Arthur Besse@lemmy.ml to Technology@lemmy.mlEnglish · 10 months ago"participants who had access to an AI assistant wrote significantly less secure code" and "were also more likely to believe they wrote secure code" - 2023 Stanford University study published at CCS23arxiv.orgexternal-linkmessage-square31fedilinkarrow-up1282arrow-down17
arrow-up1275arrow-down1external-link"participants who had access to an AI assistant wrote significantly less secure code" and "were also more likely to believe they wrote secure code" - 2023 Stanford University study published at CCS23arxiv.orgArthur Besse@lemmy.ml to Technology@lemmy.mlEnglish · 10 months agomessage-square31fedilink
minus-squareSl00k@programming.devlinkfedilinkEnglisharrow-up1·10 months agoThe jump from GPT-4o -> o1 (preview not full release) was a 20% cumulative knowledge jump. If that’s not an improvement in accuracy I’m not sure what is.
The jump from GPT-4o -> o1 (preview not full release) was a 20% cumulative knowledge jump. If that’s not an improvement in accuracy I’m not sure what is.