Just leave me behind

 I have been working on this algorithm for ordering a word list in terms of similarity between Korean words. Korean uses its own alphabet so I made a specific algorithm tailored to how the alphabet looks visually and how the letters sound. An example of two words that seem similar to me is 공항 (airport) and  공원 (park). They have the same length visually and both begin with 공. Sound wise the bottom part of the last block in airport "ㅇ" sounds like "ng" and the bottom part of park "ㄴ" sounds like "n". So they both look and sound similar. I wanted an algorithm to capture this similarity.  Usually when I do tests I do the task myself first, and then I ask the LLM to do it as well. So I looked at things like the length of the words, similar start and finish, and if there were identical blocks like 공 and I made a list of which letters could be confused with each other.  Then I asked several LLMs to make an algorithm that did this and they all only sugge...

 This blog is about my tests of several LLM models and how exactly they fail. I have been working as an algorithm designer for about 15 years and have been trained in seeing the flaws AI/automation systems make. So this is the skill I will be using in writing this blog. I have and will run test conversations on the free models of Claude, ChatGPT, Gemini and CoPilot. I have run no tests of frontier models, but the errors the models made in the earlier versions are still there in the later ones and across different vendors. So, to my eyes, the flaws seem to be inherent to the LLM architecture.  I appreciate all the hard work from people like Ed Zitron in showing that the financing of the AI industry does not add up. Also, Cal Newport and others talking about what effect AIs have on people using them and also about how the tech works and doesn't work in general. And Nik Suresh in talking about how implementing it adds very little value in practice.  I have deep respect for t...