Create your own feed. Read from the websites you follow. Cut the noise. Ditch the big tech algorithms. Self-host your own feed reader. Here are the best choices.
“Under current U.S. hacking laws, a human can face criminal charges for breaking into someone else’s computer without permission. But when an AI agent autonomously hacks into a company’s computers, determining who is liable is much murkier.” techcrunch.com/2026/08/03/w...
Convex, which provides an AI-optimized application backend for developers, raised a $57M Series B led by Insight Partners, taking its total funding to $110.5M (Maria Deutscher/SiliconANGLE)
This @cnbc.com article cites my work on the lack of staff capacity in Congress leading to increasing staff reliance on AI tools. As Marci Harris of @popvoxfdn.bsky.social explains in the piece “Staffers are famously overwhelmed, and actually AI is exacerbating that." www.cnbc.com/2026/08/03/o...
It's far easier to find security holes than to fix them, and leaving it to AI can introduce 9 times as many new vulnerabilities as developers do: AI is finding bugs faster than humans can fix them: How enterprise security teams must adapt www.zdnet.com/article/ai-i...
With Chinese company Z.ai’s #GLM5.2 open-weight model providing cheap AI access to software developers, have U.S. tech companies lost their AI edge?spectrum.ieee.org/ai-coding-as...ic
Centre for Society, Technology and Values @uofwaterloo@cstv-uw.bsky.social
The recent hack of #HuggingFace by an #OpenAI model during testing illustrates that current ideas about #security, such as #AI Kill Switch, do not address significant risks of current technology, argues Simon Blanchette
Midnight Blizzard's CaptiveCrunch campaign poisons hotel Wi-Fi DNS to deliver CornFlake RAT and steal Microsoft 365 credentials from traveling professionals.
We are a big fan of Google's Open Knowledge Format for providing structured knowledge to agents. In this Deeper Dive, we review OKF & its new v0.2 along with some use cases & domain specific applications
PeppyPanda 🐼 - Uplifting and good news@peppypanda.com
Good news for future tech stars! Police are finding creative ways to help teen hackers apply their skills positively, preventing them from falling into cybercrime. Read the uplifting BBC report on these diversion programs. ✨ #YouthInTech#CyberSafety#PositiveImpact#BBCNews
PharmaEssentia and Panco Healthcare were hit by a cyberattack affecting some systems, but operations remain largely intact. Defenses were activated, experts engaged, and authorities notified. #Taiwan#PharmaEssentia#PancoHealthcare
The iPhone maker has accused OpenAI and former Apple employees of deliberately stealing Apple trade secrets. But there’s also a long history of Apple alumni accidentally taking confidential material with them.
They've turned your calendar into a postbox. Commands arrive as events scheduled for the year 2099, stolen data leaves as encrypted attachments, all riding through Microsoft Graph...
The White House is navigating uncharted waters with AI regulations, leaving companies like Anthropic in limbo. With no clear rules, the future of innovation hangs in the balance. How should we balance safety and progress?…
From the open secret of AI use, to Ben Affleck's $587M startup sale, to Amazon MGM dropping that OpenAI biopic...we covered a lot. Read and listen here:
A critical flaw (CVE-2026-17583) in Thermo Fisher's forensic DNA software allows for nearly undetectable evidence tampering. The vulnerability impacts the integrity of the criminal justice system. Patches are available. #CyberSecurity#Forensics#DNA
UK's Police National Legal Database (PNLD) breached. Names, work emails of police officers, govt partners & public exposed on the dark web. The service supports all 43 Home Office police forces. #DataBreach#UKPolice#LawEnforcement
Russian actor Midnight Blizzard (Storm-2945) exploits Outlook XSS flaw CVE-2026-42897 to access mailboxes. Also hijacks hotel Wi-Fi in 'CaptiveCrunch' campaign to steal M365 tokens with CornFlake & ChocoShell malware. #ThreatIntel#APT
Open-source supply chains are under pressure as attackers target PyPI, npm, Docker, and GitHub Actions. Active SonicWall exploitation, N-able N-central takeovers, and new tools like AtlasRAT also surface in phishing campaigns. #PyPI#SonicWall#Nable
Sandisk and SK Hynix announce the High Bandwidth Flash (HBF) open specification that enables up to 512GB memory modules with 0.4 TB/s to 3.0 TB/s bandwidth (Anton Shilov/Tom's Hardware)
OpenAI hosted some social media influencers at a luxury retreat in upstate New York to generate positive buzz.
Instead, followers asked influencers for follow-up videos about AI data centers’ environmental impact or OpenAI’s $200M deal with the Dept. of Defense. techcrunch.com/2026/08/03/i...
Are there chatbots in your therapy sessions? @psychtoday.bsky.social discusses this growing trend as people are influenced by learnings from ChatGPT, Claude, Gemini and more.
As this continues, therapists are “compelled to make a case for ‘old school’ therapy.” tinyurl.com/2s4tdt6h
The GUARD Act's definition of an "AI companion" is broad enough to sweep in ChatGPT, Claude, and Gemini, argues Dr. Sameer Hinduja. He makes the case for six fixes that tie regulation to what a product does, not its app-store label.
Hackers of all skill levels are developing attacks and finding exploitable software vulnerabilities with the help of mostly closed AI models. www.axios.com/2026/08/04/e...
This part! “The attacks [on local water infrastructure], tentatively attributed to Iran, expose a dangerous mismatch at the heart of American cybersecurity: The threat is geopolitical, while the defense is municipal.” — Jen Easterly, former #CISA Chief www.nytimes.com/2026/08/02/o...
Sadia Mehrin Rahi, Ayesha Siddika, Istiyak Hasan Maruf, Adrita Rahman Tory, Muhammad Aminur Rahaman, Khondokar Fida Hasan
A Decade of Healthcare Cyber Threats: Empirical Analysis, Evidence-Based Prioritisation, and AI Threat Model https://arxiv.org/abs/2608.00901
Apple has expanded its trade secrets lawsuit against OpenAI, alleging that more former employees may have taken confidential company information with them after joining the…
The article discusses how adversaries are increasingly weaponizing AI for malicious purposes, as analyzed by Talos. It identifies three main categories of AI usage: creating sophisticated malicious software, scaling criminal operations, and conducting vulnerability research.
"messages appear to show Liu walking his former colleagues thru accessibility issues in Apple’s internal systems. According to OpenAI, Liu did nothing wrong here, & the blame should instead fall on Apple for “failing to properly manage system access when ppl leave”" gizmodo.com/openai-is-tr...#law
"White House hosted a meeting with representatives from top AI companies, including Anthropic, Meta & OpenAI. President Trump did not attend the meeting which was focused on voluntary AI safety testing for advanced AI models" www.nbcnews.com/now/video/wh... Rogue tech breached other companies 🤦 #law
