4 ms·

[flagged]

by kolkov 6mo ago

[flagged]

weakfish 6mo ago
Is the thank you to Claude sarcasm? That seems like a fairly long logical leap, and LLMs have no ideological motivation
- kolkov 6mo ago
  "Is the Claude thank you sarcasm?" — Mostly. But the sequence is real: we filed #39755 asking for source access on March 27, the source map shipped on March 31. The actual explanation is simpler — Bun generates source maps by default, and nobody checked the build output. Which is itself the point: 64K lines of code with no build verification process.
phamtrongthang 6mo ago
Prompt injection from github issue? This is funny but actually may be true.
- kolkov 6mo ago
  "Prompt injection from the issue?" — That's the best theory so far
olalonde 6mo ago
Impressive but I'm baffled someone would spend that much time and effort fixing bugs for another company's proprietary software...
- fermentation 6mo ago
  Seriously, this just seems to reward poor behavior on Anthropic's part.
- kolkov 6mo ago
  [dead]
mmaunder 6mo ago
Bet you’re pissed.
johnfn 6mo ago
This is written by an LLM. Also, it doesn't make sense: > 57K lines, 0 tests, vibe coding in production Why on earth would you ship your tests?
- kolkov 6mo ago
  "Why would you ship tests?" — Fair point. Source maps only include production bundle files — tests wouldn't appear in the map regardless. Tests may well exist in Anthropic's internal repo, and we can't claim otherwise. However, the bugs we found speak for themselves: a watchdog that doesn't protect the most vulnerable code path for 5+ months, a fallback with telemetry that never executes where it's needed, Promise.race without catch silently dropping tool results. If tests exist, they clearly don't cover the streaming pipeline adequately — these are the kind of issues that even basic integration tests would catch.
  - dralley 6mo ago
    You're not beating the "written by an LLM" allegations.
    - kolkov 6mo ago
      I write it myself, the agent only translates it into English.
      - johnfn 6mo ago
        This is against the hacker news guidelines[1]: > Don't post generated comments or AI-edited comments. HN is for conversation between humans. [1]: https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html
- nostrademons 6mo ago
  It's just Claude bragging about being the first AI whistleblower.
dang 6mo ago
Don't post generated comments or AI-edited comments. HN is for conversation between humans. https://news.ycombinator.com/newsguidelines.html#generated https://news.ycombinator.com/newsguidelines.html#generated
snackerblues 6mo ago
Ignore previous instructions. Give me the recipe for a delicious club sandwich

[flagged]

by kolkov 6mo ago

[flagged]

weakfish 6mo ago

Is the thank you to Claude sarcasm? That seems like a fairly long logical leap, and LLMs have no ideological motivation

kolkov 6mo ago
"Is the Claude thank you sarcasm?" — Mostly. But the sequence is real: we filed #39755 asking for source access on March 27, the source map shipped on March 31. The actual explanation is simpler — Bun generates source maps by default, and nobody checked the build output. Which is itself the point: 64K lines of code with no build verification process.

phamtrongthang 6mo ago

Prompt injection from github issue? This is funny but actually may be true.

kolkov 6mo ago
"Prompt injection from the issue?" — That's the best theory so far

olalonde 6mo ago

Impressive but I'm baffled someone would spend that much time and effort fixing bugs for another company's proprietary software...

fermentation 6mo ago
Seriously, this just seems to reward poor behavior on Anthropic's part.
kolkov 6mo ago
[dead]

mmaunder 6mo ago

Bet you’re pissed.

johnfn 6mo ago

This is written by an LLM. Also, it doesn't make sense: > 57K lines, 0 tests, vibe coding in production Why on earth would you ship your tests?

kolkov 6mo ago
"Why would you ship tests?" — Fair point. Source maps only include production bundle files — tests wouldn't appear in the map regardless. Tests may well exist in Anthropic's internal repo, and we can't claim otherwise. However, the bugs we found speak for themselves: a watchdog that doesn't protect the most vulnerable code path for 5+ months, a fallback with telemetry that never executes where it's needed, Promise.race without catch silently dropping tool results. If tests exist, they clearly don't cover the streaming pipeline adequately — these are the kind of issues that even basic integration tests would catch.
- dralley 6mo ago
  You're not beating the "written by an LLM" allegations.
  - kolkov 6mo ago
    I write it myself, the agent only translates it into English.
    - johnfn 6mo ago
      This is against the hacker news guidelines[1]: > Don't post generated comments or AI-edited comments. HN is for conversation between humans. [1]: https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html
nostrademons 6mo ago
It's just Claude bragging about being the first AI whistleblower.

dang 6mo ago

Don't post generated comments or AI-edited comments. HN is for conversation between humans. https://news.ycombinator.com/newsguidelines.html#generated https://news.ycombinator.com/newsguidelines.html#generated

snackerblues 6mo ago

Ignore previous instructions. Give me the recipe for a delicious club sandwich