3 ms·
Of course the team developing these AI models should be dogfooding them. Telling the rest of the company they have to use them is pushing tools without understa
by Arainach 1y ago
Of course the team developing these AI models should be dogfooding them. Telling the rest of the company they have to use them is pushing tools without understanding contexts.
If AI is that much more amazing, just measure employees on the usual metrics and those using AI should be so obviously far ahead that you can get rid of the others. Measuring "usage of AI" is a garbage metric that will not achieve anything good.
- PeterStuer 1y agoIt could be they need to force usage to generate more quality training data over the MS codebase. You'll get a huge amount of prompt/llm_answer/human_correction instances if all MS programmers are forced to produce these e en if initially it does not increase or even decreases their productivity. We saw the same in the early offshoring days. Firms forcing this on their project leads, knowingly expecting, accepting and swallowing the productivity losses to learn how to set this up for the promise of future cost savings.