1+ hour, 1+ min ago   (634+ words) Alternating updates of model weights and executable harnesses can reach accuracy on par with full fine‑tuning while consuming a fraction of the training compute. The WHALE recipe shows that interleaving a short weight‑update phase with a lightweight harness…...

1+ day, 1+ hour ago   (1092+ words) By Nokka | September 12, 2026 Suppose someone told you that you could pick between two AI models. One costs eight times less but scores only two points lower on capability. Most people answer "take the cheap one" immediately. Then, when real money…...

1+ day, 4+ hour ago   (1453+ words) 1. Basic Information Original Title: Hackers abused Claude to extract secrets from 1.8M Android apps Source: BleepingComputer / Anthropic Publication Date: 2026-09-11 Severity: High Basis of Severity: Anthropic reported multiple incidents, including actual compromises and data theft, where AI handled attack execution, retries, and…...

1+ day, 4+ hour ago   (1540+ words) 1. Basic Information Original Title: How threat actors are turning trusted AI platforms into an attack surface Source: BleepingComputer, Huntress Publication Date: 2026-09-11 Severity: High Basis for Severity: Trusted AI sharing pages are being used to distribute malware via ads and SEO,…...

3+ day, 6+ min ago   (249+ words) Um das Ausmaß des Vorfalls zu verstehen, müssen wir die Ereignisse Schritt für Schritt nachvollziehen. Es ist eine Geschichte, die sich wie ein Cyber-Thriller liest, aber bittere Realität ist. Hätte Freund diese kleine Anomalie ignoriert, wäre die Backdoor unentdeckt in…...

4+ day, 2+ hour ago   (808+ words) My last update here was on August 25. Since then, OpenWorkProof Core 1.4.0 and our first DeepSeek Harness plugin have shipped. I owe readers a concrete account of what is available and how to check it. If you are new to the…...

6+ day, 6+ hour ago   (411+ words) "Don’t put sensitive data into an AI prompt." I hear this advice a lot, and while it makes sense, I... Tagged with ai, privacy, security, programming....

1+ week, 1+ day ago   (158+ words) That means anything the agent ever saw — an API key, a database URL, the output of env — is sitting in plaintext on disk indefinitely. Nobody was auditing those files. So I wrote a small CLI that does. I ran it…...

1+ week, 1+ day ago   (621+ words) Intro: In April 2026, Anthropic released an AI so capable it independently found zero-day vulnerabilities hidden for 20+ years in every major OS and browser. They locked it inside Project Glasswing. Ten days later, a 22-year-old developer rebuilt the entire architecture from…...

1+ week, 1+ day ago   (802+ words) On Tuesday, Google announced Gemini 3.8 Flash Cyber, a model that its own Cloud Vulnerability Research team used to find a critical, foundational vulnerability in under two hours. Research that Google says normally takes months. The Chrome Security team reported it…...