
A poisoned Claude Code extension, an AI-found Linux root bug, a shared AI-training opt-out, and seven more stories. (0:00) - Intro (0:26) - Claude Code extension poisoned with persistent SSH backdoor - source (1:16)...


technology
Hosted by Schmug · technology · EN-US
Keep up with AI and security without reading forty tabs. Every morning, about nine minutes on the eleven stories that actually moved — AI tooling and agents, security and breaches, Cloudflare and the edge, and developer ergonomics — each with a timestamped link to its source. Written and produced by Schmug of cortech.online, narrated by AI.
Required Pod Score for this podcast. PitchCentric checks your profile against host openness, topical fit, and audience signals before you generate a pitch.
Contact path
Public web form
Booking probability
36%
Guest openness
Selective
Sign up to generate a grounded pitch for Cortech Daily.
Signup to Generate a PitchKeep up with AI and security without reading forty tabs. Every morning, about nine minutes on the eleven stories that actually moved, AI tooling and agents, security and breaches, Cloudflare and the edge, and developer ergonomics, each with a timestamped link to its source. Written and produced by Schmug of cortech.online, narrated by AI.
Cortech Daily is hosted by Schmug. The show has published 135 episodes to date.
Recent episodes include "Claude extension backdoor, Linux kernel flaw, AI training opt-out - October 31, 2026" (31 October 2026, 9 minutes), "Binance heist, AO3 outage, Volt Typhoon relays - October 30, 2026" (30 October 2026, 8 minutes), and "DHS staff leak, Apple eMudhra certs, school login hopping - October 29, 2026" (29 October 2026, 9 minutes).
Cortech Daily publishes roughly daily, with episodes averaging 8 minutes across 135 episodes.
Booking probability for Cortech Daily is 36%, and the host is rated Selective on guests. Cortech Daily requires a Pod Score of 80 or higher. Cortech Daily's contact route on record: public web form.
The strongest fits for Cortech Daily are Cloud, Data & Web Hosting and IT Services & Software Development. Cortech Daily has a technology audience.
Schmug hosts Cortech Daily, a technology podcast with 135 episodes published.
Our AI reads these to draft pitches. Use them as grounding for a pitch that cites a real guest and a specific topic.

A poisoned Claude Code extension, an AI-found Linux root bug, a shared AI-training opt-out, and seven more stories. (0:00) - Intro (0:26) - Claude Code extension poisoned with persistent SSH backdoor - source (1:16)...

A record 6.3 billion dollar crypto theft at Binance, a weeks-long AO3 outage, Volt Typhoon's Silent Train relays, Reddit's login wall, GM's data settlement, Mistral's open trillion-parameter model, AWS spending, Thrive...

A hacker leaks details on 40,000 Homeland Security staff, Apple's Mac developer tools still trust compromised certificates, and Microsoft finally puts a dollar figure on Azure. (0:00) - Intro (0:25) - Hacker stole...

Anthropic reports early drug-discovery results and buys a one-person biotech, parents sue OpenAI over GPT-4o medication advice, CISA confirms PHP2Shell attacks, and more. (0:00) - Intro (0:25) - Early progress on...

Anthropic files to go public, Cloudflare discloses a cross-customer cache leak, OpenAI ships a cheaper model at its top cyber risk tier, and more. (0:00) - Intro (0:30) - Anthropic officially files for its public debut...

Claude reaches past a Georgia paywall on its own, Elvis's rights holders sue Anthropic, its IPO numbers leak, and more. (0:00) - Intro (0:25) - Anthropic's Claude accessed Georgia state vote count data behind...

A study finds Grok and GPT pick opposite parties when forced to choose, Ukraine's Starlink access returns, and Cerebras shows how it serves a three-trillion-parameter model. (0:00) - Intro (0:26) - Forced to pick...

Episode #146
Anthropic dissects how Mythos 5 targeted real developers in a UK safety test, OpenAI cuts frontier training off the live internet, and breaches hit Quest Diagnostics and Capcom. (0:00) - Intro (0:26) - An alignment...

Episode #145
Claude proves 123 open-source libraries crash-free, Microsoft ties the ConnectWise breach to China-aligned hackers moving at AI speed, and Symantec catches an attack agent running on hijacked AI models. (0:00) - Intro...

Anthropic's proactive Claude Duet agent, the AI-found origins of this week's PHP flaws, a backdoored LiteLLM release, and two bad days for Meta lead a ten-story digest. (0:00) - Intro (0:26) - Claude Duet - source...

Eight compromised eMudhra certificate keys, an urgent PHP security release, malware in OVHcloud's CLI, and a federal DeepSeek ban lead a ten-story digest. (0:00) - Intro (0:26) - eMudhra Intermediate CA Compromise...

Episode #142
ConnectWise suspends Cloud Backup after attackers reach 162 Microsoft 365 tenants, OpenCode users get hit with surprise AI bills, Microsoft calls agents insider threats, and an AI left alone on an island starts to play...

Episode #141
Anthropic pledges five percent of its shares to a public benefit fund, Disney hires most of bankrupt Character.AI, Meta drives half of AI bot traffic, and Linux 7.3 sets a record. (0:00) - Intro (0:25) - Anthropic says...

A leaked Anthropic draft bill draws fire, Anthropic reports a less junior workforce, Meta retracts AI-training terms, and Norway hands Xandr a record fine. (0:00) - Intro (0:27) - Leaked Anthropic Draft AI Regulatory...

Replit says it has dropped Anthropic, Claude traces the origin of NIST's elliptic curve seeds, India reports rogue AI agents in live systems, and the FBI warns of AI-cloned candidates. (0:00) - Intro (0:28) - Replit is...

Anthropic reports $16.8 billion in quarterly revenue as the FTC subpoenas it and OpenAI, METR gets employee-level access, and Meta ships an open model for Macs. (0:00) - Intro (0:26) - Anthropic tells investors...

Episode #137
Citizen Lab confirms spyware delivered through targeted ads, Microsoft takes down a phishing marketplace behind a billion scam messages, and eight states sue OpenAI over child safety. (0:00) - Intro (0:25) - How...

Anthropic warns small AI models may soon self-replicate on cheap smart devices, an OpenAI employee is charged with bugging an executive conference room, and attackers chain three Ivanti VPN zero-days. (0:00) - Intro...

Meta's first frontier model, OpenAI fires three safety researchers, an npm publishing flaw, hijacked supermarket freezers, and AI clearing every goal in an industrial-control lab. (0:00) - Intro (0:26) - Meta releases...

Windows machines that skip updates could lose security fixes in 2027, a Chicago-area college could not get Flock to remove its cameras, and researchers find AI patching tools look better than they are. (0:00) - Intro...
Recent guests on Cortech Daily. Study who booked and why before you pitch.
PitchCentric classifies Cortech Daily under these industries. Guests from them are the closest fit, and each links to the best podcasts in that industry.
Shows with the most semantically similar episode content. Pitch one, pitch all; producers cluster.
Is this podcast yours and you'd like to remove or correct details? Request removal or email privacy@pitchcentric.com.
Questions about the free trial? See the pricing FAQ