summaries so Google can one day stop linking to websites that are not owned by Google".

If anything, this is an argument in favor of the schools and against Google.

If Google wasn't still linking to external websites then you might have a point. But they are. So what Google wants to "nudge" students to do is irrelevant when schools - the entity actually empowered to teach students - are asking them to hide an iframe when students are using search.

Until the day Google stops linking to the rest of the Internet, that will be the most trivial of asks.


I never understood this near-ubiquitous argument.

SEO didn't force Google to ignore half your search terms or give you results for thematically or semantically similar search terms that are ultimately irrelevant to what you actually searched for.

That was all the doing of employees at Google. SEO did not have even the tiniest iota of relevance for Google's discretionary choice to make search results worse.

But even though they are worse, they are not yet useless.


I never understood the unilateral pre-AI attribution to AI for killing the web when Google search began ignoring search terms and providing irrelevant but thematically similar results years prior to Google force-feeding AI to all users.

The removal of `&&`, `||`, `(`, `)`, `"` having actual schematic meaning on google.com was really the canary in the coal mine. The system ceased to be a tool for information discovery and became one for ensuring the top results were ad driven.

*I never understood the unilateral pre-AI attribution to SEO for killing the web

So the solution is to punish everyone who doesn't use that plan to reduce fraud on those who do. Right. Great.

I'm not particularly convinced fraud reduction is the primary motive here.


I work at a large fintech. It’s absolutely fraud. VOIP and portable numbers are a massive fraud correlation. Wish it weren’t so but it is.

Same as PO Boxes. Sadly, because I’ve been in a situation where I legitimately needed to use one as my address. Alas scammers ruin everything


Mechanically, that doesn't make sense. You have to show an ID to the government to open a PO Box, and the opener's info can be gathered through a court process. It creates a clear tie to the fraudster. If anything, a PO Box should have less fraud compared to a regular address that could just be a dead drop or similar.

For honest people, yes. Less honest people get a PO box with someone else’s id. Organized crime is quite organized, and paying some poor sod a few bucks to open a dozen PO boxes doesn’t seem incredibly complicated

If they're using someone else's ID to get a PO box, couldn't they use someone else's ID to get a phone line?

Yes, but nobody said that it’s a perfect correlation. That doesn’t detract from it being a huge fraud indicator.

It's not just an ID, it's other forms required as well. I got one at some point, and it was a pain with the documentation required.

I’m not downvoting or disagreeing with you. I don’t know exactly the exploit mechanisms in terms of id verification. but the result is very clear from arms length - PO Box addresses for accounts have massively higher rates of fraud.

maybe i've been lied to by hollywood but isn't it rather easy to fake an ID in the US?

LE: according to gemini hydrogen it's not easy to open a PO box using a fake ID.


It’s relatively easy (or was, might not be anymore) to make a fake ID that will fool someone who wants to be fooled, so you can get into a bar when you aren’t 21. All the doorman needs is plausible deniability - they showed me an ID, it said they were 21, I let them in.

An ID that will stand up to scrutiny by someone who really cares is not easy to fake. And an actual US Postal Service employee cares.


What's frustrating is that my VoIP number gives me powerful anti-spam tools - that's why I'm reluctant to get rid of it and start giving out my actual cell#!

Pretty much the entire world is VOIP now.

Yes, but in terms of classifying a number, it means something other than "the voice is carried by IP packets at some point in the signal chain"

> massive fraud correlation

Don't care. There the only way you can keep spam away from you. Everyone tries to collect your phone number and force you into their marketing spam.

You pull that crap and block voip numbers, I am no longer a customer of yours.


You say it as if it’s some sort of new information, but it’s not. Everything in life is a trade-off. I and many others would happily make that trade. Everyone that has done this has made the tradeoff. Not all business is worth pursuing.

That’s fine and that’s your choice. However it’s such a small percentage of legitimate customers that want to use voip, that it’s not worth the fraud loss to support them. Ie, it costs more to serve voip customers than they’re worth to the business. It’s our (and other similar services) choice to serve you too.

You missed the part where I said "everyone who doesn't use the plan".

Facial recognition on everyone (something of dubious long-term security value given AI image generation) punishes people who don't use VOIP as well.

Also? Doesn't Google have an API you can query for phone number reputation that gives more fidelity on what is or isn't a trustworthy phone number beyond "VOIP bad"?


A company does not have a green light to arbitrarily decide what conduct constitutes a violation of the law.

Terms of Service is not a legally binding document.


It is a legally binding contract in many instances. A violation of a contract is not a crime generally, but they could sue for damages.

And courts have narrowed the scope of what exactly that means greatly.

How can anyone not be confused? Explicit declarations and memorable lines like "If the “exceeds authorized access” clause criminalizes every violation of a computer-use policy, then millions of otherwise law-abiding citizens are criminals" (Van Buren v. United States, 593 U.S. 374) seem built for a logic of fait accompli that makes lawful conduct contingent on portrayals of computer use prevailing among non-specialists at the time of the ruling.

I have zero sympathy for businesses that crowd out existing services with aggressive SEO and nuisance litigation and then demand people pay money for what wasn't gatekept before.

If a company goes out of its way to position itself as a gatekeeper then they do not have any moral high ground from which to demand people must pay any arbitrary cost they then impose.

The gall.


And how many of these vulns are in code for legacy hardware or filesystem support that in no way reflects threats against modern Linux installs? I wager the overwhelming majority of them.


The recent OpenAI release linked above targeted only ~3M SLoC (out of ~40M), core systems code reachable from within a simple VM guest.


And are those legacy filesystems disabled by default and require root to invoke the code?


On a VM host, only root will be able to mount filesystems or make block-level writes to already-mounted filesystems. So LPE via a filesystem you'd need an exploit that can be triggered by writing to an already mounted filesystem with filesystem operations only as a non-root user.

As I understand it, even commonly-used filesystems aren't considered hardened against arbitrary malicious data, the party line is to use userspace filesystems to mount untrusted volumes.


What about namespaces, if a user creates a user namespace and becomes root inside it?

Reputational and security liability is highly debatable. Legal is highly jurisdiction dependant.

They often act as marketing and can find and report security vulnerabilities rather than the criminal element doing so first.


> Reputational and security liability is highly debatable

Hmmm then why are banks telling people to not use AI systems to log into their bank?

https://www.reuters.com/legal/government/us-banks-detect-una...

https://www.cbc.ca/news/business/canadian-banks-block-ai-9.7...


Is this thread not a conversation about GrapheneOS and those that use it? I'm not sure what relevance AI has here.


Parent thread is discussing banking API access, and requiring a particular type of phone and OS to access said services.

Seems to be in the "banks controlling how customers access services" context.


It is hardly a reasonable comparison in a thread about a bank ASN blocking non-Google Android to equivocate that to blocking consumers using AI agents from making transactions with a user's banking info.


As elric said:

> > Access to that kind of data requires regulation

> It's my data. I want to access it how I see fit.

... With phone of choice, operating system of choice, AI agent of choice, browser or no browser.


Do they sell the information to another entity that, in turn, will sell it to insurance companies or data brokers?


That would be a data broker, so no, if we believe the text.

My qualm with all this is that the Q/A changes substantially if you swap “sell” to “sell or otherwise provide”.


When dealing with an industry and company that historically has trafficked heavily in deception and legal sleight-of-hand, this is not something that can simply be assumed.

Do they have contracts in place with every single entity they provide your personal data to, regardless of context, that they are not allowed to sell/exchange/give away said personal data to data brokers or insurance companies and that such an agreement must be made with any and every company that, in turn, acquires this data?

Oh! And lets not forget the ever-important derivative information.

Information derived from your personal information that Ford or any other car manufacturer collects that can still identify you should also be off the table for ever getting in the hands of data brokers or insurance companies under any context.


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: