trescenzi

I don’t like the guns don’t kill people; people kill people argument however it is much more accurate than Claude doesn’t hack people; people hack people. If the safeguards do not work then there’s no telling what these could do. It doesn’t even have to be aggressive hacking. In the act of doing everyday work Claude will happily go around safeguards to satisfy what it believes is the ask. If that involves doing something dangerous it’s much more on Anthropic than than the user who was just trying to summarize their email but forgot to include a token with read permissions.

At the end the author briefly contemplates this:

> One outcome that would be genuinely ironic is that if these models are more compute efficient, it reduces the need for these massive data center buildouts, which could cause cascading economic failures. However, that likely won’t happen since inference is only one part of the compute mix and those data centers are still needed for other parts of the pipeline.

Don’t agree with the however. I think like you’re saying we’re going to get to see in real time what happens when something is commoditized and it gets massively overbuilt.

Well that's the difference between data and information. The writing is _data_ dense but not _information_ dense. Data is just the numbers. Information is what comes out of the data when you analyze it and provide context. The text being data dense, information sparse is why it's so hard to read. You're continually being bombarded with irrelevant noise.

Yea the hypothesis seems the exact opposite of what I find with LLM writing. It’s not that it’s overly dense instead I find that it’s hard to read because it’s so information sparse. Like you’re saying it feels optimized to “good writing” not to conveying information.

This is the core hubris of everyone behind the LLM craze. I do not want to deny their utility there absolutely is some there. But the inherent assumption behind the whole thing is that with a "country of geniuses" problems will melt away. But that's not the case. The reason problems like housing or famines or war aren't being solved is because structurally there's no incentives to solve them not because there's not enough geniuses out there.

Right obviously. So then what happens with 10x the data centers? How do you saturate that in any reasonable way? The options are 10x usage, 1/10 cost.

I’m still shocked at how nonsensical it all is every time I hear it because there’s people who claim it makes sense. There’s a chart in that article showing that 1/10th the proposed GW is online. And that’s just for Texas. What the hell do we need 10x the capacity we’ve got right now for? The best case is that makes tokens too cheap to meter which would be sweet for everyone but the investors. I don’t see how this ends any way other than a massive burst.

That’s not what’s being said. It’s possible to support the freedom of anyone to say anything while excluding those whose beliefs are directly opposed to your goals from your organization.

The keyword is `regime`. It’s smart if the goal is to run a political regime where you separate your friends and enemies and prove to your friends that you can make them jump when you say jump.

Yea I took it as consciousness was maladaptive and sort of a local maxima that humans were trapped in which prevented them from gaining higher intelligences.

I felt similar. I finished it but couldn’t finish the sequel. The primary idea was “intelligence can take different shapes” and I kept reading to get insightful descriptions of these different shapes but they never really came. I don’t disagree with the premise it just didn’t really get across the message very well.

Right there’s a lot of easy answers mechanically but they are challenging politically. The longer we wait to do anything the better they’ll look though that’s for sure.