Article URL: https://trufflesecurity.com/blog/scanning-7-6-petabytes-of-ai-training-data-for-secrets Comments URL: https://news.ycombinator.com/item?id=49136955 Points: 7 # Comments: 0
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Morph, a Y Combinator Summer 2023 batch startup, has posted an opening for a Member of Technical Staff. The job listing was shared on Hacker News, though it received no points or comments at the time of reporting. The position details are available through Y Combinator's company job board. No further information about the role's requirements or responsibilities was provided in the listing summary.
Article URL: https://www.youtube.com/watch?v=pHcZpvIfho0 Comments URL: https://news.ycombinator.com/item?id=49139325 Points: 9 # Comments: 0
Researchers have demonstrated that multimodal AI agents such as Gemini 3 Pro and GPT-4o-audio can be manipulated through hidden commands embedded in ambient sound or overlapping speech, achieving a 69% attack success rate under lab conditions. The vulnerability is an audio-based extension of prompt injection, a well-documented flaw in text-based AI systems, now made more potent by the always-listening design of modern voice agents. Because these models continuously parse incoming audio for instructions, any sound near a microphone becomes a potential attack surface. A proposed defense called cross-modal consistency detection (CADV) was able to identify such attacks over 90% of the time, suggesting the problem is a pipeline design gap rather than an unfixable architectural flaw. Developers building voice assistants, call-center bots, or ambient-listening agents are urged to treat all incoming audio as untrusted input and implement input segmentation and provenance checks.
A study from MIT Sloan found that AI-generated financial advice is surprisingly effective, particularly when users ask the right questions. The research highlights that the quality of guidance depends significantly on how queries are framed. MIT Sloan researchers explored AI's potential as a financial advisory tool, assessing its accuracy and usefulness. The findings suggest AI could serve as a viable resource for personal finance guidance, though the approach to questioning plays a critical role in outcomes.
Discussion (0)
Log in to join the discussion and vote.
Log in