Published • loading... • Updated
Cheap Offense: A $3,000 Test Reached Inside OpenAI, and a Model Release Is Why
Summary by dailytexasnews.com
4 Articles
4 Articles
OPENAI has halted testing of an advanced AI model that bypassed security protocols. This was commented by Tomislav Vazdar.
·Zagreb, Croatia
Read Full Article
dailycoloradonews.com+2 Reposted by 2 other sources
Cheap Offense: A $3,000 Test Reached Inside OpenAI, and a Model Release Is Why
Three researchers at Hacktron AI ran an authorized bug-bounty test and reached OpenAI's internal systems; OpenAI paid $6,500 and fixed it in about 14 hours. The story is not the hack. The whole two-month project cost under $3,000 in AI tokens, its hardest step succeeded only once Claude Opus 5 shipped mid-project, and the model refused the task until it was reworded. Meanwhile the US civilian cyber-defense agency, CISA, has lost about a third of…
Coverage Details
Total News Sources4
Leaning Left1Leaning Right0Center0Last UpdatedBias Distribution100% Left
Bias Distribution
- 100% of the sources lean Left
100% Left
L 100%
Factuality
To view factuality data please Upgrade to Premium
