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

Instead of having distinct tokens, you could have modifier vectors which would be added to other tokens. Think in terms of control, shift, meta etc.


Hey, Jasmine here (one of the authors) -- that's an interesting idea!

There's an interesting exploration of this here: https://www.lesswrong.com/posts/HEzNZ9gvgYwT3aZFS/role-embed....

Curious if you have additional thoughts, and thanks for reading!


12-bit log is a very common format in professional post-production.


<prefix>::1, <prefix>::2, <prefix>::3, <prefix>::4... exactly how hard can that be?

Or if you're feeling playful, <prefix>::b0d, <prefix>::bed, <prefix>::dad, <prefix>::b1d...


The prefix is too long and it's not the same at every site.

When I'm at a different location, the biggest problem is usually figuring out if they use 192.168.0 or 192.168.1 :)


lucky you, IPV6 indeed uses the same prefix everywhere: fe80::


Nitpick: The subnet fe80/10 is for link-local addresses (https://en.wikipedia.org/wiki/Link-local_address#IPv6). However, it is not an analog to the RFC 1918 IPv4 subnetworks.

192.168/16 is a private IPv4 subnet (https://en.wikipedia.org/wiki/Private_network#IPv4) and the equivalent for IPv6 is the fd00/8 subnet (RFC 4193).


And with ULAs (fd00::/8) you can pick your own prefix. Officially you should pick a random one to prevent collisions when you connect private networks, but you can choose something memorable if you don’t care about that


A changeover to your IPv5 would be just as agonizing as the changeover to IPv6. A system with a larger address space is fundamentally uninteroperable with one with a smaller address space as there is nowhere to put the extra bits in the old protocol. The lack of motivation to move to the new protocol would also be just the same.

And as for memorization: do you actually memorize MAC addresses for your interfaces? The answer is no, you don't, becase ARP handles all that for you. Well, for IPv6, DNS, mDNS and so on handles all that for both your IPv4 and your IPv6 addresses - or should, if you know what you are doing, as memorizing IPs doesn't really scale beyond a few dozen machines.

Yes, IPv6 is overengineered, but it gets the pain of having larger addresses in the packet done once and for all - the odds of needing more than 128 bits in the rest of human history are very small indeed. And if something radically new needs to replace the current IPv6 architecture, which is much more likely, the extra address bits are already there; only 2000::/3 is assigned for public use so far, and the new addresses would fit in the current IPv6 packet format already.


One big advantage of IPv6 local addresses is that you can pack a lot of semantic information in an address that's easy to remember, plus bits to help with routing and/or firewalling if you need.

DNS and mDNS don't "just work". You don't need but probably really want HA for DNS which is overkill for a homelab user, and you really want a fixed address for that DNS, because who wants to fix issues when you can't even address your services, and you really want your routers to have fixed addresses for the same reasons; you need VLAN and/or Avahi reflecting for mDNS, and if you need firewalling on your LAN, have fun dealing with the fact that mDNS clients prefer GUAs, then IPv4s, then ULAs in that order, by RFC rule, and managing GUAs sensibly when your ISP keeps changing your prefix -- well, IPv6 is almost 30 years old and home/SMB equipment still can't handle that reliably or flexibly, if it even lets you do anything besides assign /64s, and there's nothing stopping your ISP from saying "here just have a single /64, sorry if you wanted to actually use IPv6 for anything clever like having multiple subnets, who would ever want that?" So you say "I'll just use DHCPv6" and it turns out that DHCPv6 kind of sucks and it also turns out many devices don't support that by default or at all, including every single Android and Chrome device, for starters.

IPv6 is full of these design issues where you have a lot of things that are supposed to Just Work, Look It's So Much Simpler Than IPv4, and look at all these address bytes (excuse us while we take 64 of them away for no reason), except you discover that nothing Just Works with anything else in mildly nontrivial cases. You end up on a yak shave only to discover no yak underneath, and you end up just having a broken network while standing in a pile of yak hair. The whole story above is just one example. IPv6 is a migraine in RFC form, and if it weren't that I accidentally bought some expensive IOT devices that are IPv6-only, I'd be happy to never touch it. At this point, it would have been a better time-money tradeoff to have thrown those in the trash as soon as I had seen the problem.


Yeah but that ridiculous overdimensioning is something I object to. There's more IPs than is needed to give each grain of sand on this planet its whole IPv4-sized internet. That's just overkill.

And the problem seems to be solving itself as the world is turning its back on globalism. China and North Korea already have separated themselves. Iran too. China still uses the same address space but it's not like there's open connectivity with the rest of the world. We'll probably cut off Russia at some point completely as part of some sanction (they've been preparing for that for years), and Europe will break with America if things continue. We'll just have interoperability at a few controlled border points then, like China already does with its great firewall. It'll be easy to do some address translation then.

Ps that's not something I'm necessarily happy about but I do see this trend emerging of every region trying to wall itself off.


> Yeah but that ridiculous overdimensioning is something I object to. There's more IPs than is needed to give each grain of sand on this planet its whole IPv4-sized internet. That's just overkill.

People also thought that 4 byte wide IPv4 Adresses would be large enough. It's really hard to estimate how much you will need. And because numbers are effectively a free resource, it is better to overestimate.

IPv6 also gives you shortcuts to write addresses. You can abbreviate the longest run of zeroes with `::` and leading zeroes within a hextet can be omitted. This makes IPv6 address notation elastic.


Well considering the 4 bytes were designed in the time of a small research network between defense and universities, it's great foresight that they made it as big as it is. It still runs most of the internet to this day.

But it's not free, after all every packet carries this burden. I know about the annotation but it also makes it very difficult to parse.


That's basically no burden at all. If we cut the address length down to increase throughput, we would get a one-time increase of about 0.8% -- but consider how much faster Internet connections have gotten over the past 30 years. They've improved by about 0.8% per week on average. You're worrying about something that's smaller than one or two weeks of natural progress in Internet connection speeds

Rather than trying to minmax the address length, it makes more sense to sacrifice a few bits per packet to the addresses, wait a week or two on average to get the lost throughput back, and then spend those bits elsewhere to make other things easier. (For example, avoiding NAT is an obvious one, but even just "everything is a /64" removes the need to ever need to think about the size of a network.)

If you can spend a few address bits to make something easier elsewhere, that's a good trade. Maybe start worrying if the addresses were a kilobit long or something, but they aren't even close to that.


What ridiculous overdimensioning? L2 addresses are 64 bits, and L3 has to be bigger than L2 because L3 acts as an aggregation layer over L3. From staring at RFC 3194, I'd say the minimum size for L3 is about 80 bits.

v6 is 128 bits, which is the smallest power of 2 that's bigger than 80 and is only an extra 48 bits. Is that really enough to qualify for a claim of "ridiculous overdimensioning"? Especially when we really, really don't want to discover down the line that we made it too small and now need to migrate to another L3 protocol?

I get that the resulting number of IPs is big, but... so? Everything deals with the 128-bit long addresses, not the 2^128-entry long list of IPs. There's no need to care about the latter, just like there's no need to care about the number of potential RSA2048 keys or whatever.


If ipv5 worked just like ipv4 except with a larger address space, it would be easier than moving to ipv6. I shouldn't have to change my address to switch for example.


The slow adoption of IPv6 by many older companies is at least in part a paradoxical result of the success of IPv6 elsewhere, particularly in new builds where there is practically no overhead in deploying IPv6 in a green-field environment - see for example the mobile telecoms market in developing countries, where new builds are IPv6-first.

This has taken pressure off the IPv4 legacy address pool, reducing the urgency for older providers.

End-users are typically completely unaware of whether their traffic is being carried over IPv6 or IPv4, and so simply do not care one way or the other. (This particular post is more than likely being made over IPv6, since news.ycombinator.com has an IPv6 address and my OS, browser, router and ISP all support IPv6 straight out of the box, as is now true for the majority of users in my country.)


"Yes the planet got destroyed. But for a beautiful moment in time we created a lot of value for shareholders." [0]

The mindset behind private equity takeovers is exactly this. They know, but they are really well paid to do it, and they don't care.

[0] https://x.com/Benioff/status/549339156854214656


In-person written exams and classwork are really the only solution for this, either handwritten or on dumb devices locked down to only being a word processor.


This is a fascinating study that shows the massive disconnect between our moral choices and our actual hehavior, showing how many people are willing to accept the risk of killing children in exchange for a car that makes them feel good, even if they have been explicitly informed of that risk.


To be fair, the vast, vast majority of drivers will go their entire lives without hitting a pedestrian, let alone at speeds that are fatal. Where I live, pedestrians deaths seem to always be from drunk drivers, the very young or very old, and people walking across 20 lanes of interstate at night for some reason.


It does not necessary shows disconnect. Moral choices and behavior of those people are aligned they just have different values then you assume.


I would expect most would cite a thought process more akin to "I'm a safe driver, have never hit a pedestrian or cyclist before, and don't expect to in the future, either." for why they aren't swayed by this warning.

I expect very few went "I don't care about killing people with my car", for their reasoning.

Now, people are a terrible judge of their own talents at driving, and so many of those people are much less safe drivers than they think. But the point here is that most of those people believe they specifically are not at risk of causing that terrible outcome. It's faulty risk assessment rather than total lack of caring about others.


lol, check the comment above yours. It’s exactly the line of reasoning they cite.


I just took a look at the page's own 'town square'. At the moment, it's largely full of people saying offensive things for the sake of it.


After having been a moderator for a number of communities over the years—BBCs in the 90s, niche forums during the 00s & 10s, Reddit boards in the 20s—I’ve come to believe that running shadow services where disruptive individuals can comingle may be the best way to handle this.

The advent of LLMs really opens the door to shunting off these “community members “ who’d rather contribute in misanthropic ways for the lulz than either leave or not contribute at all. They can take part in an interactive echo chamber that gives just as well as they can. You don’t even need a powerful model so the overall costs to the community are probably lower than the alternatives of trying to coexist with community-arsonists.

I spent years trying to find ways to bring people productively “into the fold” but eventually realized that it is futile in some cases because there’s zero value to the individual or the community to find a middle ground. They want to see things burn, and the community simply wants them out.


But why not kick them out? Why bother with a shadow site for them?


If they don't know they're being quarantined, there's no reason for them to find a way around their block. They'll eventually tucker themselves out and move on to whatever their internet equivalent of burning ants with a magnifying glass is.


I used to run a small MMORPG. I only had to ban a player once but when I did they made it their life's mission to get around it and spam profanities. They went through every proxy in existence and I had to resort to banning whole IP ranges to keep them out.


The same reason people get shadowbanned. If they know they are banned they will attempt to work around it.


One day I'll find out that I've been shadow banned from HN, and all of you were bots all along, lolz


That defines planet earth nicely. Everybody here is shadow banned and has issues to sort out.


If they believe they’re still trolling people they will keep doing so and won’t make an alternative account and will keep harassing bots instead of real people.


This suggestion is scarier than the thought someone might say a slur.

You're talking about using a machine to detect social undesirables then quarantining them in the matrix.


A slur is rarely the problem. Trolls can (sometimes) add flavor to a community and keep things interesting. Mostly they can be an annoyance.

I’m specifically referring to people who have seemingly made it their sole purpose to create as much indiscriminate damage as possible.

You can ban them, block routes for them to attempt to Sybil themselves back to having accounts, etc. but even with great moderation tools and systems, it’s extremely difficult to set up a strong enough set of controls which don’t adversely impact everyone else who you want to have participate in the community.

Yes, a shadow environment is dystopian. It’s not my nature to want to even consider using one.

But we’re talking about privately run communities which also deserve to exist to serve their purposes.

So given the choice between anarchy which drives away people who contribute to make the community what it is and a shadow option for those actively working against its interests, I’ll consider the community first.

You may have misinterpreted my comment. I’m not suggesting you use LLMs as moderators. I’m talking about using LLMs as participant “members” of this shadow board to interact with someone whose account was flagged by a human moderator.


If someone is able to get around a ton of banning techniques, they will definitely notice they're shadowbanned.


In some cases, I’m sure you are right. That said, the online communities which currently employ shadowban systems continue to use them. That tells me there’s a value greater than their implementation and operation costs.

There is no perfect tech solution to a human problem. But in my opinion, having access to a partial mitigation is better than no mitigation.


There’s a difference between someone who is occasionally abrasive and mentally ill people want to destroy things other people enjoy.


If it's your forum you can do whatever you want with it.


Edit: Welcome to Hacker News btw! https://news.ycombinator.com/item?id=48618488 looks like a great project and I put it in the second-chance pool (explained at https://news.ycombinator.com/item?id=26998309).

---

Can you please edit out swipes like "Do you hear yourself?". This is in the site guidelines: https://news.ycombinator.com/newsguidelines.html.

If you wouldn't mind reviewing the other guidelines too, we'd appreciate it. Note this one, for example:

"Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith."


this comment immediately replying to "You're talking about using a machine to detect social undesirables then quarantining them in the matrix." raises interesting questions about the nature and purpose of dangbot


I would be happy to edit but I don't think I yet have the ability. Cheers.


Normally I'd say not to bother, but since your GP post got some interesting replies, I think it's a good idea so I've re-opened it for editing.

If you do that, let me know and I'll unflag the post and collapse the moderation bits ;)


Appreciated! It's edited.


Ok, all done. Thanks.


Yes this is the way, give no affordance that they are banned


Within about 300ms I thought “moderation is going to be a problem”.


Why? Messages disappear after a few seconds. So all messages are self moderating!


AI can moderate given specific instructions to prevent anything that can damaged the brand's image.


Yeah, I had exactly the same thought.


Maybe a really lightweight fast LLM could moderate messages in realtime. Not sure how pricey that would get though.


No matter how cheap per request, someone will figure out a way to DoS that endpoint, and it will be extremely pricy unless you have effective rate limiting.


The OpenAI moderation endpoint is free and it doesn't count toward the usage limit.


Extremely if not behind substantially cheaper anti spam measures.

Further it may still make sense to use human reports to gate some automation even if it slows response.


So it’s a twitter plugin?


This sort of boring grind is where the real work of systems engineering is done. Big infrastructure projects like this work on making the Linux kernel more reliable while still keeping it workable throughout the process move on the scale of decades, not months.


On one hand I understand why it’s decade scale (the long tail of users/dependencies is really, really long) but on the other hand it doesn’t feel like a tenable pace at which we can make meaningful long term progress. Less of a gripe and I guess more a paradox of critical infrastructure.


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

Search: