😺 NVIDIA's Apple moment?!

Welcome, humans.

We have exciting news for any of you wanting to get your AI product or service (or AI-adjacent product / service) in front of 500K+ daily Neuron readers: We have sponsorship slots available in Q1!

The world’s leading AI companies advertise with The Neuron because of the trust we’ve built with you, our audience.

And trust us, we got some of the NICEST comments on our first email back of the new year… can’t wait to share them all in A Cat’s Commentary. We have THE BEST readers!

Anyway, click the button below, fill in a few deets, and we’ll reach out to follow up!

Advertise in The Neuron here.

Here’s what you need to know about AI today:

• NVIDIA released a $3K AI supercomputer and new robot training model.
• OpenAI lost money on ChatGPT Pro despite $200 price tier.
• Samsung, Qualcomm, Google, and AMD all launched new AI products at CES.
• Apple secretly opted us all into AI photo analysis back in 2023.

NVIDIA wants to put an AI supercomputer on your desk (and it only costs $3,000).

Late last year, NVIDIA announced a tiny, $250 AI supercomputer for pet AI projects—and it released a pretty hilarious launch video featuring CEO Jensen Huang’s actual pets. The idea of this little chip was to make AI more accessible to everyone, not just big tech.

Remember when putting a personal computer on every desk and in every home changed the world? NVIDIA's more or less trying to do the same thing with AI supercomputers.

And last night at CES 2025, Jensen Huang announced phase two of this plan: Project DIGITS, a $3K personal AI supercomputer that runs 200B parameter models from your desk. Guess we now know why Apple recently developed an NVIDIA allergy…

DIGITS is powered by NVIDIA's new GB10 Grace Blackwell Superchip, delivering 1 petaflop of AI performance—fancy talk for “you can run ChatGPT-sized models on your desk using a regular power outlet.”

But NVIDIA doesn’t just want its “Apple PC moment”… it also wants its OpenAI moment. NVIDIA also announced Cosmos, a platform for building physical AI (think: robots and self-driving cars)—which Jensen Huang calls “the ChatGPT moment for robotics.”

It generates synthetic training data for robots and autonomous vehicles, processes 20 million hours of video in just two weeks (versus 3.4 years using CPUs), and is already being used by companies like Uber, Waabi, and Figure AI.

The big news for us regular folks here is clearly the $3K AI PC, though. Specifically, how it takes what we like about Macbook Pros and Mac Minis, and hyper-optimizes for AI-specific workflows to run bigger local models.

Speaking of big models, DeepSeek V3 dropped while we were on holiday, and it's currently the strongest open-source AI model available, outperforming Meta's Llama and even matching some of OpenAI and Anthropic's closed models on certain tasks.

It’s basically ChatGPT-4 quality but open source—it's so similar that it sometimes even tells the same jokes and forgets it's not actually ChatGPT. We're not kidding, it literally thinks it's ChatGPT sometimes.

But here's the catch: it's massive. We're talking 671B parameters-massive.

To run it optimally, you'd need either 18 A100 GPUs ($540K-$720K), 9 H100 GPUs ($270K-$360K), or 4 AMD MI300X GPUs (~$60K).

Why this matters: Until now, training physical AI systems has been incredibly expensive and time-consuming. While NVIDIA's $3K supercomputer is a huge step toward democratizing AI, we're still a ways off from running the most advanced models on our desks.

FROM OUR PARTNERS

💼 Want to become an AI Consultant?

This stat is WILD: The AI consulting market is projected to explode from $6.9 billion to $54.7 billion by 2032. Want to capture this opportunity?

So how do you, an AI enthusiast, become a handsomely compensated AI consultant?

Well, The AI Consultancy Project from Innovating with AI has already helped 300+ students build successful consulting businesses.

Current students love:

• Tools and frameworks to find and serve clients.
• 6-month roadmap to building a 6-figure business.
• Methods to land your first client in as little as 3 days.

As a Neuron reader, you get early access to the next enrollment cycle!

Take advantage of the AI boom—request early access to The AI Consultancy Project ASAP!

Level up your AI for 2025, part one: NotebookLM

Learn 80% of NotebookLM in Under 13 Minutes!

This 12-min NotebookLM tutorial might be the best AI explainer we've seen this year. In it, creator Jeff Su breaks down exactly when (and when not) to use Google’s powerful NotebookLM research tool.

The TL;DW? NotebookLM shines when you need to analyze scattered info across different formats with near-zero hallucination tolerance.

It’s not going to replace your ChatGPT, Claude, or Gemini usage, but it’s a great research assistant to collect your findings + feed them into your other AI tools.

Treats To Try.

1. *Attention is our (Noah & Grant) secret weapon for transforming weekly syncs into perfect action items and follow ups—click here to book a demo ASAP.
2. Trickle helps you build AI chatbots, 3D games, and database-powered web apps just by typing what you want—get 5 messages a day for free, but be warned: no custom domain support yet.
3. Gensmo lets you snap a photo or type what you need to get instant visual collages of matching gifts, outfits, and decor (not better than Pinterest…yet).
4. Cobalt downloads videos, audio files, and GIFs from social media websites by simply pasting a URL.
5. Peony creates professional file-sharing pages where you can track who views your documents and protect them with passwords and watermarks—here’s a demo we made based on yesterday’s email.
6. PicturePicker is a Chrome extension that saves and organizes images from any webpage with one click, automatically categorizing them so you can instantly find what you need (free to try, but there’s a paid version too).
7. QuickMock is a Chrome extension that turns any LinkedIn job posting into an instant mock interview, giving you personalized feedback and practice with job-specific questions.

See our top 51 AI Tools for Business here!

*This is sponsored content. Advertise in The Neuron here.

Around the Horn.

• OpenAI Round-up
  • In a new blog post, Sam Altman claimed OpenAI “knows how to build AGI” and will deploy OpenAI’s AI agents into the workplace this year, while confirming ChatGPT has grown to 300M+ weekly users.
  • OpenAI is losing money on its $200/month ChatGPT Pro subscriptions due to heavy usage, with ChatGPT costing an estimated $700K daily to run, yet still projects hitting $100B in revenue by 2029.
• CES 2025 Round-up
  • Samsung announced Vision AI, bringing AI features like live translation and personalized wallpapers to its entire TV lineup, including a partnership w/ Microsoft to integrate Copilot.
  • AMD and Qualcomm launched competing AI laptop processors, with major manufacturers set to release new devices using both chips in Q1.
  • Google announced an AI-powered TV feature for Google TV that uses Gemini to create "News Briefs," summarizing news from across the web and YouTube, set to launch in late 2025.

FROM OUR PARTNERS

Ready to accelerate your video creation?

Join Dell experts on January 9th to see how Precision workstations and NVIDIA RTX™ transform pre and post-production workflows. Get practical tips from Foundry experts. Save your spot!

Tuesday Ticker

Here are the results from last year’s poll:

Here’s what you said gave it away as being AI:

• S.E: “The nose on the shoulder of the pineapple (?) when the AI mistakes the shoulder armor for a head. Also, their faces taking on the texture of the fruit before the helmet comes down...that makes no sense.”
• J.L: “The fruit and how it doesn't spray or reach to the physics of what it's going through.”
• G.S: “The fact that the guy has a pineapple for a face is a dead giveaway.”

A Cat's Commentary.