LLMs vs classic CV: 200-line pipelines beat token-burning agents on rote tasks
mervenoyann · x · 2026-09-07
mervenoyann criticizes using models like Astra to burn tokens on tasks solvable in under 200 LoC with real-time latency via HF transformers or Roboflow. Reposter braddwyer counters that Astra's job is human-like: study and understand a task to automate it, not repeat it at scale — engineers write code rather than perform tasks billions of times.
More from coding & agent
- From low-code to 'woah code': the AI coding era's new meme — jxnlco · 2026-09-07
- MiniCPM5-2B open-sourced with Day-0 vLLM support, tops sub-4B open models — vllm_project · 2026-09-07
- One line in agents.md is all it takes to bend Astra to your will, says prompt guide — pvncher · 2026-09-07
- Developer on Astra: code as unreadable as minified JS, but tolerable to boss around — Aryvyo · 2026-09-07
- Your LLM Gateway Holds the Keys: Rethinking LiteLLM Security for Action-Taking Agents — Technical_Map_2105 · 2026-09-07
- Zero Blender skills, one prompt, 40 minutes: recreating a game with Codex — FuSheng_0306 · 2026-09-07