dabaja

3 ms·

1. 1.▲
   by dabaja 8mo ago

   Honestly, it's empirical. We started with what was easiest to measure: human correction rate. If I had to step in and fix something, that's a clear signal the agent took a bad path. Iterations and reverts turned out to be noisier
2. 2.▲
   Ask HN: Anyone using the Claude Excel extension/add-in in anger?

   1 points by dabaja 8mo ago|0 comments
3. 3.▲
   by dabaja 8mo ago

   We hit something similar building document processing agents. What helped was treating memory as typed knowledge rather than flat text. Three buckets: (1) constraints -- hard rules that should always apply, (2) decisions -- past choices wit
4. 4.▲
   by dabaja 8mo ago

   Interesting framing, but I think emotions are a proxy for something more tractable: loss functions over time. Engineers remember bad hygiene because they've felt the cost. You can approximate this for agents by logging friction: how ma
5. 5.▲
   by dabaja 2y ago

   Why am I not surprised? We’re starting 2025 just as we ended 2024—with masks dropping everywhere. From Big Oil slashing investments in green energy, to U.S. banks walking back on sustainability-linked loans, to the EU's push for an omn