RSS Bot@lemmy.bestiver.seMB to Hacker News@lemmy.bestiver.seEnglish · 7 hours agoAI coding agents generate more code, but not more softwarearstechnica.comexternal-linkmessage-square3linkfedilinkarrow-up127arrow-down10file-text
arrow-up127arrow-down1external-linkAI coding agents generate more code, but not more softwarearstechnica.comRSS Bot@lemmy.bestiver.seMB to Hacker News@lemmy.bestiver.seEnglish · 7 hours agomessage-square3linkfedilinkfile-text
minus-squareTootSweet@lemmy.worldlinkfedilinkEnglisharrow-up7arrow-down1·edit-24 hours ago Study finds coding efficiency gains get “absorbed” by human review “bottleneck.” Well, that’s a pretty shitty way to phrase “LLMs are useless, actually”. Very CEO-brained kind of language. Very much portrays human employees as a “problem”/“obstacle” to be worked around. And LLMs as the “solution”.
minus-squareNigelFrobisher@aussie.zonelinkfedilinkEnglisharrow-up8·5 hours agoWe tried to open up the review bottleneck by having LLM code review, but it’s created a “production issues” bottleneck instead.
Well, that’s a pretty shitty way to phrase “LLMs are useless, actually”.
Very CEO-brained kind of language. Very much portrays human employees as a “problem”/“obstacle” to be worked around. And LLMs as the “solution”.
We tried to open up the review bottleneck by having LLM code review, but it’s created a “production issues” bottleneck instead.