In partnership with
☀️ TRENDING AI NEWS 🚨 Nvidia Probe: The board opened an independent investigation into CEO Jensen Huang after reports he was the target of a foreign intelligence effort. 🏢 Apple Siri: CEO John Ternus is reportedly pressuring Craig Federighi to speed up an agentic Siri that can compete with OpenAI's Dots and Meta's Muse. 🎵 Suno Sued: Elvis Presley rights holders and an Eminem publisher filed separate suits claiming Suno generates music and lyrics that mimic their artists. 🛠️ 0sec Labs: The open-source security harness '0' runs code review and pentesting locally via browser, CLI and MCP server. |
Seven days ago, the biggest question around Nvidia was how many chips it could ship. This morning, the question is whether its CEO survives a board investigation. The AI industry's power structure is getting reshuffled in ways nobody scripted.
| 🛠️ Tool of the day | |
Perplexity - An AI answer engine that gives you sourced answers to questions in plain language. It helps anyone who wants quick research without opening ten tabs.
| ⚡ Quick hits | |
A new SHL survey says 19.1% of US workers replaced by AI were later asked to return by their employers. Read more →
Microchip's shares fell 10% after its CEO said AI companies "took over the whole factory," stretching lead times to 6 to 32 weeks. Read more →
Apiiro found 2,539 secrets across 1,994 public repositories linked to the Claude Code "don't ask again" settings file. Read more →
🤓 AI Trivia
In a Mixture-of-Experts language model, what does the 'router' (or gate) decide?
The answer is hiding near the bottom of today's newsletter... keep scrolling. 👇
| 🏢 Nvidia's Board Investigates Its Own CEO | |
A suspected spy, an affair, a memo
On Sunday, Nvidia said its board is "initiating an independent investigation in light of recent reporting that founder and CEO Jensen Huang has been the target of a foreign intelligence effort." The New York Times reported in October that Trump had been briefed on intelligence about Huang, and the company's statement followed fresh reporting this weekend.
The center of the story is Megaspeed executive Alice Huang, who a US intelligence memo reportedly says was recruited by China's Ministry of State Security to spy on Nvidia's CEO. In a memo to staff, Jensen Huang admitted he had an affair. Yale's Jeffrey Sonnenfeld puts it at "about a 50-50" whether Huang keeps his job.
The bottom line
Nvidia sits at the center of the AI supply chain, so a leadership crisis tied to China reaches well past one company. Watch whether the board's findings land before the next chip-export decisions do.
You know the problem. Now build the fix.
You've spent years close to a broken industry or process. You know exactly what would fix it. But you’ve never had the time or resources to make it happen.
With Lovable, your expertise is all you need. Describe the solution to your problem — a website, a dashboard, an internal tool — and get professional-grade software back. You make every call on how it works, because you're the one who knows what it needs to do.
Start today and take $5 off your first month.
Discount of $5 applies to first month of any paid plan. You need to sign up via this webpage in order to claim the discount. The discount applies to new and returning users without an active paid plan. Terms apply.
| 🤖 Apple Is Racing to Make Siri Agentic | |
Pressure from the top, rivals pulling ahead
According to Bloomberg's Mark Gurman, CEO John Ternus is pushing AI czar Craig Federighi to speed up the Siri roadmap and ship a powerful AI agent that can compete with OpenAI's Dots or Meta's Muse.
Federighi has delivered a functional version of Siri AI on an incredibly short timeline, but Apple has no illusions that it's keeping pace with its rivals. Gurman reports two significant improvements are expected in the coming releases.
The bottom line
Apple's advantage has always been owning the device, but an assistant that can't act on your behalf is a liability when rivals ship agents. Expect Siri announcements to get a lot more urgent.
| ⚠️ Elvis and Eminem Rights Holders Sue Suno | |
Two new suits, two different legal theories
AI music company Suno is facing yet another pair of lawsuits. Rights holders of Elvis Presley's music and the publisher of some of Eminem's biggest records claim the service illegally generates music and lyrics that mimic those artists.
The Presley lawyers argue Suno violated Tennessee's ELVIS Act (Ensuring Likeness Voice and Image Security), a state law aimed squarely at AI voice and likeness cloning. It's a notable test of whether state-level likeness statutes can bite where copyright claims have moved slowly.
The pressure isn't limited to the US. The head of the global music body IFPI warned Australia's Albanese government that weakening AI copyright law to help AI companies would be "ridiculous and dangerous" and a "disaster for creativity." More on the broader fight over AI copyright is on our tag page.
The bottom line
Likeness laws give artists a second front beyond copyright. If the ELVIS Act claim holds up, other states' voice-protection statutes become real leverage against AI music tools.
| 🛠️ A Local AI Security Harness Goes Open Source | |
Pentesting and code review, on your own model
0sec Labs has open-sourced 0, an AI security harness for code review, pentesting, reverse engineering and fixes. It runs locally through a browser console, a headless CLI and an MCP server, and you bring your own model provider.
The timing fits a rough few weeks for defenders. The Guardian rounded up recent incidents, from a fake nuclear alert to a $6bn heist, and some experts say AI exposing flaws in IT systems means cybersecurity events like these could soon be common. Meanwhile, if you're launching a site to showcase a project like this, our sponsor 60sec.site is an AI website builder that gets you live fast. And for more daily AI coverage, visit dailyinference.com.
The bottom line
Running an AI security tool locally with your own model keeps sensitive code off third-party servers. For small teams without a security budget, that's a genuinely useful starting point.
| 🔬 Inside Mixture-of-Experts: How 10 Open Models Route Tokens | |
Same idea, very different gating choices
MarkTechPost published a config-verified comparison of how 10 open MoE models route tokens, covering total and active parameters, expert counts, top-k, shared experts, gating functions and load-balancing methods. Models in the mix include DeepSeek, Qwen, Mixtral, Llama 4 and gpt-oss.
A companion guide explains seven routing methods, from Mixtral's top-2 softmax to DeepSeek-V4's hash routing. The router is the small component that picks which experts handle each token, and it's where much of the efficiency (and the engineering headaches) of these models live.
The bottom line
If you're choosing an open model to self-host, active parameter count and routing design affect speed and memory more than the headline size does. These two references make that comparison easy.
🌎 Trivia Reveal
The answer is: the router decides which experts process each token. Only the chosen experts activate for a given token, which is how MoE models keep active parameters far below their total size.
💬 Quick Question
Would you trust an AI agent from Apple's Siri to act on your behalf, or would you rather it only answer questions? Hit reply and let me know - I read every response!
That's it for today - see you tomorrow with more.
| 🎁 Share Daily Inference | |
Refer a friend to get the AI Tools Starter Kit. Refer 5 to unlock the AI Insider Briefing.