In the context of the discussion of *sshhole, it’s interesting to see yet another issue caused by the in-band signalling so common in C APIs, in this case using 0 as a sentinel rather than as 0.
"in-band signalling" -- thanks, exactly the phrase I've been looking for. I've argued against multiple instances of in-band signalling in recent PRs at my workplace in the previous few weeks, in C# no less! There are a variety of better options!
The topic seems interesting but the more and more obvious involvment of LLM makes me question the reality of this. It seems plausible but it's the LLM taint: the style makes you wonder if the content has been hallucinated.
I can believe the contents are true. I wouldn't mind it that much if the actual attempts have been done by an LLM agent. However I don't like that the poor writing by an LLM incurs a toll on the reader.
I can think of two possibilities:
lots of data collected by hand but making a report from that is long and tedious so an LLM was used: then give the raw data!
data collected by LLM and an LLM's natural output is prose and that page was either the direct output, or a secondary generation.
tclancy | 12 hours ago
I’m sorry, what in the Suicidal Tendencies is that header image and font? I am all the way in.
val | 4 hours ago
icefox | 12 hours ago
Good to see that some of the best traditions of hackers are still out there: stainless steel rats, nibbling on the concrete wainscotting of society.
rau | 6 hours ago
In the context of the discussion of *sshhole, it’s interesting to see yet another issue caused by the in-band signalling so common in C APIs, in this case using 0 as a sentinel rather than as 0.
frogulis | 2 hours ago
"in-band signalling" -- thanks, exactly the phrase I've been looking for. I've argued against multiple instances of in-band signalling in recent PRs at my workplace in the previous few weeks, in C# no less! There are a variety of better options!
janiczek | 4 hours ago
Shame the AI prose writing tells get more and more severe as the article progresses.
adrien | an hour ago
Felt the same.
The topic seems interesting but the more and more obvious involvment of LLM makes me question the reality of this. It seems plausible but it's the LLM taint: the style makes you wonder if the content has been hallucinated.
I can believe the contents are true. I wouldn't mind it that much if the actual attempts have been done by an LLM agent. However I don't like that the poor writing by an LLM incurs a toll on the reader.
I can think of two possibilities:
In any case, sigh, so much for trust in writings.