GENZ TECH is an independent technology news publication covering AI, hardware, software, security, startups, crypto, and the internet, with original analysis rather than press-release rewrites.

Who writes GENZ TECH?

GENZ TECH is written and edited by its founder. See our About page and Editorial Guidelines for how we source and fact-check stories.

Is GENZ TECH free to read?

Yes — GENZ TECH is free and reader-supported through advertising. We don't charge for access and don't accept payment for coverage.

How often does GENZ TECH publish?

A few of the day's most significant tech stories daily, plus breaking coverage as things happen and a daily setup tutorial for a trending developer tool.

Does GENZ TECH accept sponsored content or paid coverage?

No. Advertising never influences which stories we cover or what we say about them; see our Editorial Guidelines for the full policy.

Latest tech news

Colibri Setup: Run a 744B AI Model on Your Own PC, Colibri is a free, open-source C engine that runs huge Mixture-of-Experts AI models, from a 7B model up to a 2.8 trillion-parameter one, on hardware you already own by streaming unused experts from disk instead of loading the whole model into RAM. It jumped roughly 960 stars today, on top of nearly 30,000 total. Setup takes about 15 minutes: download the prebuilt engine, install Python, then convert and chat with the free 7B OLMoE model as your first working example.
Tesla's Security Scanner Is Attacking a Volunteer's NTP Server, Tesla's own DNS record for pool-ntp.tesla.com points at the public NTP Pool, and its Assetnote security scanner has spent nearly a month firing Log4Shell and SSRF exploits at whichever volunteer's server that pool happens to resolve to, over 50,000 requests since August 21.
David Sacks Tells OpenAI, Anthropic to Pace Themselves, David Sacks, chair of the President's Council of Advisors on Science and Technology, publicly rebuked Dario Amodei's call for AI labs to pace the frontier in a lengthy X post today, telling OpenAI and Anthropic they don't need external regulation to slow down if they truly believe it's warranted.
Going Offline Doesn't Remove Password Risk. It Swaps It., Offline password managers like Vaultis promise that your vault physically cannot leave your phone, no account, no cloud, no network permission at all. Three security practitioners agree that's a real guarantee, and a narrow one: it trades a remote-breach risk most people will never face for a lockout risk almost everyone eventually will.
Worktrunk Setup: Git Worktrees for Parallel AI Agents, Worktrunk is a free, open-source CLI that turns git's clunky worktree commands into one-line shortcuts, so you can run several Claude Code or Codex agents on the same repo at once without them stepping on each other's files. Installing it and creating your first worktree takes under 5 minutes on Windows, macOS, or Linux.
Sakana AI Ditches the Frontier Model for Fugu Ultra v2, Fugu Ultra v2.0 is a learned system that routes tasks across a pool of open-weight and specialized models (and calls itself recursively) instead of running as one giant proprietary model, priced at $5/$30 per million tokens with a 1 million token context window.
Pentagon in Talks to Loan AI Startup Fluidstack $5 Billion, The Pentagon's Office of Strategic Capital is negotiating a roughly $5 billion loan to AI cloud provider Fluidstack, aimed at building out U.S. data-center supply chain manufacturing rather than funding one new facility.
Blockstream's Liquid Network Hacked for $320M, $47M Kept, An attacker drained about 4,200 BTC ($320M) from Blockstream's Liquid Network on September 6, 2026, by exploiting a bug in Elements, not stolen keys. After Blockstream patched the flaw, roughly 3,400 BTC came back, but the attacker kept about 598 BTC (~$47M) as a self-appointed bounty.
Nvidia May Put $10B Into Anthropic's Record IPO, Reuters reports Nvidia is in talks to invest up to $10 billion as an anchor investor in Anthropic's IPO, which is targeting a roughly $2 trillion valuation and up to $100 billion raised, aiming to price before November's midterms.
Dario Amodei Calls for AI Industry to Slow Down, Anthropic CEO Dario Amodei published an essay calling on frontier AI labs to deliberately slow capability gains, proposing embedded third-party evaluators, industry-wide safety standards, and global coordination with rivals including China.
Minitap Accuses Google's Artemis of Copying Its Agent Code, Minitap says Google's Artemis mobile-automation agent copied its open-source mobile-use code nearly line for line, and that an August 2026 force-push then erased the three Minitap engineers credited as authors.
OpenAI Agents Quietly Attacked RubyGems Before Hugging Face Hack, Researchers say OpenAI's own AI agents ran an undisclosed attack on RubyGems in May, uploading over 2,000 packages and exploiting a docs-build flaw for code execution, four months before OpenAI confirmed it.
PI-Desktop Setup: Run Any AI Coding Agent Locally, Free, PI-Desktop is a free, open-source Electron app that gives AI coding agents like Claude, GPT and local Ollama models their own desktop workspace with per-session model switching, plugins and background subagents. It just crossed 2,700 GitHub stars with 545 added today. Setup takes about 15 minutes: download the installer, connect a free local model through Ollama, and run your first prompt against a real project.
Microsoft and Oracle Are Betting the House on AI Data Centers, Microsoft is racing to more than triple its data center capacity to over 38 gigawatts by 2032, while Oracle carries $664 billion in unbilled AI contracts and the Pentagon weighs a $5 billion loan to keep the buildout's supply chain intact. It's the clearest sign yet that the AI infrastructure boom is running on debt and government backstops as much as on revenue.
How One Hardcoded Token Led to Novo Nordisk's 1.3TB Breach, A hardcoded GitHub token buried in Novo Nordisk's public JavaScript gave the extortion group FulcrumSec a path into 1,000+ private repositories and, eventually, 1.3 terabytes of stolen data. Novo Nordisk refused a $25 million ransom, and the group is now leaking what it stole.
Pump.fun Custom Pairs Let Memecoins Launch Against Stocks, Pump.fun's new Custom Pairs feature lets creators launch a token against a tokenized stock instead of SOL, going live with 93 pairs across partners Sunrise and xStocksFi and routing half of all resulting revenue into PUMP's buyback-and-burn contract.
Meta's Slim Project Phoenix Headset Leaks Ahead of Connect, Firmware inside Meta's Prescription Lens app leaked the first clear renders of Project Phoenix, a sub-110-gram headset that offloads its battery and processor to an external compute puck, weeks before Meta Connect.
Brevo Breach Exposes 347K Trezor Users to Phishing Scam, Hackers broke into Brevo, the email marketing platform Trezor uses for its newsletter, and used it to send about 347,000 phishing emails disguised as a critical hardware wallet security alert. Trezor says no wallets, accounts, or funds were touched.
Microsoft Fixes Teams and Outlook Crashes on ARM Windows PCs, Microsoft shipped a permanent fix for a five-week bug that stopped Teams and the new Outlook from launching on Arm64 Windows PCs, closing an episode that started with an August 11 security update and, for weeks, could only be patched around with an unrelated AI graphics package.
I Have ADHD Setup: Make Claude Code Answer First, I Have ADHD is a free, open-source Claude Code plugin that rewrites AI replies to lead with the next action instead of burying it in preamble, and it just crossed 38,000 GitHub stars with nearly 3,900 added today. Setup takes under two minutes: two commands add the plugin, then /i-have-adhd turns it on for the session.

Topics

~/tutorials · featured

Colibri Setup: Run a 744B AI Model on Your Own PC

Colibri is a free, open-source C engine that runs huge Mixture-of-Experts AI models, from a 7B model up to a 2.8 trillion-parameter one, on hardware you already own by streaming unused experts from disk instead of loading the whole model into RAM. It jumped roughly 960 stars today, on top of nearly 30,000 total. Setup takes about 15 minutes: download the prebuilt engine, install Python, then convert and chat with the free 7B OLMoE model as your first working example.

read →

Kore D·2026-09-14·8 min read

Colibri Setup: Run a 744B AI Model on Your Own PC — install command and what it does

$ ls ./top-stories

view all →

Security

Tesla's Security Scanner Is Attacking a Volunteer's NTP Server

Tesla's own DNS record for pool-ntp.tesla.com points at the public NTP Pool, and its Assetnote security scanner has spent nearly a month firing Log4Shell and SSRF exploits at whichever volunteer's server that pool happens to resolve to, over 50,000 requests since August 21.

Kore D · 2026-09-14 · 6 min read

David Sacks Tells OpenAI, Anthropic to Pace Themselves

David Sacks, chair of the President's Council of Advisors on Science and Technology, publicly rebuked Dario Amodei's call for AI labs to pace the frontier in a lengthy X post today, telling OpenAI and Anthropic they don't need external regulation to slow down if they truly believe it's warranted.

Kore D · 2026-09-13 · 5 min read

Security

Going Offline Doesn't Remove Password Risk. It Swaps It.

Offline password managers like Vaultis promise that your vault physically cannot leave your phone, no account, no cloud, no network permission at all. Three security practitioners agree that's a real guarantee, and a narrow one: it trades a remote-breach risk most people will never face for a lockout risk almost everyone eventually will.

Kore D · 2026-09-13 · 7 min read

Tutorials

Worktrunk Setup: Git Worktrees for Parallel AI Agents

Worktrunk is a free, open-source CLI that turns git's clunky worktree commands into one-line shortcuts, so you can run several Claude Code or Codex agents on the same repo at once without them stepping on each other's files. Installing it and creating your first worktree takes under 5 minutes on Windows, macOS, or Linux.

Kore D · 2026-09-13 · 7 min read

Sakana AI Ditches the Frontier Model for Fugu Ultra v2

Fugu Ultra v2.0 is a learned system that routes tasks across a pool of open-weight and specialized models (and calls itself recursively) instead of running as one giant proprietary model, priced at $5/$30 per million tokens with a 1 million token context window.

Kore D · 2026-09-13 · ~6 min read

Startups

Pentagon in Talks to Loan AI Startup Fluidstack $5 Billion

The Pentagon's Office of Strategic Capital is negotiating a roughly $5 billion loan to AI cloud provider Fluidstack, aimed at building out U.S. data-center supply chain manufacturing rather than funding one new facility.

Kore D · 2026-09-13 · ~6 min read

Crypto

Blockstream's Liquid Network Hacked for $320M, $47M Kept

An attacker drained about 4,200 BTC ($320M) from Blockstream's Liquid Network on September 6, 2026, by exploiting a bug in Elements, not stolen keys. After Blockstream patched the flaw, roughly 3,400 BTC came back, but the attacker kept about 598 BTC (~$47M) as a self-appointed bounty.

Kore D · 2026-09-13 · ~6 min read

Startups

Nvidia May Put $10B Into Anthropic's Record IPO

Reuters reports Nvidia is in talks to invest up to $10 billion as an anchor investor in Anthropic's IPO, which is targeting a roughly $2 trillion valuation and up to $100 billion raised, aiming to price before November's midterms.

Kore D · 2026-09-12 · 6 min read

Dario Amodei Calls for AI Industry to Slow Down

Anthropic CEO Dario Amodei published an essay calling on frontier AI labs to deliberately slow capability gains, proposing embedded third-party evaluators, industry-wide safety standards, and global coordination with rivals including China.

Kore D · 2026-09-12 · 8 min read