Andrew Godwin on Nostr: My current prediction of where "vibe" (LLM-assisted) coding goes from here: - ...
My current prediction of where "vibe" (LLM-assisted) coding goes from here:
- Management says all PRs must be reviewed by senior engineers
- It's a massive time waste, so new metrics are imposed on original authors/LLM users over %age of accepted PRs
- Authors realise it is often quicker to write themselves than review all the code in detail for a mistake that hits their metrics
- LLMs end up being only used for smaller-scale tasks the author understands
Published at
2026-06-04 14:51:42 UTCEvent JSON
{
"id": "1798161f6d771d202d238b82410a2ef8c5a7ab90d535a0f5e047f1e72311eaf6",
"pubkey": "d3c537be60d55b1d776f75a337d4ec14693ea2d0f07474895d5a95d3679c50ac",
"created_at": 1780584702,
"kind": 1,
"tags": [
[
"proxy",
"https://fedi.aeracode.org/@andrew/116692399059926347",
"web"
],
[
"proxy",
"https://fedi.aeracode.org/users/andrew/statuses/116692399059926347",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://fedi.aeracode.org/users/andrew/statuses/116692399059926347",
"pink.momostr"
],
[
"-"
]
],
"content": "My current prediction of where \"vibe\" (LLM-assisted) coding goes from here:\n\n- Management says all PRs must be reviewed by senior engineers\n- It's a massive time waste, so new metrics are imposed on original authors/LLM users over %age of accepted PRs\n- Authors realise it is often quicker to write themselves than review all the code in detail for a mistake that hits their metrics\n- LLMs end up being only used for smaller-scale tasks the author understands",
"sig": "6d45d0d5efff3696a0f3b47c40405a25e65bbc7ec23b3ce2083e9080fdda57eff052ff26400daa5416a5cb17d701bba855f51aa0bbf21268d92b57a52cf9a111"
}
