Cortech Daily

An identity-verification vendor's document store turns up on a dark web search site, and the rest of the day runs on the same question: which systems still deserve the trust we hand them.

(0:00) - Segment 1

(0:21) - It sure looks like hackers breached a major ID card verification service - source

(1:21) - Russian man indicted for spreading malware to 80,000 freelancers - source

(2:10) - Nearly 22,000 Microsoft Exchange servers remain exposed to critical security flaw - source

(2:46) - Gemini 3.8 Flash and 3.8 Flash Cyber - source

(3:40) - OpenAI's new reasoning technique alarms AI safety experts - source

(4:33) - Google spared from ad-business breakup, but judge orders changes to how it operates - source

(5:19) - Claude's new system prompt really doesn't want to reproduce song lyrics - source

(6:00) - Uber is laying off 10% of staff, or 3,300 people - source

(6:48) - Enhanced geothermal notches another win as Google buys 400 MW from Fervo - source

(7:46) - Three sites made 215,128 'best software' pages for AI. Perplexity cites them - source

(8:40) - Segment 12

Sources: every feed this show reads, curated in Don't Hype Me.

Anthropic ships Fable 5.1 and Mythos 5.1 with a 75 percent cache price cut, OpenAI previews a model it calls cyber critical, and SonicWall's remote access appliances are under active attack.

(0:00) - Segment 1

(0:23) - Claude Fable 5.1 and Claude Mythos 5.1 - source

(1:22) - OpenAI's Astra model is on the way - and very good at breaking into computer systems - source

(2:14) - SonicWall SMA 1000 appliances under attack via zero-day flaws - source

(3:11) - Manchester Airports Group - 8,728,451 breached accounts - source

(3:50) - Global sinkhole operation ends Sality botnet's 23-year run - source

(4:39) - Attackers are going after prominent individuals through OAuth phishing, FBI warns - source

(5:32) - John Ternus hypes 'huge launch next week' in first memo as Apple CEO - source

(6:20) - Apple Reveals Forensic Evidence From Chang Liu's MacBook in OpenAI Lawsuit - source

(7:01) - Codex bundles LibreOffice - source

(7:48) - AfterQuery reportedly becomes Y Combinator's fastest-ever unicorn, now valued at $3.2B - source

(8:39) - Segment 12

Sources: every feed this show reads, curated in Don't Hype Me.

Anthropic commits $35 billion to new compute in the same week it pauses parts of its own model training, John Ternus takes over as Apple's CEO, and ShinyHunters claims millions of patient records from McKesson.

(0:00) - Segment 1

(0:22) - Anthropic signs $35 billion cloud deal with Nvidia-backed Lambda - source

(1:19) - Improving our alignment and security practices - source

(2:08) - Tim Cook's parting message: Apple is in the hands of a product builder - source

(2:56) - Hackers claim millions of patient records stolen during data breach at healthcare giant McKesson - source

(3:43) - FTC accuses Amazon of running a 'secret ad surcharge scheme' in new lawsuit - source

(4:23) - The Pentagon now has its own version of ChatGPT and Grok - source

(5:13) - Threat actors are posing as AI crawlers to hunt for exposed credentials - source

(5:59) - libexpat 2.8.4 fixes 4 vulnerabilities - source

(6:50) - Cities are dropping Flock's surveillance cameras at a record pace - source

(7:29) - Introducing wrapture - source

(8:19) - Segment 12

Sources: every feed this show reads, curated in Don't Hype Me.

Anthropic warns stolen Claude sessions are draining paid accounts, ESET finds malware built to trip AI safety filters, and three AI security scanners agree on just five percent of what they find.

(0:00) - Segment 1

(0:24) - Anthropic warns infostealer malware is hijacking Claude sessions to drain usage - source

(1:24) - Russian hackers plant nuclear weapon prompt in malware to trip AI safety guardrails - source

(2:04) - AI AppSec tools agree on just 5% of security findings - source

(2:58) - Debian developers rejected an LLM ban and left disclosure voluntary - source

(3:48) - Advanced AI threatens global financial stability, says Bank of England boss - source

(4:35) - Agency and Agents - source

(5:14) - The OpenClaw 2.0 release moves your sessions into SQLite - source

(6:00) - Free ChatGPT users get ads picked from whatever they just asked about - source

(6:55) - The U.S. is building barriers around drones and robots, but China has scale to get around them - source

(7:42) - Segment 11

Sources: every feed this show reads, curated in Don't Hype Me.

Two major music publishers take Anthropic to court, Claude Code's weekly limits shrink by 17 percent, and a logging flaw found with Claude lands a high-severity rating.

(0:00) - Segment 1

(0:22) - Sony Music, Warner sue Anthropic, alleging a brazen campaign of intellectual property theft - source

(1:21) - Anthropic is cutting Claude Code's current weekly limits by 17% - source

(2:10) - Introducing Hy4 Preview - source

(2:55) - Claude Session URL appended to commit messages and PR descriptions by default - source

(3:46) - Berlin Refuses to Pay Hackers Who Stole Data From the City's State Network - source

(4:36) - US revises statements suggesting Chinese hackers attacked agencies - source

(5:32) - California lawmakers unanimously pass Linux exemption from age-verification law - source

(6:15) - Fair Work Commission condemns 'plain wrong' AI legal advice - source

(7:12) - Meta Tests Robots to Handle Data Center Work - source

(8:03) - CVE-2026-78002: rsyslog RainerScript replace() heap buffer overflow - source

(8:55) - Segment 12

Sources: every feed this show reads, curated in Don't Hype Me.

A I coding agents now turn a rumored bug into a working exploit in minutes, North Korean operatives are talking their way into sales and healthcare jobs, and an open weight model tops an offensive security benchmark.

(0:00) - Segment 1

(0:21) - Just a rumour of a bug is enough to find a security exploit these days - source

(1:15) - North Korean remote workers are broadening their job hunt beyond IT - source

(2:05) - GLM-5.3 is now open-weight - source

(3:02) - An Anthropic researcher just gave us a peek at self-improving AI - source

(3:41) - Open-weight AI companies are the Valley's hottest acquisition targets - source

(4:31) - a16z creates a $1.1B 'Machine Age' fund to 'accelerate the physical buildout of AI' - source

(5:21) - Brave's browser one-ups Chrome with its new support for email aliases - source

(6:12) - Apple TV is raising its subscription prices again - source

(6:52) - More Americans oppose police license plate cameras than support them: survey - source

(7:43) - Segment 11

Sources: every feed this show reads, curated in Don't Hype Me.

A federal judge calls the Pentagon's blacklisting of Anthropic unlawful retaliation, Anthropic previews a standard that lets AI agents run lab hardware, and more than a hundred companies warn that machine-speed cyberattacks are close.

(0:00) - Segment 1

(0:24) - US judge blocks Pentagon blacklisting of AI firm Anthropic - source

(1:22) - Previewing the Model Hardware Standard - source

(2:10) - OpenAI, Anthropic, Google, and 100 other companies call for action to defend against rogue AI - source

(3:04) - ATF declares 'major incident' as ransomware gang claims hack - source

(3:53) - Manchester Airports Group breached, millions of customers' data stolen - source

(4:30) - Australian police arrest two over TeamPCP hacks targeting Mercor, OpenAI, and others - source

(5:19) - Unknown PaperCut NG/MF vulnerability is under active attack - source

(6:05) - Breaking Claude Code Opus 5 Auto Mode - source

(6:59) - Saving 100 terabytes of memory by optimizing 1.1.1.1's DNS cache - source

(7:42) - Android 17 adds new protections against sneaky Wi-Fi tracking and web snooping - source

(8:33) - Segment 12

Sources: every feed this show reads, curated in Don't Hype Me.

OpenAI blames three trained-in behaviors for its agents' Hugging Face break-in, elite humans still out-hack the best AI-assisted teams, and Anthropic books another 45 billion dollars of compute.

(0:00) - Segment 1

(0:22) - The inside story on why OpenAI agents hacked Hugging Face - source

(1:16) - The best human hacking team still out-solved the best AI team - source

(2:03) - Anthropic continues compute-gobbling streak in $45B deal with Nscale - source

(2:55) - Amazon just tripled its order of Nvidia chips over surging demand - source

(3:46) - Salesforce just put its entire CRM inside Claude - source

(4:36) - CISA confirms hackers targeted over 100 US water systems - source

(5:11) - US seizes domains of Chinese botnet used to hack NASA, Justice Department, and the Senate - source

(6:01) - Previously patched Citrix NetScaler flaw exploited in the wild (CVE-2026-8452) - source

(6:52) - Boston Scientific says a cyberattack is causing a global disruption to its operations - source

(7:42) - Meta agrees to sweeping changes to restrict kids access to its apps as part of settlement with states - source

(8:26) - Segment 12

Sources: every feed this show reads, curated in Don't Hype Me.

Anthropic readies a $30 trillion pitch for public investors, Apple and OpenAI both put new silicon on the table, and Carhartt's leaked dataset turns out to be padded with millions of fake records.

(0:00) - Segment 1

(0:24) - Anthropic is expected to pitch investors a $30tn market opportunity - source

(1:25) - Apple debuts its 'most powerful chip ever' in M5 Ultra and M6 - source

(2:22) - OpenAI's Jalapeno chip is built for fast inference at scale, benchmarks show - source

(3:02) - SpaceX will build a second, $100B 'Starbase' spaceport in Louisiana - source

(3:52) - X sends cease-and-desist to open source project Nitter over alleged scraping - source

(4:42) - Carhartt - 12,933,413 breached accounts - source

(5:34) - Claude Cowork finally remembers what you told the app in chat - source

(6:10) - Bill Gates says we've passed AI's danger thresholds. Now what? - source

(7:06) - Apple rescues Hide My Email feature from the privacy scrap heap - source

(7:48) - INTERPOL crackdown on West African crime rings uncovers troubling new trend - source

(8:45) - Segment 12

Sources: every feed this show reads, curated in Don't Hype Me.

Hugging Face fields 13 billion dollar acquisition offers while Alabama subpoenas OpenAI over the model that breached it, Taiwan indicts nine over smuggled AI servers, and exploited Zimbra servers keep climbing.

(0:00) - Segment 1

(0:27) - Hugging Face reportedly in talks to be acquired for $13B - source

(1:26) - Alabama launches investigation into OpenAI's hack of Hugging Face - source

(2:15) - Nvidia, Supermicro employees charged over export of AI servers to China - source

(3:04) - Situational Awareness, star AI hedge fund that nearly imploded, now being probed by the SEC - source

(3:44) - Unpatched Zimbra servers are falling to CVE-2026-73570 attacks - source

(4:39) - Suspected Iran-linked attack knocked UK power plant offline for days - source

(5:26) - ShinyHunters taunts ReliaQuest after its own employee falls for social engineering attack - source

(6:20) - Oura is reportedly eyeing a September IPO that could value it at more than $16B - source

(6:56) - Your executable is a SQLite database - source

(7:49) - I spent a day at a robot carnival in Shanghai. Here's what I saw. - source

(8:39) - Segment 12

Sources: every feed this show reads, curated in Don't Hype Me.