3 ms·
The problem in most of those cases is not specifically AI. Many of the issues you cited are related to Anthropic specifically and many could have been avoided w
by michaelmior 5mo ago
The problem in most of those cases is not specifically AI. Many of the issues you cited are related to Anthropic specifically and many could have been avoided with better testing.
- pluc 5mo agoYes, I am assuming the AI/LLM of choice you've implemented in your software engineering org is Claude because as far as I can tell there aren't really alternatives that come close to its quality in software.
- michaelmior 5mo agoWe have some very heavy users of Codex in my org and we're very happy with the quality (politics aside).
- pluc 5mo agoSo your thinking in light of this information is just "don't use Claude"?
- michaelmior 5mo agoI don't think it's that simple, but I do think a lot of the problems mentioned are not inherent to the use of AI.