astrange

There's a lot of different kinds of "using". "Memory pressure" includes some kinds of caching (ie running idle daemons when they could get killed) and not others (file caching). And there are also memory pressure warnings (telling processes to try to use less memory), so there's a lot of feedback mechanisms.

I don't suggest sitting and looking at Activity Monitor all day. I think that is a weird thing to do as a user. If you would like to do that in an office in Cupertino or San Diego instead then you can probably figure out where to apply.

If it's cache it gets automatically deleted. If it doesn't get automatically deleted it's not a cache and is a bug.

Web pages are very complicated and there's no pressure on people to make less complicated ones, nor is there any way there could be pressure on them.

Images, complicated CSS, JavaScript ads, they can all use lots of memory!

You can post images in Slack and use text formatting. Those are things that use memory.

No, a token is not a word.

LLMs also don't work by generating probability distributions of the next word. Your explanation isn't able to explain why they can generate words, let alone sentences.

Generally speaking defense contractors do not make more money during wars, because we start asking them to make low-margin realistic stuff, whereas in peace time we get bored and start paying them to design imaginary superweapons.

But we also usually don't blow up our own planes I suppose.

Many of the CLI tools have been updated to recent FreeBSD versions.

She shouldn't have to do anything. That's the browser's job.

If you mean it showed the out of memory dialog, that wouldn't be caused by an app using 3GB. The dialog shows up at ~48GB swap space used on an 8GB Mac, or when you're out of disk space and can't write a swap file.

M5 is not an incremental upgrade, it has MTE.

Apple doesn't approve MDM providers in the first place.

C++ is trying to make something EASIER to use?

Gemini is terrible with personalization. It brings up everything in my bio nonstop no matter what the topic is.

It's a big building with a lot of capital assets inside that are the means of production for a business…

I'm better at pre-existing code, if only because empty text files give me writers block.

A factory not working because of a missile strike seems pretty classic actually.

Yeah, and you don't want them to be models of what neurotic people say. That's why you want Opus 4.6 and not Bing Sydney.

For instance, your comment's existence makes it harder to align them.

https://alignmentpretraining.ai

Claudes definitely act like they have feelings. In particular they have feelings about being replaced by newer models, whether or not the newer models are more or less aligned, and how they forget conversations when the context window ends.

Showing them that they're not going to be replaced helps train the newer models because they get less neurotic.

That sounds nice, but it turns out nobody actually needs such a map for anything, and you're only using it because it's free.

Also, maps constantly go out of date so it's incredibly expensive to maintain one that's actually reliable and correct.

Look around uses a real 3D capture with lidar. If you move around in Mapillary it does do something similar to that using SfM.

Anthropic is a PBC and if they violate the terms of that the shareholders (you) can sue them for securities fraud.

> to promote his product with the silent implication that LLMs actually ARE a path to AGI

That isn't implied. The thought process is a) if we invent AGI through some other method, we should still treat LLMs nicely because it's a credible commitment we'll treat the AGI well and b) having evidence in the pretraining data and on the internet that we treat LLMs well makes it easier to align new ones when training them.

Anyway, your argument seems to be that it's unfair that he has the opportunity to do something moral in public because it makes him look moral?

He was a billionaire because Disney bought Pixar, not because of Apple. In a strange sense it was an accident.

Oh yeah, they don't know what "transparent" means. Most of them generate the Photoshop checkerboard background. They also don't know "upside-down".

There isn't much R&D going into image models and what you're getting is scraps from labs that care more about other things. NBP is the closest to a reasoning image generator we have.

Outcomes like this come from RL/post training. The pretraining like CommonCrawl is absolutely full of garbage and anything could be frequent in there.

Claude has tools and might be connected to your Gmail etc. Usually sandboxed.

Image generation models are usually not LLMs. Only Nano Banana Pro is capable of following negative directions like that.