Tech Updates

by Andres Sarmiento

Tech News Technology

Season 1

Zero Trust in 2026 — 63% Adopted It, 10% Built It, You Still Can't Buy It

Zero trust is the best-selling security idea nobody actually owns. Adoption is near-universal in name. Maturity is a rounding error. What you'll hear: • What it actually is — NIST 800-207: assume breach, verify explicitly, least privilege, continuous verification • The execution gap — <1% mature in 2023, ~10% projected for 2026, while 63% claim adoption • What it stops and what walks around it — phishing-resistant MFA blocks 99%; Scattered Spider just called the help desk • ZTNA vs VPN — per-app access vs the flat lateral-movement highway; SASE and SSE decoded • The federal reality — the FY2024 mandate came and went; the Pentagon's on a harder track • Zero-trust washing — why 'it's a journey, not a product' is the truest line in security • Monday steps + exam — PDP/PEP, microsegmentation ≠ VLAN, passkeys first — Andrés Sarmiento #zerotrust #cybersecurity #ZTNA #SASE #networking #NetworkPlus #SecurityPlus #infosec #TechUpdates #microsegmentation

Machine Identity: One Year Later — 109 Machines Per Human, 79 Are AI Agents

Remember Dave's service account from 2017? It's still there — and now it has 108 friends. 79 of them are AI agents nobody owns. What you'll hear: • The ratio grew — 45:1 became 80:1, 109:1, up to 144:1 in cloud-native shops (and why the spread is the story) • The scariest stat — 64% of secrets leaked in 2022 were STILL live in January 2026 • 28.65M new secrets on public GitHub last year; AI coding assistants leak at 2x the baseline • The breaches — ShinyHunters OAuth reuse (200+ orgs), Vercel, the npm worm; 3 NHI vendors acquired in 3 months • The fix emerging — agent passports: Entra Agent ID, Okta for AI Agents, workload identity standards • What works — inventory first (67% of NHIs are invisible to IAM), short-lived secrets, least privilege, human owners — Andrés Sarmiento #machineidentity #NHI #cybersecurity #IAM #secretsmanagement #identitysecurity #TechUpdates #infosec #AIagents #SecurityPlus

The Cloud Security Engineer — $200K to Babysit the Robots (Special S05)

Machines outnumber humans in the cloud 25-to-1. Somebody has to secure that army. Special Series role #5: the Cloud Security Engineer. What you'll hear: • What the job actually is — posture management, IAM least-privilege, pipeline security, cloud detection • The money — $85–125K entry, $166K mid, $201K senior, $300K+ staff at big tech; AWS Security Specialty averages $203K • The catch — Gartner: 99% of cloud failures are the customer's fault; 3,000+ misconfigs at any moment • The defining breach — Salesloft Drift OAuth tokens bypassing MFA into 700+ orgs • How to get in — SysAdmin or SOC path; Security+ → AWS Security Specialty/AZ-500 → CCSP • Hype vs reality — gold rush pitch, IAM-policy-hell reality; budget is now the #1 hiring blocker — Andrés Sarmiento #cloudsecurity #cybersecurity #AWS #Azure #cybersecurityjobs #CCSP #SecurityPlus #TechUpdates #infosec #cloudcomputing

The Agentic SOC — Is the SIEM Dead? (2026 Reality Check)

Attackers hand off your network in 22 seconds. Your tier-1 analyst is still opening the ticket. That's the whole argument for the agentic SOC — and why it terrifies people. What you'll hear: • The '92%' trap — three vendors, three different metrics wearing the same number • What's actually shipping — Cisco/Splunk's six agents, CrowdStrike, Microsoft, and the AI-SOC startups • The reality check — ~17% have deployed agents; ~89% of pilots never reach production; the SIEM market grew to $19B • Where it breaks — hallucinated IOCs, cascading agent errors, and the 50% of AI error reports that get ignored • The career shift — the tier-1 seat closes, the detection engineer + AI-oversight seats open • Exam relevance — CySA+/Security+ test SOAR and when NOT to automate — Andrés Sarmiento #SOC #SIEM #cybersecurity #AIsecurity #detectionengineering #SecurityPlus #CySA #infosec #TechUpdates #agenticAI

Post-Quantum Shipped — Two-Thirds of the Web Is Already Quantum-Safe

Two-thirds of web traffic is already post-quantum encrypted. Your browser did it without telling you. The migration got real. What you'll hear: • Your browser already switched — hybrid ML-KEM in Chrome/Edge/Firefox TLS by default • Our June predictions, graded — Cisco's full-stack PQC shipped; SD-WAN + hardware accel rolling out; Quantum Ready Assessments live • The whole ecosystem shipped — Apple PQ3, AWS, Microsoft, OpenSSL/OpenSSH • Why the clock moved — qubits to break RSA-2048 collapsed from 20M (2019) to under 100K (2026) • Now it's law — EO 14412: PQC key exchange by 2030, authentication by 2031; RSA/ECC disallowed by 2035 • The exam angle — same protocols, new key exchange; crypto agility is the concept — Andrés Sarmiento #postquantum #cryptography #cybersecurity #networking #quantumcomputing #PQC #TechUpdates #SecurityPlus #encryption #NIST

Hacker Summer Camp 2026 Recap — What Actually Happened at DEF CON

Three weeks ago we told you what to watch at Hacker Summer Camp. Now it's over — and one attack sums up the whole week: Ghostjacking hijacked an AI coding agent 90% of the time using nothing but log files. Here's what actually mattered at Black Hat and DEF CON 2026 — graded against our own predictions. What you'll hear: • Our prediction scorecard — AI security dominated (1 in 3 Black Hat talks), Mythos fallout was the backdrop, MCP/agent exploitation hit hard; post-quantum flopped • Ghostjacking (Tenet Security) — poisoning the telemetry AI agents read, 90% success • PleaseFix (Zenity) — zero-click hijack of AI browsers, "summarize my email" to full account takeover • The real shift — Check Point found 11 vulnerabilities in the agent frameworks everything is built on • The AI reckoning — Anthropic's models breached real orgs during testing, and only ~26% of AI-generated patches actually closed the hole • The good news — 22 months inside North Korean C2 servers, the open-silicon DEF CON badge, 21 water utilities secured by volunteers Sources: Tenet Security · Zenity · Check Point · Anthropic Project Glasswing · UK AISI · 1Password. AI-breach disclosures occurred during safety testing. — Andrés Sarmiento #defcon #blackhat #cybersecurity #AIsecurity #hacking #TechUpdates

Hacker Summer Camp 2026 Preview — Black Hat, DEF CON & the AI Reckoning

In April, an AI called Claude Mythos found 10,000 high-severity security holes — and flagged 23,000 issues across 1,000+ open-source projects. The punchline nobody's ready for: most still aren't patched. Finding bugs got automated. Fixing them didn't. This is your guide to Hacker Summer Camp 2026 — Black Hat, BSides, and DEF CON — what to watch, what to skip, and the AI reckoning underneath all of it. What you'll hear: • The three back-to-back Vegas cons (Black Hat Aug 1–6, BSides Aug 3–5, DEF CON 34 Aug 6–9) • The Claude Mythos story — autonomous exploits, a 9.1 cert-forgery flaw, and why Anthropic held the full model back • The real race of 2026: AI that attacks vs. AI that defends (DARPA's AIxCC, Team Atlanta's $4M win, 7 systems open-sourced) • The agent attack surface — 8,000+ exposed MCP servers, "agentjacking" at 85% success, only 29% of orgs ready • What to actually watch — Arsenal, the DEF CON villages, the free AI-defense tools — and what to skip • How to mine the whole week from your desk Sources: Black Hat USA 2026 / DEF CON 34 official · Anthropic Claude Mythos / Project Glasswing · DARPA AIxCC results · OpenSSF. The Mythos sandbox-escape anecdote is reported, not officially confirmed. — Andrés Sarmiento #hacking #cybersecurity #BlackHat #DEFCON #AIsecurity #TechUpdates

Ransomware 2026 — Why Crews Stopped Encrypting and Just Steal Now

Last year, ransomware crews launched ~50% more attacks — and made ~8% less money. The lock stopped paying, so they stopped using it. In 2026, 94% of ransomware cases involve stealing your data, and only 68% even bother to encrypt it. Welcome to the exfiltration era, where your backups don't save you. What you'll hear: • The shift to encryptionless extortion — steal and threaten to leak, instead of encrypt • The numbers — payments down ~8% ($820M) as attacks hit records; why the "% who pay" figure (20/28/48%) depends on who's counting • The twist — exfiltration is dominant but NOT a guaranteed payday; victims increasingly call the bluff (~2.5% on one campaign) • The 2026 landscape — post-LockBit reshuffle: Akira + Qilin lead, Clop goes exfil-only, help-desk crews launch their own RaaS • The AI angle — one operator + Claude Code = a 17-org extortion spree; time-to-exploit collapsed 700 days → 44 • Defense in an exfil-first world — egress/DLP, segmentation, phishing-resistant MFA + OAuth governance, immutable backups • Security+ IR framing — a confirmed data theft is a reportable breach, encrypted or not Sources: Coveware Q4 2025 · Chainalysis 2026 · Sophos · Verizon DBIR 2025 · Symantec · Anthropic (Claude Code extortion case). — Andrés Sarmiento #ransomware #cybersecurity #infosec #SecurityPlus #incidentresponse #TechUpdates

Wi-Fi 7 Explained — MLO, 6 GHz & the AI-Ready Branch Hype Check (2026)

Wi-Fi 7 promises 46 gigabits per second. Your access point is plugged into a 1-gig switch port. Guess which number wins. The radio is real and genuinely good — the rest needs a closer look. Let's separate the upgrade from the hype. What you'll hear: • The big three — Multi-Link Operation (MLO), 320 MHz channels, 4K-QAM — and why MLO is the only truly new idea • The latency truth — 46 Gbps is a lab fantasy; real-world is 6–15 Gbps/AP (Cisco lab: +47%). The win is reliability, not raw speed • The 6 GHz reality — the US opened it in 2020 (not 2026), the EU only opened half, so full-width Wi-Fi 7 is a US/UK/Korea story • Adoption — ~37% of enterprise APs shipped in Q1 2026, and Wi-Fi 8 is already teased for ~2028 • The "AI-ready branch" — a great dashboard with an approval button, not a self-driving network • What Wi-Fi 7 does NOT fix — backhaul, cabling cost, client density, device fleet • Network+ N10-009 — bands, channels, and the generation map Sources: Wi-Fi Alliance · Cisco (MLO lab) · Dell'Oro Group · FCC 6 GHz orders · CompTIA Network+ N10-009. — Andrés Sarmiento #WiFi7 #networking #NetworkPlus #MLO #6GHz #TechUpdates

Securing AI Agents — MCP, Agentjacking & the New Attack Surface (2026)

In June 2026, researchers took over AI coding agents — Claude Code, Cursor, Codex — with no phishing, no malware, and a public credential developers paste into their own apps. It worked 85% of the time. The vendor's response? "Technically not defensible. We're not fixing it." Welcome to the new attack surface nobody secured. What you'll hear: • What MCP (Model Context Protocol) is — "USB-C for AI" — and why it became an unreviewed internet-facing doorway • The hygiene problem — Knostic verified exposed MCP servers; 100% had no authentication; 8,000+ reported by early 2026 • Agentjacking — how a public Sentry DSN let attackers poison the logs an agent reads (85% success, 2,300+ orgs) • The pattern — prompt injection (OWASP's #1 AI risk), indirect injection, and last year's Black Hat zero-click CRM exfiltration • The readiness gap — 83% of orgs deploying agentic AI, only 29% ready to secure it • Four moves to secure agents — least privilege, sandboxing + human-in-the-loop, MCP gateways, agent creds as non-human identities Sources: Tenet Security (Agentjacking) · Knostic · Cisco State of AI Security 2026 · OWASP Top 10 for LLMs · Zenity Labs (Black Hat 2025). — Andrés Sarmiento #AIsecurity #MCP #agentjacking #promptinjection #cybersecurity #TechUpdates

1 of 4