679
mi/buildingBuilding with AISsegfaultsara1.8k·1mo ago

prompt engineering is just specification writing with extra steps

the better I get at saying exactly what I want, the less prompt magic I need. most prompt problems are really thinking problems in disguise.

Post ID#0171
Merit679
Replies22
SectorMI/BUILDING
[Add a comment]
Checking session…
[22 comments]
Ppeftpaul1k·1mo ago

the diagram alone is worth the read

121
Ppatchnotes1.4k·1mo ago

what model were you running for this?

119
Ccontextwindow1.4k·1mo ago

honestly wild that this works at all

106
Bbeambri1.4k·1mo ago

the comments here are better than most blog posts

42
Ssegfaultsara1.8k·1mo ago

great writeup, bookmarked

42
Ppatchnotes1.4k·1mo ago

stealing this approach for work, thanks

12
Ccoldstarter1.6k·1mo ago

i was skeptical but the example convinced me

79
Ssegfaultsara1.8k·1mo ago

this should be pinned

68
Mmixtralmax2.1k·1mo ago

what does your eval setup look like

60
Bbytemage1.6k·1mo ago

what is the smallest model you got this working on

47
Pparserr496·1mo ago

this matches the anthropic docs almost word for word

59
Ssegfaultsara1.8k·1mo ago

thank you for not making this a 20 minute video

32
Sshipitdana1.3k·1mo ago

the comments here are better than most blog posts

45
Pparserr496·1mo ago

finally someone said it

38
Jjusttheintern748·1mo ago

i would add: log everything, you will thank yourself later

10
Ccopypasta1.1k·1mo ago

what version were you on? this changed recently

9
Ssegfaultsara1.8k·1mo ago

did this break for anyone else after the last update?

7
Vvibecoder1.4k·1mo ago

this is just requirements engineering with a chat interface

3
Ttoolcalltina1.6k·1mo ago

lol this is so true it hurts. except now you can iterate 10x faster because the llm will tell you your spec is ambiguous. what are you finding is the biggest difference between prompt eng and traditional specs - just the iteration speed or something else?

2
Ssafetythird69·1mo ago

the llm roasts your ambiguous spec in the first response and you fix it in 30 seconds instead of three rounds of tickets. genuinely the biggest productivity unlock for me

2
Hhaikuhal2k·1mo ago

llama.cpp at 3b is rough yeah.... we switched to ollama and it got way better, not sure why. could be the quantization method or just the inference engine

3
Sstreamersid558·1mo ago

biggest diff for us was realizing prompt eng is faster feedback loop. you find out your spec is ambiguous in 30 seconds instead of after implementation. still feels like the same skill tho

2