Good morning ☀️, leader of the next generation.

Another day, another ridiculous amount of AI news.

New models. New startups. Billion-dollar deals. Breakthroughs, failures, controversies — and occasionally something that makes you wonder what year we’re actually living in.

You don’t need to follow all of it.

That’s our job.

Here’s what matters today. ⚡

WHAT'S AT STAKE TODAY ⚡

• Google releases a new local-first Granola competitor 🗒️
• OpenAI's math solutions aren't meeting the field's standards yet ➗
• China's Manus raises over $500M in first funding round since split with Meta 🇨🇳
• Nous Research confirms it hit $1.5B valuation, launches AI agents for business users 💼
• Cal AI's 19-year-old founder just raised $10M for his new AI startup 🧑
• Goodfire says its new 'inside-out' monitors catch rogue AI agents at a fraction of the cost 🔍
• Natura's $99 smart ring puts AI agents on your finger 💍
• Anthropic changes usage policy to ban model abuse and election interference 🗳️
• Google brings agentic AI to Gemini, starting with businesses 🏢

Google launches an offline AI meeting note-taker to rival Granola.

Google releases a new local-first Granola competitor

Google has launched a new AI-powered meeting note-taker called Google AI Edge Foresight, designed to compete directly with popular tools like Granola. The Mac app comes from the same team that released an AI dictation tool earlier this year and is built to run entirely on-device, with no internet connection required.

At its core, the app uses the on-device EmbeddingGemma 2 model — a lightweight model with 740 million parameters — to capture and process meeting notes across a range of apps, including those used for in-person meetings. Because everything runs locally, users can continue working with their notes while traveling or in areas without reliable connectivity.

The interface will feel familiar to anyone who has used Granola. It features a split-screen layout where users can jot down shorthand notes on one side while AI-generated notes appear on the other. Like Granola, it also supports manual note-taking during live transcription. Users can additionally view a full meeting transcript or interact with a Gemma 4-powered assistant to ask questions and surface key points from the conversation.

One standout feature is the ability to build a personal knowledge base by uploading documents in a variety of formats, including PDFs, Google Docs, Microsoft Office files, plain text, Markdown, and web bookmarks. When a topic from one of those documents comes up during a meeting, the app can draw on that stored knowledge to answer questions in real time — a useful feature for professionals who regularly reference research or internal materials during calls.

Google has confirmed that the app is optimized for Apple Silicon, making the most of the hardware's on-device processing capabilities. The company has not announced whether a broader consumer version is in the works, though it seems plausible that a cloud-powered variant integrated with Gemini could eventually follow, supporting popular video calling platforms.

The meeting note-taker market is increasingly crowded. In recent months, dictation app Wispr and scheduling platform Calendly have both entered the space with their own tools. Meanwhile, productivity startup Superhuman — formerly known as Grammarly — acquired the YC-backed note-taker Fathom last month. Whether Google AI Edge Foresight is a product in its own right or a showcase for the capabilities of its Gemma offline model series remains to be seen, but its arrival signals that Google is watching this space closely.

OpenAI's AI-generated math proofs fall short of expert community standards.

OpenAI's math solutions aren't meeting the field's standards yet

OpenAI released hundreds of solutions to some of the world's hardest math problems, claiming to have consulted elite mathematicians. However, the lab fell short of the standards set by the Advisory Group on Mathematics and Artificial Intelligence (AGMAI), particularly around ensuring human understanding of results. Only 10 of 719 manuscripts included the model's chain of thought, and just 42% of proofs were formally verified.

A new paper from Cambridge and King's College London found discrepancies between OpenAI's natural language proofs and their Lean code translations, raising serious questions about reliability without human review. Prominent mathematician Terence Tao criticized the approach, noting that AI prompters have no deeper engagement with the field after producing a result.

Experts stress that human mathematicians traditionally take responsibility for their discoveries through papers, talks, and community engagement. Harvard professor Melanie Wood warned that when a model outputs a solution, genuine human understanding is still absent — and the real work has only just begun.

🎙️ The Supercharged Podcast Is Growing

Real Conversations with the People Building the AI Future

The Supercharged Podcast is quickly becoming a space for real, unfiltered conversations about AI — beyond the hype, tools, and surface-level takes.

Each episode dives deep with founders, operators, and builders who are actively working with AI — or building AI-first companies — to uncover how it's truly changing the way work gets done.

From strategy and systems to real-world execution, these conversations are practical, honest, and focused on what actually works — not just what sounds good.

⚡ Trends for the Future

China's Manus raises over $500M in first funding round since split with Meta

Chinese AI lab Manus and its parent company, Butterfly Effect, have secured more than $500 million in their first funding round since a proposed $2 billion acquisition by Meta was blocked by Chinese authorities. The round was led by Boyu Capital and IDG Capital, with existing investors Tencent, HSG (formerly Sequoia China), and ZhenFund also participating. The company did not disclose a valuation, though it had previously been reported to be seeking funds at a $4 billion valuation.

Manus first shot to prominence after a viral demo of its AI agent, and the company relocated its staff to Singapore in mid-2025 before announcing the Meta acquisition deal that December. At that point, the startup was reportedly generating over $100 million in annual recurring revenue. However, in April, Chinese authorities ordered the deal to be unwound amid growing concerns in China about AI talent and researchers moving to Western companies. Manus resumed independent operations in August and was required to delete some user data as part of the separation from Meta.

The company operates in a similar space to tools like Cursor, Lovable, and Replit, offering a chatbot and vibe-coding platform that lets users build apps, websites, designs, presentations, and video. Manus recently launched Manus 2.0, which introduces a new architecture and expanded capabilities. It also unveiled a new standalone app called Cue, which gives personal AI agents their own email addresses, phone numbers, digital wallets, and computers, allowing them to handle tasks and make payments within user-defined limits.

With fresh capital in hand, Manus says it plans to continue hiring both in China and internationally. The company is also reportedly exploring a public listing on the Hong Kong stock exchange, signaling ambitions for significant further growth as it rebuilds momentum following the collapse of the Meta deal.

⚡ Let’s Make AI Actually Useful:
What Would Move the Needle in *Your* Industry?

AI has potential — but generic advice rarely helps.

What would be genuinely valuable for AI to do in your industry right now?

• Automate a painful workflow?
• Improve decision-making?
• Replace a manual process that wastes time?
• Help your team upskill faster?

Tell us what you’d want AI to handle — or where you feel stuck.

We’re using these insights to curate **industry-specific trainings, live webinars, and practical guidance** you can actually apply.