Richi JenningsAug 15, 2023

#ChatGPT is wrong more than half the time—makes many conceptual errors, but sounds confident, authoritative.

So, hard to spot the errors, say researchers. In this week’s #SSBlogwatch we can’t say we’re totally surprised. For #ReversingLabs: https://reversinglabs.com/blog/ai-coding-helpers-get-failing-grade?utm_source=richisoc&utm_medium=social #AI #DevOps

AI coding helpers get FAILing grade

Purdue researchers expose generative AI tools like Copilot's frequent errors when asked basic development questions. Only fools rush in.

ReversingLabs
Richi JenningsAug 8, 2023
yes, even over Zoom, Twitch or Slack Huddle. In this week’s #SSBlogwatch we turn up the music. For #ReversingLabs: https://www.reversinglabs.com/blog/ai-deep-learning-enables-attackers-to-hear-your-password?utm_source=richisoc&utm_medium=social
Listen up, devs: AI trained to overhear passwords

Deep learning model knows what keys you press — “with 95% accuracy.” The password's days are numbered.

ReversingLabs
Richi JenningsJul 25, 2023

Google employees are to be protected from themselves. In what’s being described as a pilot program, they’ll lose internet access at work and/or root privileges.

The idea is to stop break-ins by bad actors. In this week’s #SSBlogwatch we try not to imagine the horror. For #ReversingLabs: https://www.reversinglabs.com/blog/no-net-for-some-no-root-for-devs-google-cuts-off-staff-internet-for-safety?utm_source=richisoc&utm_medium=social

No net for some, no root for devs — Google pilot walls off staff internet, access for ‘safety’

It’s an optional trial program (for now). How would your dev team cope? Is this the future of zero trust?

ReversingLabs