3 ms·
The tools can _answer_, but they tend to hallucinate functions and arguments that seem plausible but are wrong upon closer inspection. This output is an exampl
by al2o3cr 4y ago
The tools can _answer_, but they tend to hallucinate functions and arguments that seem plausible but are wrong upon closer inspection.
This output is an example: the generator made up an option for Macro.escape and completely misunderstood what Macro.to_string does:
https://elixirforum.com/t/chatgpt-generated-code-for-eval-string-alternative/52544 https://elixirforum.com/t/chatgpt-generated-code-for-eval-st...
- hrpnk 4y agoIt's true that not all code generated by ChatGPT will work out of the box. https://twitter.com/sergiorocks/status/1600489868936749059 https://twitter.com/sergiorocks/status/1600489868936749059 lists some examples of real-life interview questions. https://medium.com/@streetsofboston/did-chatgpt-ace-its-android-dev-job-interview-fd99262d21e3 https://medium.com/@streetsofboston/did-chatgpt-ace-its-andr... shows a full interview for an Android developer. Those two are more of the examples that call for re-evaluating approach to interviewing moving forward.