for me claude-code's agentic loop tooling + plan mode was th...

jb55

npub1xtscya34g58tk0z605fvr788k263gsu6cy9x0mhnm87echrgufzsevkk5s

hex

81d0445614e8b5fd1c1205a1741a417bc37a75603af33e0327829b5404d79e56

nevent

nevent1qqsgr5zy2c2w3d0arsfqtgt5rfqhhsm6w4sr4ue7qvnc9x65qnteu4sprpmhxue69uhhyetvv9ujuem4d36kwatvw5hx6mm9qgsr9cvzwc652r4m83d86ykplrnm9dg5gwdvzzn8ameanlvut35wy3gmvl9gl

Kind-1 (TextNote)

2026-01-28T20:37:02Z

↳ 回复 Leo Wandersleb (npub1gm7tuvr9atc6u7q3gevjfeyfyvmrlul4y67k7u7hcxztz67ceexs078rf6)

Guess I should try out more tools, not only many LLMs. I used cline with opus4.5 but it runs into context limits fast. So Sonnet it is mostly.

for me claude-code's agentic loop tooling + plan mode was the game changer. at this point its now just about scaling up the agent parallelism while remaining in the loop

原始 JSON

{
  "kind": 1,
  "id": "81d0445614e8b5fd1c1205a1741a417bc37a75603af33e0327829b5404d79e56",
  "pubkey": "32e1827635450ebb3c5a7d12c1f8e7b2b514439ac10a67eef3d9fd9c5c68e245",
  "created_at": 1769632622,
  "tags": [
    [
      "client",
      "Damus Notedeck"
    ],
    [
      "e",
      "ad71510e6b23f0cb72ec0a3f7a1c792594811a7c6f8ec07772bab164ec8189b9",
      "",
      "root"
    ],
    [
      "e",
      "9304585f1eb650c8dce51583e5d4a0d84bd4ba0e84da3b4418a45bc3eea8a109",
      "",
      "reply"
    ],
    [
      "p",
      "46fcbe3065eaf1ae7811465924e48923363ff3f526bd6f73d7c184b16bd8ce4d"
    ],
    [
      "p",
      "32e1827635450ebb3c5a7d12c1f8e7b2b514439ac10a67eef3d9fd9c5c68e245"
    ]
  ],
  "content": "for me claude-code's agentic loop tooling + plan mode was the game changer. at this point its now just about scaling up the agent parallelism while remaining in the loop",
  "sig": "7a92236ce683dc0631091d9c592232eea4ef641fc5c9f771050e57f11ad3f22c460ee0c238a045ce4456c534499987e0063edec50ee7f6650a28a3254c559879"
}