fix(llm): switch to deepseek-chat (standard model, no reasoning overhead)

deepseek-v4-flash is still a reasoning variant and burns tokens on
chain-of-thought before producing content. deepseek-chat produces output
directly — 82 tokens all went to content in testing, vs 85-102 tokens
that produced empty observations on flash.

lib/towerops/llm/deepseek.ex

@@ -13,7 +13,7 @@ defmodule Towerops.LLM.DeepSeek do
   @behaviour Towerops.LLM.Behaviour
 
   @default_base_url "https://api.deepseek.com/v1"
-  @default_model "deepseek-v4-flash"
+  @default_model "deepseek-chat"
   @receive_timeout 120_000
 
   @impl true