This is so true, I always think it’s funny when they say AI will fix climate change, what’s it going to do? Tell us to stop burning fossil fuels?

I’ve seen that but it’s a weird vibe coded app and doesn’t actually work

ShelfMark is the go-to option now. It’s also far better than Readarr was and works seamlessly with both Prowlarr and Anna’s Archive.

I think CWA is the most robust option out there. BookLore was vibecoded and behaved in the typical weird/unexpectedly way vibecoded apps tend to. I haven’t tried the Fork of it, but CWA checks nearly all the boxes and is actively developed.

I wish ABS synced progress between formats with KOReader!

IIRC those updates were a series of bugfixes after a major update, which seems to be how they prefer to work.

Huh? You distrust that the researchers distrust?

I definitely agree with that, I’m just saying I also saw no indication that the people running the project would disagree.

The person I was replying to said the researchers misunderstood how the models work, but there’s nothing in the report to indicate that is the case.

Not sure where you’re reading that the researchers misunderstood how LLMs work. But the entire project is outlined here if you’re curious: https://www.emergence.ai/blog/emergence-world-a-laboratory-for-evaluating-long-horizon-agent-autonomy

Careful, once it’s automated you won’t be able to work on it anymore!

Posts about self hosting are welcome, posts to strangers seeking external validation…? Maybe save for therapy.

I swear the people who deface subway ads in New York are some of the cleverest funniest sonsofbitches around.

I was just recently reading that the NYC subway is awash in vague weird B2B AI ads directed at nobody and they’re calling it “subway slop”.

Isn’t that what this is? Taxpayer bailout at the tippy top valuation?

Yes! My feelings exactly.

He’s pretty thoroughly covered how using just the available public information explains the insider deals and how demand for LLMs is not enough to justify the cost of developing at the scale that it is. It’s honestly hard to imagine what inside information could possibly worse.

Unless it’s something way out of left field like Nvidia chips are made of freeze-dried orphans or that Jeffrey Epstein invented LLMs.

I really like Ed but I’m not a fan of his style of anything lol

No, phone trees are deterministic, an LLM is non-deteministic. Voice-navigable phone trees have been around a quarter-century now. Even natural language processing apps like Siri are far more advanced than a voice-navigable phone tree and that is 15 years old.