Hacker Newsnew | past | comments | ask | show | jobs | submit | jtbayly's commentslogin

Of course we would!

How in the world can you double check the AI-generated tax filing without going back and preparing your taxes by hand?

You might skim an ai-written email.


But did the tractor owner win the lawsuit? Anybody can sue anybody. Winning is what matters. That anecdote doesn't tell me anything, and the whole article reads as FUD. I wonder if it is paid for by John Deere?

Not 100% sure about that one - though even that they were sued is an issue.

Note that part of this is also that worker's comp excludes farm labor. So you can't get compensated through workers comp.

https://nationalaglawcenter.org/workers-compensation-for-agr...

> Whether or not someone is eligible for workers’ compensation depends on their state and the industry they work in. There are no requirements at the federal level that mandate states to have workers’ compensation laws. Nevertheless, every state, except for Texas, requires most employers to carry workers’ compensation insurance. However, the majority of state workers’ compensation laws specifically exclude or limit agricultural employers from the workers’ compensation requirement.

Product Liability in the Farm Equipment Industry - https://youtu.be/NdN577BbnSY

Take note of 42:23 where it goes to who is at fault and the liability for it (the case study starts at 21:21 - jumping to 42:23 you can get the "who is paying for it" in two minutes). The entire video is interesting (if you're interested in product liability for industrial (farm) equipment ... but I can understand someone not wanting to watch an hour long video about it). It boils down to "every piece of farm equipment is dangerous from the insurance perspective and a manufacture allowing unapproved modifications to the equipment is still at fault, even if the modification was made by another party."


>But did the tractor owner win the lawsuit? Anybody can sue anybody.

Yes, but this requires the accused to pay for legal counsel and go through an expensive trial.

Maybe the US should fix this.


Agreed! A fix would be nice. But owning a tractor could get you sued. Going outside could get you sued. Being a good neighbor could get you sued. There's no reason to think that modifying your tractor makes you more likely to get sued. Let alone the dealership.

The very act of existing as a corporation with a lot of money is what makes you a lawsuit magnet. Nothing you can do will prevent that. Trying to prevent people from making modifications to their tractors they bought from you is just as likely to get you sued as not trying to stand in their way.


There are corporations that are well-known to use lawsuits as a way of silencing people. They have lots of money to keep lawyers on the payroll and bankrupt their opponent.

It is 100% FUD, as evidenced by the fact they were afraid to link the actual case.

https://www.govinfo.gov/content/pkg/USCOURTS-ohnd-1_12-cv-01...


I doubt it. Good luck.

I think I have the same problem on Tahoe.


I agree completely with your first paragraph, but I'm not sure what privatization has to do with it. Also, I agree that more regulation of private parties is needed. Or even better, break up the private companies that are like multi-state governments in terms of power.

modeless linked to this article earlier today:

https://james.darpinian.com/blog/apple-imessage-encryption/

My current understanding of the facts:

1. Google defaults to encrypted backups of messages, as well as e2e encryption of messages.

2. Apple defaults only to e2ee of messages, leaving a massive backdoor.

3. Closing that backdoor is possible for the consumer, by enabling ADP (advanced data protection) on your device. However, this makes no difference, since 99.9% of the people you communicate will not close the backdoor. Thus, the only way to live is to assume that all the messages you send via iMessage will always be accessible to Apple, no matter what you do.

It's not like overall I think Google is better for privacy than Apple, but this choice by Apple is really at odds with their supposed emphasis on privacy.


Enabling ADP breaks all kinds of things in Apple’s ecosystem subtly with incredibly arcane errors.

I was unable to use Apple Fitness+ on my TV due to it telling me my Watch couldn’t pair with the TV.

The problem went away when turning off ADP.

To turn off ADP required opening a support case with Apple which took three weeks to resolve, before this an attempt to turn off would just fail with no detailed error.

Other things like iCloud on the web were disabled with ADP on.

I just wanted encrypted backups, that was it.


That chimes roughly with my experience, but to be fair ADP is designed not just for encrypted backups, but to harden the ecosystem for people who may be under the greatest threat. Worth noting that it has been outlawed in the UK and cannot be enabled, which makes me think it's pretty decent

> Worth noting that it has been outlawed in the UK and cannot be enabled

For the record, there is an ongoing court battle between Apple and UK government about getting it overturned.

Which also says many positive things for Apple that they are willing to put their money where their mouth is and put up a fight.


And that’s a significant PR and marketing posture for Apple.

You wouldn’t happen to work in North Norfolk would you?

FWIW I've been running ADP for over a year now, and so far I haven't noticed any problems.

iCloud on the web not being available is kind of expected; how would it work with E2EE?


Huh, that’s crazy because ADP doesn’t break anything for me. Then again, I’m not trying to connect an Apple Watch to a tv. What a simple life I live.

Apple's other emphasis is customer experience, and there are more "I forgot my code, help me recover my stuff" people than you can imagine.

It would be bad PR for Apple if everybody constantly kept losing their messages because they had no way to get back into their account.


That’s all fine, but then show the sender whether their connection is actually end to end encrypted, or whether all their messages end up in Apple’s effective control.

One might consider differently colored chat message bubbles… :)


You think there are fewer people who forget using Google devices? I don’t it. The article talks about how Google prevents that from happening.

ADP isn’t the default, and almost nobody who isn’t a journalist/activist/potential target turns it on, because of the serious (potentially destructive) consequences.

How does Google manage this, such every normie on earth isn’t freaking out?


Nobody expects their text messages to be backed up.

They get deleted and people shrug.


Or IOW, Googles solution affects only messages. Apple’s solution affects your whole digital life so the consequences are a lot more dire.

> Apple’s solution affects your whole digital life

I don’t know if that’s generally true. I could lose my apple account and not really give a a damn. Not that I see how such a thing would happen, save for apple burning down all their datacenters. I’m running ADP


Google’s solution also ensures that they know all the metadata of your messages, except the content of the message itself.

Apple too collects unencrypted metadata but now promises to reduce its scope.

How convenient...... indeed

I keep my messages and would like them to not go away.

Why?

I reference them every so often

> because of the serious (potentially destructive) consequences

Huh? What are you talking about? I don’t see anything destructive about it.


People don't always have enough Apple devices to justify confidence that they couldn't lose them all at the same time, which with ADP is a permanent death sentence if you don't have your recovery key.

(Apple says you can also use a device passcode; I'm not sure if this works if the device is lost. Maybe it does?)


I have 2 or 3 yubikeys associated with my account. I think apple does a decent job at communicating the importance of having recovery keys to the point where they deter those who can’t be bothered.

Yubikeys are great


I'm always put off by the incredibly low limits on yubikeys. What's the point of having a security key if you can only have 25 accounts in its lifetime? What are you supposed to do, buy tons of keys and then figure out a system to remember which key each account is? Like fucking hell just let me use passkeys in iCloud Keychain. My bank's mobile app specifically supports only security keys and explicitly not passkeys for literally no reason because passkeys are practically just as secure as any security key. It's actually harder to specifically exclude passkeys and allow only security keys than it is to just use passkeys which automatically include security keys.

How do maps changes make sense to subscribe to when they are on OSM?

And what do you even mean by subscription to changes to the law?


If OSM is up to date - many places it is very outdated. (others it is very good).

Law - when a government changes the driving laws. Government can be federal (I have driven to both Canada and Mexico. Getting to Argentina is possible though I don't think it has ever been safe. Likewise it is possible to drive over the North Pole to Europe), state (or whatever the country calls their equivalent). When a city changes the law they put up signs, but if a state passes a law I'm expected to know even if I have never driven in that state before. Right turn on red laws are the only ones I can think of where states are different - but they are likely others.

Laws also cover new traffic control systems that may not have been in the original program. If the self driving system can't figure out the next one (think roundabout) then it needs to be updated.


> sometimes they don't have one when you arrive.

Or, if they are Hertz, they might have one but refuse to give it to you. This happened to my wife. In spite of payment already being made to Hertz corporate online, the local agent wouldn't give up a car for a one-way rental. Hertz corporate was less than useless, telling us their system said was a car available, and suggesting we pay them hundreds of dollars again and go pick it up. When I asked the woman from corporate whether she could actually guarantee we would be given a car, she said she couldn't. When I suggested she call the local agent, she said she had no way to call the local office. Unbelievable.

Since it was last minute, there were... as you said, no cars available at any of the other rental companies. So we had to drive 8 hours to pick her up. Then 8 hours back, which was the drive she was going to make in the rental car in the first place.

Hertz will hurts you.


I personally would have changed it to a round-trip then just returned the car to the other Hertz location and let them figure it out.

Hertz this time, but things like that have happened with every rental company I know of.

I don’t understand why you’d try to use an LLM for that step if there is already a tool that you can call to check it. Help me out.

I just ran into the problem of extremely slow uploads in an app I was working on. Told Gemini to work on it, and it tried to get the timing of everything, then tried to optimize the slow parts of the code. After a long time, there might have been some improvements, but the basic problem remained: 5-10 seconds to upload an image from the same machine. Increasing the chunk size fixed the problem immediately.

Even though the other optimizations might have been ok, some of them made things more complicated, so I reverted all of them.


Actually, you're allowing a much higher percentage of cheaters if you read the paper. They optimized to avoid false accusations. It's only ~45-75% accurate at detecting AI writing. It's closer to 90% accurate at detecting human writing. Half the cheaters get through, and you still fail 10 percent of the people who didn't cheat.


> It's closer to 90% accurate at detecting human writing.

I know that's what they wrote, but I heavily disagree. It got 28/30 (93%) correct, but out of the two it got "wrong":

- one was just straight up not rated because the file format was odd or something

- the other got rated as 11% AI-written, which imo is very low. I think teachers would consider this as "human-written", as when I was being evaluated with Turnitin that percentage of "plagiarism" detected would have simply been ignored.


At this point the most basic users of could be easily picked off and that style and list will grow yearly.


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

Search: