Dark Hacker News
new
|
best
|
ask
|
show
|
jobs
undefined | Dark Hacker News
16 points
by
auspiv
3 hours ago
pavlov
2 hours ago
|
next
[−]
This post is obviously AI-written and it’s practically incomprehensible without some background context that’s missing here.
Oarch
2 hours ago
|
next
[−]
Just tried this in Opus 5:
"Looks like your message came through empty, just a dash. What did you want to ask?"
Insanity
2 hours ago
|
parent
|
next
[−]
I couldn’t replicate it either. Not sure if the user who discovered it, accidentally did this in an environment with “memory” like an OpenClaw.
toastau
2 hours ago
|
parent
|
next
[−]
To the API, without any instructions / user settable "system" prompt?
minraws
2 hours ago
|
root
|
parent
|
next
[−]
You can use pi and empty out context and system prompt for it that I think should be a fair comparison too.
Tiberium
2 hours ago
|
next
[−]
People should just try base models, or models where you have access to prefills. I guess so many who use LLMs nowadays have never actually experimented with a text completion endpoint locally/over API.