I think the Westphalia thing is somewhat overblown there were lots of sovereignty analogs throughout human history all over the world before that.

Why so? Do you think Monaco should be part of France? Do you think Singapore should be part of Malaysia? A lot of big countries respect the sovereignty of neighboring smaller countries, although that is unfortunately becoming less true now.

It isn't about colonialism. I have never seen anyone seriously argue it should go back to the British. It is about a framework to ensure they maintain their rights. It would be great if that looked like expanded rights for all of China but it can also look like some degree of sovereignty, which was in place for quite some time.

Comments URL: https://news.ycombinator.com/item?id=45958002

Points: 2

# Comments: 1

The manager example is a good case study. There are a lot of examples here where there might be genuine repercussions for raising an issue with a manager. I wouldn't give this as blanket advice.

Unfortunately, I don't think there's a simple rule about whether or not you should raise an issue and it needs to be decided case by case.

I am a "CTO" and I always put that in air quotes because I have one direct report and I spend the lion's share of my time doing IC work. I know what I do is not what people picture when they hear the title and I feel weird saying it. I use it because I do have to make the strategic technical decisions, there is no one else. When people are marketing technical B2B SaaS I am the one they are looking for.

From my perspective there just isn't nearly enough for me to do as a CTO to justify me not coding. If I were to hire someone just to manage them that would be an unjustifiable expense at this point. But I also get that as soon as we get to a reasonable size this would be totally unsustainable.

I am even more skeptical on a conceptual level. The LLM memories aren't constructing a self-consistent and up to date model of facts. They seem to remember snippets from your chats, but even a perfect AI may not be able to get enough context from your chats to make useful memories. Things you talk about may be unrelated or they get stale but you might not know which memories your answers are coming from but if you did have to manage that manually it would kind of defeat the purpose of memories in the first place.

This is only because these projects only became consumer facing fairly recently. There was a lot of incremental progress in the academic language model space leading up to this. It wasn't as sudden as this makes it sound.

The deeper issue is that this future-looking analysis goes no deeper than drawing a line connecting a few points. COVID is a really interesting comparison, because in epidemiology the exponential model comes from us understanding disease transmission. It is also not actually exponential, as the population becomes saturated the transmission rate slows (it is worth noting that unbounded exponential growth doesn't really seem to exist in nature). Drawing an exponential line like this doesn't really add anything interesting. When you do a regression you need to pick the model that best represents your system.

This is made even worse because this uses benchmarks and coming up with good benchmarks is actually an important part of the AI problem. AI is really good at improving things we can measure so it makes total sense that it will crush any benchmark we throw at it eventually, but there will always be some difference between benchmarks and reality. I would argue that as you are trying to benchmark more subtle things it becomes much harder to make a benchmark. This is just a conjecture on my end but if something like this is possible it means you need to rule it out when modeling AI progress.

There are also economic incentives to always declare percent increases in progress at a regular schedule.

Will AI ever get this advanced? Maybe, maybe even as fast as the author says, but this just isn't a compelling case for it.

Dependencies do suck but it is because managing a lot of complicated code sucks. You need some way to find issues over time and keep things up to date. Dependencies and package managers at least offer us a path to deal with problems. If you are managing your own dependencies, which I imagine would mean vendoring, then you aren't going to keep these dependencies up to date. You aren't going to find out about exploits in the dependencies and apply them.

You could easily say that the AI hype is a cope as well. The tech industry and investors need there to be be a hot new technology, their career depends on it. There might be some truth to the coping in either direction but I feel you should try to ignore that and engage with the content of whatever the person is saying or we'll never make any progress.

It's always interesting to see this take because my perception is the exact opposite. I don't think there's ever been an issue for me personally with a bigger mismatch in perceptions than AI. It sometimes feels like the various sides live in different realities.

As for practical problems they are extremely vague. What counts as harm? Could a robot serve me a burger and fries if that isn't good for my health? By the rules they actually can't even passively allow me to get harmed so should they stop me from eating one? They have to follow human orders but which human? What if orders conflict?