David Nowak
LightNews LightNews
banner
davidnowak.me
David Nowak
@davidnowak.me
80 followers 41 following 1.7K posts
I bridge technical expertise with human understanding. Built solutions for millions. I help organizations question assumptions before costly mistakes. Connecting dots, creating impact.

🌐 davidnowak.me
🗞️ thestrategiccodex.com
Posts Replies Media Videos
davidnowak.me
ICE is deploying facial recognition apps that agents point at teenagers' faces—no warrant, no consent, stored 15 years. Plus spyware that hijacks phones via text. This isn't immigration enforcement. It's infrastructure for total control... 🧵
www.npr.org/2025/11/08/n...
Immigration agents have new technology to identify and track people
The Department of Homeland Security is adopting powerful new tools to monitor noncitizens. Privacy advocates are worried they erode privacy rights for all Americans.
www.npr.org
November 12, 2025 at 1:49 AM
davidnowak.me
Google just analyzed five AI-generated malware families, and here's the gap between hype and reality: they're embarrassingly bad. Easy to detect, missing basic features, no novel techniques... 🧵
arstechnica.com/security/202...
5 AI-developed malware families analyzed by Google fail to work and are easily detected
You wouldn’t know it from the hype, but the results fail to impress.
arstechnica.com
November 11, 2025 at 7:40 PM
davidnowak.me
Boeing paid $6.1B in losses. Wells Fargo's stock never recovered after $3B in fines. VW paid $33.3B in settlements. Legal compliance didn't save them. They lost something you can't buy back: permission from the people affected by their business... 🧵
davidnowak.me/when-compani...
When Companies Forget They Need Permission - DAVID NOWAK
Your business has a legal license, but does it have permission to operate? Wells Fargo and Boeing learned the difference the hard way, costing them billions. This isn't just theory; it's a repeatable ...
davidnowak.me
November 11, 2025 at 1:30 PM
davidnowak.me
Denmark just joined Australia in banning social media for kids under 15—the most ambitious global pushback yet against platforms designed to capture young attention. But here's what caught me: enforcement? They haven't figured that out... 🧵
www.engadget.com/social-media...
Denmark set to ban social media for users under 15 years of age
The Danish government is set to ban social media for users under 15 years of age.
www.engadget.com
November 11, 2025 at 1:31 AM
davidnowak.me
Google's AI weather model just lapped the US government's GFS in hurricane forecasting. We're watching a private tech company solve what government agencies—with bigger budgets—couldn't. That tension deserves real attention... 🧵
arstechnica.com/science/2025...
Google’s new hurricane model was breathtakingly good this season
Meanwhile, the US Global Forecasting System continues to get worse.
arstechnica.com
November 10, 2025 at 8:11 PM
davidnowak.me
AI is screening for tuberculosis in places with zero radiologists—nomadic camps in Nigeria, refugee settlements in Chad. A mother in Mali got her TB diagnosis in seconds, not weeks. It's working. And it raises questions... 🧵
www.npr.org/sections/goa...
AI steps in to detect the world's deadliest infectious disease
There's a global shortage of radiologists. Now artificial intelligence is helping speed up the diagnosis of tuberculosis in hard-to-reach communities.
www.npr.org
November 10, 2025 at 1:31 PM
davidnowak.me
Character.AI just banned everyone under 18 from their chatbots after teen suicides. The move raises a question nobody wants to ask: If your product is unsafe for a third of the human lifespan, is it safe at all? 🧵
arstechnica.com/information-...
After teen death lawsuits, Character.AI will restrict chats for under-18 users
AI companion app faces legal and regulatory pressure over child safety concerns.
arstechnica.com
November 10, 2025 at 1:38 AM
davidnowak.me
The Snap-Perplexity deal is fascinating for what it really says about how search is evolving. $400M isn't just about integrations—it's about recognizing that answers live where conversations happen. Early 2026 launch... 🧵
www.engadget.com/social-media...
Snap and Perplexity sign $400 million deal to put AI search directly in Snapchat
Snap and Perplexity AI have struck a $400 million deal that will bring the AI search engine directly to Snapchat sometime in "early 2026," the two companies announced.
www.engadget.com
November 9, 2025 at 7:42 PM
davidnowak.me
Anthropic's introspection research just dropped, and the headlines are spinning it as "AI self-awareness." But here's what matters: they've shown we can actually detect when AI models lie about their own reasoning. That changes how we build trustworthy systems... 🧵
arstechnica.com/ai/2025/11/l...
LLMs show a “highly unreliable” capacity to describe their own internal processes
Anthropic finds some LLM “self-awareness,” but “failures of introspection remain the norm.”…
arstechnica.com
November 9, 2025 at 2:03 PM
davidnowak.me
Apple paid $1B to Google for Gemini because it couldn't build competitive AI on its own—and the tech industry noticed. When a company known for flawless execution starts admitting defeat quietly, that tells you something deeper about where it stands... 🧵
www.engadget.com/ai/apple-wil...
Apple will reportedly use a custom version of Gemini to power the new Siri
Apple's updated Siri assistant will reportedly rely on a version of Gemini for some of its new functionality at launch.
www.engadget.com
November 9, 2025 at 1:38 AM
davidnowak.me
A $38B compute deal between OpenAI and Amazon just crystallized what many suspected: two companies with deep accountability deficits now intertwined at the infrastructure layer. The public sees it... 🧵
arstechnica.com/ai/2025/11/o...
OpenAI signs massive AI compute deal with Amazon
Deal will provide access to hundreds of thousands of Nvidia chips that power ChatGPT.
arstechnica.com
November 8, 2025 at 7:46 PM
davidnowak.me
The UK just ruled on AI copyright training—and the outcome exposes something most people missed. Getty won on trademark, lost on what mattered. But here's the real story: the court never actually decided if training AI on copyrighted images is legal. 🧵
apnews.com/article/gett...
Stability AI largely wins UK court battle against Getty Images over copyright and trademark
Stability AI has mostly prevailed against Getty Images in a British court battle over intellectual property.
apnews.com
November 8, 2025 at 2:00 PM
davidnowak.me
The Internet Archive just survived billion-dollar copyright lawsuits. Brewster Kahle says "we survived, but it wiped out the library." 500,000+ books gone. Here's what happened. 🧵
arstechnica.com/tech-policy/...
Internet Archive’s legal fights are over, but its founder mourns what was lost
“We survived, but it wiped out the library,” Internet Archive’s founder says.
arstechnica.com
November 8, 2025 at 1:16 AM
davidnowak.me
346 people dead. Boeing's guilt established. Then... silence. A federal judge just approved dropping all criminal charges, saying himself it "fails to secure necessary accountability." How did we get here? 🧵
www.npr.org/2025/11/07/n...
Boeing won't face criminal charge over 737 Max crashes that killed hundreds of people
As part of a deal to dismiss the case, Boeing agreed to pay or invest an additional $1.1 billion in fines, compensation for the crash victims' families, and internal safety and quality measures.
www.npr.org
November 7, 2025 at 1:46 PM
davidnowak.me
The human brain runs on roughly 20 watts—your toaster uses more. Yet it outthinks our best AI at a fraction of the energy cost. Surrey researchers just cracked why. Their topographical sparse mapping mimics how brains wire efficiently. Worth paying attention... 🧵
www.bbc.com/news/article...
University of Surrey researchers mimic brain wiring to improve AI
Researchers say this new method rethinks how AI systems are wired at their most fundamental level.
www.bbc.com
November 7, 2025 at 2:54 AM
Reposted by David Nowak
docnicola.bsky.social
My agent will talk to your agent: Multi agent research raises new questions about how well AI agents will perform when working unsupervised techcrunch.com/2025/11/05/m...
Microsoft built a fake marketplace to test AI agents — they failed in surprising ways | TechCrunch
The research raises new questions about how well AI agents will perform when working unsupervised — and how quickly AI companies can make good on promises of an agentic future.
techcrunch.com
November 6, 2025 at 1:42 PM
davidnowak.me
Polyurethane degradation just got cracked via neural network design. Researchers couldn't find natural enzymes, so they built computational tools to design one from first principles. The real insight? We're learning to ask different questions now... 🧵
arstechnica.com/science/2025...
Neural network finds an enzyme that can break down polyurethane
Given a dozen hours, the enzyme can turn a foam pad into reusable chemicals.
arstechnica.com
November 6, 2025 at 1:39 PM
Reposted by David Nowak
markriedl.bsky.social
OpenAI updates its terms of service for ChatGPT: can’t use the service for “tailored advice that requires a license, such as legal or medical advice, without appropriate involvement by a licensed professional.”

www.ctvnews.ca/sci-tech/art...
ChatGPT users can’t use service for tailored legal and medical advice, OpenAI says
OpenAI dispelled suggestions that it’s changing its terms around legal and medical advice.
www.ctvnews.ca
November 6, 2025 at 2:32 AM
davidnowak.me
YouTube just announced "voluntary" buyouts as part of an AI-first reorganization. I've seen this pattern before—and it rarely ends well for the people who built these platforms or the communities depending on them... 🧵
www.engadget.com/big-tech/you...
YouTube is offering employees buyouts as part of an AI-focused reorganization
As part of an AI-focused reorganization, YouTube CEO Neal Mohan told employees that it will offer voluntary buyouts.
www.engadget.com
November 6, 2025 at 1:20 AM
davidnowak.me
The AI authorship crisis isn't about cheating—it's about cognitive offloading. MIT found students using ChatGPT showed reduced brain activity while writing and couldn't recall what they'd "written." Learning dissolved... 🧵
theconversation.com/where-does-h...
Where does human thinking end and AI begin? An AI authorship protocol aims to show the difference
Students – and all manner of professionals – are tempted to outsource their thinking to AI, which threatens to undermine learning and credibility. A philosophy professor offers a solution.
theconversation.com
November 5, 2025 at 1:55 PM
davidnowak.me
OpenAI just released gpt-oss-safeguard—reasoning models for content moderation that let you write your own policies. Sounds promising. But dig into the research and you find something else entirely... 🧵
openai.com/index/introd...
Introducing gpt-oss-safeguard
New open safety reasoning models (120b and 20b) that support custom safety policies.
openai.com
November 5, 2025 at 1:21 AM
Reposted by David Nowak
thestrategiccodex.com
Supply chain attacks are like dominoes—breach one vendor, and the rest fall. Your payment processor gets hit. Your email host gets compromised. Their security gaps become your ransomware. 41.4% of ransomware now starts through third parties. You can't defend what you don't control... 🧵
November 4, 2025 at 11:49 PM
davidnowak.me
Google just lost an antitrust case. Then they won compliance theater. The court said "open the Play Store." Google opened it—then built a verification system that gates 95% of Android devices. Security framing, gatekeeping function. The irony cuts deep... 🧵
davidnowak.me/the-quiet-lo...
The Quiet Lock Behind Android’s Open Door - DAVID NOWAK
Google complied with a court order to open Android—but built new gatekeeping into the infrastructure. Starting 2026, developers need Google verification and a $25 fee to publish to 95% of Android devi...
davidnowak.me
November 4, 2025 at 2:01 PM
davidnowak.me
New research shows something unexpected: our social media landscape is quietly collapsing. Not with a bang, but through mass withdrawal—especially among the young and old. The implications run deeper than headlines suggest... 🧵
arxiv.org/abs/2510.25417
Shifts in U.S. Social Media Use, 2020-2024: Decline, Fragmentation, and Enduring Polarization
Using nationally representative data from the 2020 and 2024 American National Election Studies (ANES), this paper traces how the U.S. social media landscape has shifted across platforms, demographics,...
arxiv.org
November 4, 2025 at 1:28 AM