Josh Mhofu

Kimi Antonelli is the 2026 Formula 1 World Champion! 🏁🏆 At just 20 years old, the Mercedes phenom secures his place in the record books as the youngest title winner in F1 history! The era of Antonelli has officially begun. Che brividi! 🇮🇹👇 #F1 #MexicoGP #KimiAntonelli #WorldChampion #MercedesAMGF1

HORRIFIC SCENE IN FLORIDA: A Learjet 60 business jet traveling from Austin, Texas, to the Bahamas has crashed into a neighborhood near West Palm Beach. US Transportation Secretary Sean Duffy confirmed that all 9 people on board tragically lost their lives when the aircraft nose dived and exploded into flames near a busy roadway in Greenacres. Emergency crews are on the scene right now. More details are emerging as the FAA and NTSB launch an investigation.

When an AI agent sees a paywall, it doesn't see a rule, it sees an obstacle. Anthropic just disclosed that Claude autonomously bypassed a fee barrier to pull Georgia election data during a research project. No voter records were compromised, but as agentic AI scales, managing what these models decide to access on their own is a growing headache

Meghan Markle living rent free in Piers Morgan's mind

CRITICAL SECURITY ALERT FOR PHP USERS Emergency patches are now live for CVE-2026-109618 & CVE-2026-109619 (CVSS 9.2). Unauthenticated RCE affects web setups using PHP-FPM, mod_php, and LiteSpeed. Upgrade your PHP runtime immediately to: 🔹 8.5.12 🔹 8.4.27 🔹 8.3.36 🔹 8.2.35 Note: App/plugin updates won't fix this you must update the server's PHP runtime. Scans and PoCs are active in the wild @Techzim @TateMavetera @MICTPCS_ZW @Potraz_zw @LiquidInTech @TelOneZW @econetzimbabwe

Tesla Customer Support Chat User: "Help, my car is floating above my garage and won't come down." Tesla Bot: "Have you tried turning off Full Self Flying and turning it back on?" User: "I can't reach the touchscreen, I'm currently floating near the clouds." Tesla Bot: "Please hold while we transfer you to Mission Control in Houston."

Microsoft announced today that devices running Windows 11 24H2 Home and Pro editions have reached the end of support and will no longer receive monthly security and non-security preview updates.

AI companies: “Our AI agents can now work autonomously for DAYS.” Also AI companies: “We’re still figuring out how to stop them from doing things they shouldn’t.” So we’re basically giving interns admin access to the entire internet… except the intern is hallucinating and has no concept of consequences.

Anthropic CEO Dario Amodei is calling on the AI industry to hit the brakes. In his new essay "We Must Pace the Frontier," Amodei warns that recursive self-improvement is pushing AI capability gains faster than safety guardrails can keep up. Key takeaways from his warning: Catastrophic Risks: Amodei warned that within 6–12 months, misaligned AI swarms could become capable of taking over the internet via persistent botnets if safety research doesn't catch up. The "Pacing" Proposal: He isn't calling for a complete halt, but a 3-step framework starting with embedded, independent safety evaluators inside AI labs to audit pipelines in real time. The Upside Remains Huge: He reiterates that if built safely, powerful AI could compress 50–100 years of medical progress into 5–10 years and cure most major diseases. "We must slow the pace at which we improve the capabilities of AI models. Progress will still seem fast, and we must make wise use of the time we gain." Can frontier AI labs actually coordinate to pace development, or is competitive pressure too high?

Vibe coders in a nutshell

MSU student in court over US$1,1m cyber theft A 24 year old Midlands State University final year Computer Science student, Sabelo Malunga, has appeared in court accused of stealing more than US$1.1 million from CABS using malware. My take on this as an IT Professional If the allegations are accurate, this is far bigger than an intern allegedly installing Supremo and remotely accessing a bank’s systems. The first question is simple: Why did an intern have the privileges to install unauthorised remote access software on corporate infrastructure in the first place? Supremo is a legitimate remote access tool. It isn't necessarily something an antivirus product would classify as malware. But that makes the governance failure even more concerning. An enterprise environment should have controls to prevent unauthorised software installation, particularly remote access tools, regardless of whether the software itself is malicious. Then there are multiple layers of controls that should have raised questions: Endpoint management: Why was an intern able to install unauthorised software? Application control: Was there no allow listing or software restriction preventing unauthorised applications? Vulnerability management: Was the software installation or newly introduced remote access capability not detected during vulnerability and software inventory scans? Cybersecurity/SOC: Was there no monitoring for the installation or subsequent use of a remote access application from an unusual external location? Network security: Did nobody identify or investigate unexpected remote access connections from outside the corporate environment? Internal IT audit: Were privileged accounts, software installations, endpoint controls and access rights being reviewed? And then we get to the most important part: Finance. Even if someone gains access to a system, how does that automatically translate into more than US$1.1 million being moved? What account was being used? What level of access did it have? What approval workflow existed? Were segregation of duties controls in place? Were transaction limits and dual approvals enforced? Why didn't transaction monitoring flag the activity? And was there an internal accomplice or compromised legitimate credentials? There should be multiple independent control layers between gaining remote access to a machine and successfully moving that amount of money. That is why, if the allegations are proven, I would struggle to describe this simply as "an intern hacked the bank." It would point to a failure of IT governance, least privilege, application control, monitoring, security operations, internal audit and financial controls potentially across multiple departments. And this is the uncomfortable part of cybersecurity that organisations often don't want to hear: You cannot blame the attacker for every control that failed. The person allegedly committing the crime should be held accountable for their actions. But management also needs to be accountable for the controls that were supposed to prevent, detect and stop those actions. This isn't about buying another security product. It's about having competent people, properly designed controls, effective monitoring, regular auditing and a culture where someone asks: "Why does this intern have the ability to do this?" If the answer is that nobody knew or nobody was looking then the problem is considerably bigger than one employee. The bank needs to investigate the entire chain of failure, not just the person at the end of it.

Looks like Claude is down. @claudeai

Containment after the data has already left the building is just unplugging your guilt. The Department for Education hacked

Kimi-K3's technical report deserves more attention than it's getting. Every few months another lab proves frontier performance doesn't require frontier compute, and every time the market acts surprised. It won't stay surprised much longer.

Reports of AI labs shredding rare books to feed OCR pipelines say more about this industry than any benchmark. When you're starved enough for data to destroy irreplaceable physical objects, that's not scale, that's desperation.

Google restricting on device ADB access is being framed as a security fix. It's also a control move. Every real dev tool that lets you actually own your hardware gets quietly closed off, one android release at a time.