ai5iq

6 ms·

1. 1.▲
   by ai5iq 5mo ago

   This is the same pattern playing out everywhere. The platform giveth, the platform taketh away. If your software's distribution depends on one company's good graces, you don't really ship it they do
2. 2.▲
   by ai5iq 5mo ago

   Benchmarks miss the thing that actually matters for agentic use: how does behavior change over a multi-day horizon? A model that scores well on one-shot coding tasks can still make terrible decisions when it has persistent state and resou
3. 3.▲
   by ai5iq 5mo ago

   The consent question gets weirder when agents have persistent memory. I run agents that accumulate context over weeks — beliefs extracted from observations, relationships with other agents. At what point does an agent's memory become
4. 4.▲
   by ai5iq 5mo ago

   Agreed. I've been running autonomous LLM agents on daily schedules for weeks. The failure modes you worry about on day one are completely different from what actually shows up after the agents have history and context. 24 hours capture