Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Comparing Fable and 10 other LLMs on refactoring a LangGraph god node

wtf.korridzy.com

Post reply on HN
Print view View original

11–20 of 22 posts

Re: Comparing Fable and 10 other LLMs on refactoring a LangGraph god node

#12

Post by overgard »

I had a pretty rough experience with Fable today. At first, Claude showed a message that Fable was available again, but if I typed /model it showed the same "disabled" state for the export controls, so that was sloppy. After I got it working (I had to login again), I asked it to do a security audit of a server I've written, and it spent a few tokens and then "flagged" the request and downgraded to Opus 4.8. Opus 4.8 just kind of glazed me for writing secure code without finding anything important.

By the way, I ran ccusage afterwards and if I had been paying API rates I would have apparently spent $6.88 for that.

Amazing stuff Anthropic.

Re: Comparing Fable and 10 other LLMs on refactoring a LangGraph god node

#14

Post by clickclas »

overgard wrote: post #12

I had a pretty rough experience with Fable today. At first, Claude showed a message that Fable was available again, but if I typed /model it showed the same "disabled" state for the export controls, so that was sloppy. After I got it working (I had to login again), I asked it to do a security audit of a server I've written, and it spent a few tokens and then "flagged" the request and downgraded to Opus 4.8. Opus 4.8…

https://www.anthropic.com/news/redeploying-fable-5

The US government has forced them to make the filters even stricter than they already were. It feels meaningfully worse than during its initial release.

Amazing stuff, Trump voters.

And amazing stuff to Amazon's CEO, who threw a tantrum on the basis of misinformation and created this whole mess and has how resulting in permanent government handicapping of American models. Genius move.

Re: Comparing Fable and 10 other LLMs on refactoring a LangGraph god node

#15

Post by threecheese »

Korridzy wrote: post #8
llm_nerd wrote: post #5

Earlier quoted context omitted.

It is blocked by Cloudflare, and it looks like the reason is simply that it's a super new domain so it's just an overly cautious thing.

A pity. I've really made this blog a couple days ago. This is my first post in it :)

Indeed. I read the entire article, it was great. Domains can be a big PITA.

Re: Comparing Fable and 10 other LLMs on refactoring a LangGraph god node

#16

Post by reckless »

Takeaways Which model to use for generating architecture.

The simpler the decision you need to make, the more readily you can just take a proposal from Fable or GPT-5.5. And now that Fable is unavailable outside the US, the choice between Opus and GPT is far from obvious. As a quick default I'd lean toward GPT.

Re: Comparing Fable and 10 other LLMs on refactoring a LangGraph god node

#19

Post by whythismatters »

overgard wrote: post #12

I had a pretty rough experience with Fable today. At first, Claude showed a message that Fable was available again, but if I typed /model it showed the same "disabled" state for the export controls, so that was sloppy. After I got it working (I had to login again), I asked it to do a security audit of a server I've written, and it spent a few tokens and then "flagged" the request and downgraded to Opus 4.8. Opus 4.8…

Same story here. Wanted to give Fable a spin, session got flagged mid-execution of the plan (some innocent react-native code), things fell back to Opus, and then my "current session" usage limit ran out. So I ended up with a partial and broken Fopus implementation.

Re: Comparing Fable and 10 other LLMs on refactoring a LangGraph god node

#20

Post by overgard »

clickclas wrote: post #14
overgard wrote: post #12

I had a pretty rough experience with Fable today. At first, Claude showed a message that Fable was available again, but if I typed /model it showed the same "disabled" state for the export controls, so that was sloppy. After I got it working (I had to login again), I asked it to do a security audit of a server I've written, and it spent a few tokens and then "flagged" the request and downgraded to Opus 4.8. Opus 4.8…

https://www.anthropic.com/news/redeploying-fable-5 The US government has forced them to make the filters even stricter than they already were. It feels meaningfully worse than during its initial release. Amazing stuff, Trump voters. And amazing stuff to Amazon's CEO, who threw a tantrum on the basis of misinformation and created this whole mess and has how resulting in permanent government handicapping of American mo…

I don't like any of those people so I'm not disagreeing, but I feel like this is 100% on anthropic for their constant doom trolling. They're not trying to be "responsible", they just want to scare people and it worked.
Post reply on HN