Re:Show HN: JavaFactory – IntelliJ plugin to generate Java code
by winrid on Tuesday May 20, 2025 @11:35AM
(#44042795)
I've always wondered how long until we reach this. If every pc can run models locally, with a given seed and prompt it could be the ultimate compression. It's also hilarious.
[ Parent
| on HN ]
Re:Show HN: JavaFactory – IntelliJ plugin to generate Java code
by imhoguy on Tuesday May 20, 2025 @11:46AM
(#44042922)
Although very lossy compression, each invocation will be different, so that will inevitably circle back to "strong-static-LLM" prompts. What? wait..!
Re:Show HN: JavaFactory – IntelliJ plugin to generate Java code
by woodrowbarlow on Tuesday May 20, 2025 @12:59PM
(#44043642)
LLMs at their core do produce reproducible results with a given seed. it's all the workflow stuff people do on top that tends to break reproducibility.
Re:Show HN: JavaFactory – IntelliJ plugin to generate Java code
by dibujaron on Tuesday May 20, 2025 @06:00PM
(#44046359)
This is not the case for LLMs running on GPUs (which is most of them); GPUs are non-deterministic for this use-case due to the floating point math involved. there is no way to get perfectly deterministic output from OpenAI despite the presence of seed and temperature parameters.