
Welcome back to the Neural Net! Friday’s here, and so is a long weekend. We’re off for Labor Day too, but we’ll be back in your inbox next Friday.
In today’s edition: Nvidia’s Q2 earnings breakdown + why it matters, Claude used for massive cyber attacks, Microsoft releases home-grown models, a new AI wearable, and more.
▼
The Street

note: stock data as of last market close
▼
👑 Nvidia’s Earnings: Heavy Is the Crown That Powers AI

Nvidia just delivered another record quarter. The company now accounts for 7.5% of the S&P 500 and is the only publicly traded company valued over $4 trillion.
$46.7B in total revenue vs $46.2B expected
$30.2B in profit vs $29.4B expected
Impressive 72.7% gross margin
A forecast of $54B for the next quarter, and that’s without counting a single chip sold to China
But instead of rallying, the stock fell about 3% after hours.
📉 Why Wall Street Was Disappointed
Nvidia’s revenue jump of 56% year over year is a massive gain by most standards, but it’s actually the slowest growth rate seen in over two years.
A large part of Nvidia’s recent performance was shaped by policy decisions, not market forces:
The company was unable to ship its H20 AI chips to China due to export restrictions.
While the U.S. reversed the ban in July (in exchange for a 15% revenue share), Nvidia says it hasn’t shipped any of those chips yet and may not this quarter.
Chinese regulators are pressuring local companies to use domestic chips, like those from Huawei, which would further reduce reliance on the American AI stack.
⚖️ Is the Reaction Fair?
Behind the headlines, there is plenty of reason for optimism. Nvidia’s Q2 story was all about data centers, and that isn’t expected to slow down:
Big Tech is set to pour $320B into AI infrastructure this year
About 50% is expected to go straight to Nvidia
And while fears of an AI bubble are rising, the stock’s earnings multiple looks relatively reasonable at 51—especially next to Palantir, which trades at a sky-high 527.
🔮 Reading the Forecast in Nvidia’s Chips
There’s a potential shift underway: a transition from hypergrowth to something closer to normalization. But what’s most important isn’t just Nvidia’s numbers, it’s what those numbers represent.
▼
💡 How To AI: The Gadget That Never Forgets
AI wearables are having a moment. Maybe you skipped Meta’s smart glasses, or watched the Humane pin flame out, but there’s a new contender worth your attention. And if the trend is any indication, this kind of tech could soon become the norm in your office.
The World’s Most Wearable AI
Limitless is your new superpower - an AI-powered pendant that captures and remembers every conversation, insight, and idea you encounter throughout your day.
Built for tech leaders who need clarity without the clutter, Limitless automatically transcribes and summarizes meetings, identifies speakers, and delivers actionable notes right to your fingertips. It’s securely encrypted, incredibly intuitive, and endlessly efficient.
Order now and reclaim your mental bandwidth today.
▼
Heard in the Server Room
Anthropic says its chatbot Claude has been moonlighting for hackers and even North Korean operatives. It’s been caught drafting ransom notes, plotting attacks, and helping scammers fake résumés to land U.S. tech jobs. The company shut it down and is beefing up defenses, but what a long way we’ve come from the days of Nigerian princes asking for wire transfers.
MIT says 95% of GenAI pilots are belly-flopping. Maisa AI thinks the fix isn’t flashier chatbots but dependable “digital workers” you can trust. With a fresh $25M seed, its Maisa Studio lets you describe a task in plain English, and the AI builds a step-by-step workflow to execute it. With built-in guardrails to cut down on hallucinations and errors, it aims to be a smarter, modern upgrade to old-school process automation. Maisa’s CEO sums up the company’s philosophy bluntly: “The future belongs to transparent, maintainable AI systems. Everything else is just expensive technical debt with good marketing.”
Microsoft just unveiled its first in-house AI models: MAI-Voice-1 and MAI-1-preview. Unlike Copilot, which mostly runs on OpenAI’s GPT models, these models were fully built by Microsoft’s own AI team. MAI-Voice-1 can spit out a minute of audio in under a second on a single GPU, while MAI-1-preview, trained on 15,000 Nvidia H100s, tackles text queries with a consumer-first focus. The launch gives Microsoft more control over costs and product direction, while also hinting that its partnership with OpenAI may shift from reliance to balance. Hopefully the model is more interesting than its name, which sounds like somebody fell asleep on the keyboard.
▼
In Partnership With ClickUp
Meet ClickUp, the all-in-one AI-powered productivity platform that brings your tasks, docs, and teams together.
Skip the AI Learning Curve. ClickUp Brain Already Knows.
Most AI tools start from scratch every time. ClickUp Brain already knows the answers.
It has full context of all your work—docs, tasks, chats, files, and more. No uploading. No explaining. No repetitive prompting.
It's not just another AI tool. It's the first AI that actually understands your workflow because it lives where your work happens.
Join 150,000+ teams and save 1 day per week.
▼
👷 AI Cuts Jobs But Fuels Growth

AI’s reshaping the U.S. job market, and not everyone’s winning. A Stanford study found entry-level workers in AI-exposed fields like customer service, accounting, and software development have seen employment drop 13% since 2022. Young workers (ages 22–25) are hit hardest, while older employees, and those in less AI-exposed sectors like healthcare, have held steady or grown.
Why young workers? AI is quick to replace “book learning” skills fresh grads rely on, but struggles to mimic experience-based judgment.
Not all bad news: In jobs where AI boosts efficiency instead of replacing workers, employment is strong.
Big picture: This helps explain why overall job growth looks fine, even as young workers struggle to land roles.
On the flip side, AI is juicing the economy in less obvious ways.
“It’s not because of how companies are using the technology, at least not yet. Rather, the sheer amount of investment — in data centers, semiconductor factories and power supply — needed to build the computing power that A.I. demands is creating enough business activity to brighten readings on the entire domestic economy.”
In fact, this year marks the first time spending on data centers will surpass office buildings.
👉The long-term story is less about loss and more about transformation—disruption in some areas, but growth for the economy and society as a whole.
▼
That’s it for today! Have a fantastic weekend, and we’ll catch you next Friday with more neural nuggets.




