Pink40mmHEDP_Potato on Nostr: <li><p>General ranting about some of our finest and brightest in various ...

<li><p>General ranting about some of our finest and brightest in various Cybersecurity niches are willing to let a non-deterministic application, that CAN and DOES lie to the user (the software is NOT hallucinating, as it is not a living, thinking being):</p><ul><li>Create applications for them, and trust that they're engineered well</li><li>Reverse Engineer malware for them, and trust that the commands were ran correctly, the output is valid, and that the report wasn't in any way made up</li><li>Do patch diffs for them and generate an 0-day for giggles</li></ul></li><li><p>Doctors are more frequently willing to let the clankers annotate office meetings. </p><ul><li>Between shit ass integrated microphones in most laptops being the absolute most bottom of the barrel shit they can get away with installing, and the fact that the clankers CAN and will lie when they don't understand or can't interpret what is being said, its a terrifying thought. </li><li>For consideration, google is a major AI player, and there are options for youtubers to auto-generate closed captioning for their videos. These clanker closed captions are laughably bad. Now, think about that, and apply it to your healthcare. Your doctor is using this shit to annotate your health.</li><li>Even assuming you opt out, and make them take notes manually, how do you know they aren't having the AI summarize everything? </li><li>What is the status on HIPAA protected data and the use of AI? </li><li>Where is the data stored? </li><li>Is the AI ingesting the annotations? Is that data being fed back to the techbro oligarch company that provided the AI?</li></ul></li><li><p>Screaming and gnashing of teeth</p></li>