4 ms·
lol 5000 tests. Agentic code tools have a significant bias to add versus remove/condense. This leads to a lot of bloat and orphaned code. Definitely something t
by maddmann 10mo ago
lol 5000 tests. Agentic code tools have a significant bias to add versus remove/condense. This leads to a lot of bloat and orphaned code. Definitely something that still needs to be solved for by agentic tools.
- oofbey 10mo agoOh I’ve had agents remove tests plenty of times. Or cripple the tests so they pass but are useless - more common and harder to prompt against.
- maddmann 10mo agoAh true, that also can happen — in aggregate I think models will tend to expand codebases versus contract. Though, this is anecdotal and probably is something ai labs and coding agent companies are looking at now.
- oofbey 10mo agoIt’s the same bias for action which makes them code up a change when you genuinely are just asking a question about something. They really want to write code.
- nosianu 10mo ago> Agentic code tools have a significant bias to add versus remove/condense. Your point stands uncontested by me, but I just wanted to mention that humans have that bias too. Random link (has the Nature study link): https://blog.benchsci.com/this-newly-proven-human-bias-causes-you-much-unnecessary-work https://blog.benchsci.com/this-newly-proven-human-bias-cause... https://en.wikipedia.org/wiki/Additive_bias https://en.wikipedia.org/wiki/Additive_bias
- maddmann 10mo agoGreat point, interesting how agents somehow pick up the same bias.