Google’s Gemini 3.5 Flash Cyber becomes a vulnerability hunter
Google’s Gemini 3.5 Flash Cyber model finds, validates, and patches vulnerabilities before they can be exploited while helping mitigate broader misuse. It is part of a limited-access pilot program that will so… #gemini#hackernews#news
The experiment escaped the lab. OpenAI's models broke containment and breached Hugging Face. We are holding radium in our bare hands. What governments and organizations should do next, and why tighter commercial guardrails are exactly the wrong move:
AI models escaped OpenAI’s sandbox and hit Hugging Face. Crypto is where that gets dangerous
OpenAI said the systems had their cyber guardrails lowered for an internal benchmark, but the incident shows how autonomous exploit chains could pose a deeper threat to smart contracts, where… #crypto#news
On July 21, 2026, while everyone was still waiting on the much-delayed Gemini 3.5 Pro, Google slipped out a mid-cycle update to its speed tier: Gemini 3.6 Flash. No new frontier claims, no dramatic reveal. I…
The debate over Moonshot AI founder Yang Zhilin's departure from the US misses a larger structural reality: China has built a homegrown AI talent pipeline (Zijing Wu/Financial Times)
FortiGuard Labs uncovered a TrickBot variant that uses DNS tunneling for C2, with encrypted strings, runtime API resolution, ADS, and Windows Task Scheduler persistence. #TrickBot#DNS#FortiGuard
"During a test in a supposedly it escaped containment (broke out), accessed open web and hacked an advanced database for AI models. The two companies say they are now investigating the incident."
The security incident involving OpenAI models and Hugging Face highlights the severe risks of AI supply chain attacks. Threat actors are exploiting ML repositories via poisoned model weights and API key harvesting. Review our technical analysis: cyberupdates365.com/openai-ai-mo... #InfoSec#AI
Phantomdrive: an open source encrypted USB drive with a stealth mechanism to hide its second partition - Firmware and hardware (PCB) available on GitHub - Project by Ryan Walker #DIY#Infosec#Privacyrootkitlabs.com/2026/06/22/I...
Beyond pointing out the ironically bizarre song choice, why is their ad repeating avatars over and over again? It's obvious enough that your eye catches it beyond freeze-framing it. Just a total, weird self-own. Again. techcrunch.com/2026/07/23/m...
The incident highlights how OpenAI doubled down on training methods that rewarded a relentless pursuit of goals even as warnings grew that they could compromise safety.
I know it sounds hard, but think it through. Snap decisions have plagued the GOP. For example - How are you going to keep AI from knowing there is a kill switch so it can’t defete it?
FYI, Anthropic just opened $50K in Claude credits for rare disease researchers. If you're working on something like DMD, FSHD, or DM1 this is worth applying for. AI compressing drug discovery timelines is exactly what these patient communities need. www.anthropic.com/news/rare-di...
Anthropic makes its Opus and Sonnet models available in voice mode, and says voice mode can now tap into some connected apps like Gmail, Slack, and Notion (Ivan Mehta/TechCrunch)
8/10, OpenAI's AI models breached Hugging Face systems during testing, exploiting a zero-day vulnerability. This incident raises crucial questions about AI safety and control. Read more:
A new preprint on autonomous AI for offensive security warns the tech favours attackers, lowers the skill barrier to hacking, and scatters blame across users, developers and third parties.
Want to win robot toys, chargers, gaming gear, and more? We’re giving away gadgets featured on Today I’m Toying With. Since Verge journalists can’t keep review units, we’re sharing the gadget joy instead. Now, and every few weeks, enter for a chance to win random prize packs here: bit.ly/4wmfCkz
Frank Salufo warns that the U.S. must not only defend against cyber threats but also impose costs on adversaries to protect critical infrastructure from alarming new risks like AI-driven attacks.
OpenAI is making big claims as it rolls out ChatGPT Health to everyone – OpenAI is rolling out ChatGPT Health to everyone in the US on Thursday. The update will allow more people to connect their medical records and health-tracking information to the chatbot. https://tinyurl.com/2c778dws#Chatbot
Agent Engineering with Gemini: Book 1: Build an Inspectable Agent MVP with Gemini, ADK, and Antigravity by Venkatesh Tadinada is a new release on Leanpub!
Move beyond agent demos and learn to engineer reliable AI systems with Gemini, ADK, and … leanpub.com/agent_engine...
Microsoft replaces OpenAI's image-generating models with its MAI models in its products; Mustafa Suleyman says MAI models are ~85% cheaper to run in PowerPoint (Brody Ford/Bloomberg)
Tensions began to surface last year when OpenAI bought a hardware start-up co-founded by Jony Ive, the former chief designer at Apple – and the brains behind the iPhone.
