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

Where is Claude's checkout? Do you have them all share the same local files or does each use its own copy?

People are doing this lots of different ways. Some run it in its own containers or in instances on the web. Some are using git worktrees. I use a worktree for anything large, but smaller stuff is just done in the local files.

Sloppy? Perhaps, but Claude has never made such a big mess that it has needed its work wiped.


> Sloppy? Perhaps, but Claude has never made such a big mess that it has needed its work wiped.

I think a key thing to point out to people here is that Claude's built in editing tools won't generally allow it to write to a file that has changed since last time it read it, so if it tries to write and gets an error it will tend to re-read the file, adjust its changes accordingly before trying again. I don't know how foolproof those tests are, because Claude can get creative with sed and cat to edit files, and of course if a change crosses file boundaries this might not avoid broken changes entirely. But generally - as you said - it seems good at avoiding big messes.


Perhaps it should be by user base?

"You have 100M MAU, you need to be keeping up with the standards"


How do I prevent my citizens from sharing their certificates in order to bypass the block?


From what I can gather from the EU proposal[1], they rely on such details to be stored in a hardware crypto module on a phone or similar.

Thus the user never has direct access, yet can use it to issue proof of age.

[1]: https://ageverification.dev/av-doc-technical-specification/d...


This is even more Draconian because it demands that citizens rely on manufacturers. After all the government itself doesn't make phones or hardware crypto modules.


You don't, it's up to citizens to make sure whatever authentication they use can only be used by them, just like how it works for other services today where you authenticate online somehow and the government service assumes you're you since you were able to authenticate.


My point is that this is either a bearer token (in which case it will be obtainable by proxy) or tied to your identity.

What is the incentive for the citizen to make sure their authentication isn't shared?


On the government endpoint, which returns X that the platform uses as "evidence" for you being an adult, yes, that's tied to your identity, as the certificate/whatever is tied to your identity.

But as long as the platform who need to validate that you're an adult don't get your identity, but just the proof, I don't see what the problem is?

> What is the incentive for the citizen to make sure their authentication isn't shared?

What incentives do people today have for keeping their identifications to themselves? Why aren't we all sharing CC numbers? Because we realize some data is "personal" and isn't to be used by others, like our username+passwords or whatever. This isn't exactly a new concept, just look at how it works for anything else that is tied to you.


> On the government endpoint, which returns X that the platform uses as "evidence" for you being an adult, yes, that's tied to your identity, as the certificate/whatever is tied to your identity.

In this scenario the government knows all the age-restricted sites I've visited. I'd argue that is worse than if all the age-restricted sites I've visited know who I am...

(FTR I don't know what I think about age restrictions in general, but I'm pretty sure there's no implementation that comes without negative side effects)


Not necessarily. The age verification proof doesn't need to be site-specific. But again, that reduces the incentive "for the citizen to make sure their authentication isn't shared" because there's nothing tying it to them.

I also kinda hate the whole idea of needing explicit permission from the government to access the open web, regardless of whether or not they know which specific sites they're giving me permission to access.


There's actually a much better idea that's been floating around. Require over-18 sites to set a certain header. Then anyone who wants to can install a browser on their kid's device that will block pages with the header. There's no privacy implications, no surveillance implications, no need to make VPNs illegal as long as they pass it through; it's just a plain old parental block with a regulation keeping it always up to date. Yes, you may have to stop your kid installing random software on the device to bypass whatever blocking you set up, but you had to do that anyway. If it's Apple or Google they could easily enough require everything in the app store to respect the flag when the device is set to kid mode.

(If the government does the incredibly overbearing thing and does not do the simple and effective and unintrusive thing, it proves their motivations are surveillance)


Already exists; the industry called it RTA (Restricted To Adults). Nobody used it... and it's 19 years old. Complete failure categorized under "we already tried that."

https://www.rtalabel.org

You can use it too, just put this in as a meta tag:

<meta name="RATING" content="RTA-5042-1996-1400-1577-RTA" />

Or send the following header:

Rating: RTA-5042-1996-1400-1577-RTA


Was it legally mandated? I think that's the main idea GP is proposing. Obviously without any incentive to actually implement it there's no point.


I don't think that it matters. The big porn sites have served RTA tags for many years. Android, Windows, macOS, and iOS can all be configured to block adult content tagged with this system. That still hasn't stopped a bunch of states from passing age verification laws ostensibly targeted at protecting children from these sites.


Because the laws aren't about protecting children. If they were, they'd say browsers have to enforce this header and sites have to send it.


In order to accomplish that working, you'd have to legally mandate parents put the blockers on their kids devices.

Similar things exist that block based upon lists and content keywords and such.

Most parents do not want to block stuff from their kids or they would be.

If thousands of them demanded that devices came with blockers then the market would provide such devices.

Many moons ago you could argue parents did not know what the youngins would find on the internet. Today's parents definitely know, and most do nothing to restrict access.


It doesn't have to be mandated - parents could choose.

Even if it's mandated that kids can only use phones with a special "kid mode" turned on, even if you had to present ID to turn it on or off or buy a phone with it turned off, that would still be way less bad than what's being rammed through parliaments right now.


Germany let's the ID cards themselves mint such proofs in a way similar to how Intel's attestation doesn't leak the CPU's serial number.


> What incentives do people today have for keeping their identifications to themselves?

Not being liable for loans they didn't take out themselves, being the recipient of government benefits they are owed, etc. I'm sure you have heard of identity theft before, but it sounds like you haven't heard of why it's a bad thing. It's not just a privacy thing.


If you share your CC number, someone could steal your money. If you share your anonymous age verification token... someone could pretend to be 18? And by design that token is anonymous and there's no way to prove you were the one they got it from? Doesn't seem like much of a disincentive.


> obtainable by proxy

So no different to the rules around buying an 18+ DVD.


How do they solve this for e-voting?


I'd love to get a cargo bike and use it for kid transport.

I would be worried about collision safety though, I am not going to persuade everyone in my neighborhood to stop using cars in a hurry and there are not bike routes between me and school, library, shops, ...


This is the classic urbanist anti-bike tragedy of the commons that is referenced often.

People use cars because they are (rightly) concerned about safety. People avoid using bikes because there are so many cars. It’s very hard to ban cars or restrict car usage because it seems like no one wants to use bikes, but it’s a self-reinforcing system.


And the losers that neither (gross generalization here) the bike people or the car people care about are the pedestrians. We are lucky to live in an urban environment where our family of 5 usually walks everywhere. Crossing the street with children is an unwelcome adventure. But you are absolutely right I would ride an ebike with the kids if it did not seem so dangerous.


Solution: make cars safer (as in, less likely to harm others) through road design, and things like automated emergency braking and self-driving.

Then the cars are safer both for the occupants and pedestrians/cyclists, so paradoxically people might be more inclined to walk or cycle.


I'd sooner see a more practical solution that works on the roads and cars we already have.

Aggressively limit speed and enforce it until you're onto the fast roads.

If cars could only roll on at 10mph I'd feel a lot safer and I'd probably be able to use my bike and make better time for the local stuff.


Yeah, I think I’ve seen a statistic that getting hit by a car going 30km/hr, you’re probably going to be fine. If you get hit by a car going 50km/hr, you are most likely going to die. I can’t think of any reason to allow cars to travel over 30km/hour in urbanized areas, outside of designated arterial roads.


Enforcing speed limits requires constant labor and doesn't scale.

Making adjustments to roads requires some upfront capital and scales to every car on those roads.

Check out this great resource of traffic calming measures from the Institute of Transportation Engineers[1].

Chicanes, Chokers, and Corner Extensions are just three examples of measures that can be taken temporarily and cheaply.

1: https://www.ite.org/technical-resources/traffic-calming/traf...


Well, while we're talking sci-fi: personal teleporters.


I mean it's happening already, cars are getting safer aren't they? Driver assistance features are gaining adoption. It takes time.


I assume by "getting safer" you mean that they're getting safer than they used to be, which, I suppose I agree. I don't know what you therefore think is "happening already", though. I see no indication that autonomous cars will ever be safer than human drivers.


You obviously do not live in Chicago or Minneapolis. Winters make this a nonstarter


I have an Urban Arrow since the beginning of last February and live in Minneapolis. I’ve used it about 5 days a week for getting my pre-k kid and newborn to the places we need to be, since we bought it. Each day it’s between 6 and 26 miles. That upper end comes from busy days, probably once or twice a month I get above 20 miles in a single day. Definitely not doing that on the coldest days. Studded tires really help.

One unexpected benefit is that the muddy/wet boots don’t muss up the bike like they would if I was loading them into a car. Just drips out the bottom grate.

Lots of other small benefits but not so related to winter.


You think that people don't ride bikes in the winter in Finland or Sweden? Riding in the snow with "snow tyres" for your bike is normal. There are whole YouTube videos about the phenom.


Minneapolis average January temps: about 9°F / –13°C (low) to 24°F / –4°C (high).

Stockholm January average: 25°F / –4°C (low) to 32°F / 0°C (high) — closer to Chicago than Minneapolis.

Having been splashed by busses in winter you have to be a special kind of crazy to ride a bike or motorcycle in Minneapolis or Chicago.

There is usually a week in January in Minneapolis where the high temperature for the day does not break -10F. Air temperature, not wind chill.

Minneapolis at least has a skyway for pedestrians in winter. Chicago loop, not so much.



Interesting choice since Minneapolis comes up pretty often as a pretty good biking city, by US standards at least. https://cityratings.peopleforbikes.org/cities/minneapolis-mn


It is quite good in comparison to many US cities. There is a fairly good distribution of segregated bike pathways through the city, and despite the skepticism in this thread - those pathways do get plowed quickly in the winter. There are fewer people on the paths in winter of course, but winter commuting comfort mostly comes down to wearing appropriate clothing.


> It’s common to rack up 4 hours or more of screen time a day on your phone. Here’s one way to see the cost of that: every 20 years, you lose 5 years of your waking time looking at your phone.

If you spend 4 hours/24 hours on your phone then every 20X you'll have lost 3.33... X.

I think the author is using year and waking-years but it doesn't parse well for me because you don't get close to 20 waking-years for every 20 years.


Is it? I remain unconvinced AI will replace Devs in the foreseeable but I don't think this article is passing comment one way or the other.

The way I read it TFA is claiming that the Devs who built the Natascha AI suite did solid work.

Contrary to prior reporting there really was an LLM based product at Builder in addition to their earlier business of a large pool of contractors cranking out apps on demand.


Ah, I thought this might be for turning tiny/tracking URLs into their targets like my Hacktoberfest project from a few years back

https://off-the-rails.netlify.app/


Cow practical joke anecdotes, please! :-)


Real simple stuff. Luring a cow over near a bank and then pushing her down it. Pretending there's some lush grass in the gap between trees. Nudging each other into the electric fence, or into the gatepost when going through a gate.


Doesn't this give you 2xSPF? Or can I use my local copy of the source to kick of Travis/Circle?


Ideally you don't ever rely on CI specific automation tooling to actually accomplish anything and instead just use it as a dumb "not my dev machine" to execute workflows.

You should always engineer things so you can fall back to something akin to:

./scripts/deploy_the_things

Ideally backed by a real build system and task engine ala Bazel, Gradle, whatever else floats your boat.

It also means you are free to move between different runners/CI providers and just have to work out how to re-plumb secrets/whatever.

GH actions/friends really provide minimal value, the fact they have convinced everyone that encoding task graphs into franken-YAML-bash is somehow good is one of the more egregious lies sold to the development community at large.


Well, git ops might not be impacted on Github, usual Github outages tend to be through Actions/the site, not the actual git operations. Doesn't seem like you can use a local copy but you can use Gitbucket/Gitlab.


I have fairly high local guide status. Do you think multiple submissions make a difference? Why?


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

Search: