lyricat

I started building this after one day i found claude's "well the models are technically still usable, but not really usable anymore" moments.

Their API still worked, but it could still work as good as before, especially in actual agent tasks, it felt noticeably worse.

Less careful. More guessing. Sometime it try to do something without looking plausible before you noticed it had barely read the documents or codes.

That was the part I wanted to measure. LLMSnare measures llm's behavior benchmark, not result quality. It includes read before write, reuse existing information, recover from errors, and keep that discipline across repeated runs.

A live arena is set up for me for watching the SOTA llms:

https://mistermorph.com/llmsnare/arena/

I also want to listen to your feedback, especially on case design and anti-gaming rules.

rather than listing advantages over other products to "trick" people into using it, it's better to clearly explain direction of iteration.

Great wizards often work alone, as long as the elements in the air still respond to their spells and calls

beancount.

I think it depends on the product field you are facing.

Personally, I will recommend Donella’s “Thinking In systems”

It's interesting that the behaviors began since 2013 and nobody recognized.

I used to enrol this course on Coursera. It called Learning How to Learn. It would be helpful for you. http://coursera.org/learn/learning-how-to-learn

I recall that there were lots of special-use Linux distributions which include specified packages.

Not true. Anaconda/miniconda also provide a convenient way for package control which include binary packages and source code packages

For me, condo is good solution to install necessary dependences with little pain.

Or phone stolen.