😺 Advanced Vision Mode

PLUS: Will AI kill 10% of web traffic in 2025?!

Welcome, humans.

The Neuron will always be free—well, almost free. All we need is one tiny thing from you: share this newsletter with a coworker, friend, or real cool cat you know and get them to sign up. If you do that, we’ll call it even.

You’ll find your unique referral link right below. Just copy, share, and you’re in the running.

We’ll even add a prize: we’re just under 15K subscribers away from hitting half a million before 2025. For the next five days, if you share The Neuron with one friend (no duplicate emails or actual neural nets—we can spot those), you’ll be entered to win a full year of ChatGPT, Claude, or Perplexity ($20 tier... not Pro—what, you think we’re made of money?!).

We’ll pick three winners who help make our holiday dreams come true. Let’s do this!

Share this link for a chance to win

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

You can now share your screen and video camera with ChatGPT.
Meta launched a suite of new tools for video, movement and language.
Google started testing AR glasses with translation features.
Microsoft released Phi-4, a compact math-focused AI model.

ChatGPT’s Advanced Voice Mode now has Video + Screen Sharing (and also Santa?)

ChatGPT giving a live coffee tutorial…

On Day 6 of OpenAI’s 12 days of Shipmas, the company announced it’s rolling out video chat and screen sharing to Advanced Voice Mode, letting you show ChatGPT what you're working on in real-time (FINALLY!!!!)

Picture this: You're making coffee in an unfamiliar kitchen, and need a little help. Instead of typing out “help me make pour-over coffee” and getting a wall of text back, you can just show ChatGPT your setup and get real-time feedback on your technique.

Doesn’t that sound like how AI is supposed to work?

Well, that’s what ChatGPT can do now. Here’s the announcement video to prove it!

Here's how the new features work:

Open ChatGPT on mobile and tap the Advanced Voice Mode icon (bottom right of the composer).
You'll see new buttons for video and screen sharing.
Click video to start a face-to-face chat, or screen share to show ChatGPT what's on your display.

The coolest part? ChatGPT actually pays attention to what it sees. In the demo, it noticed people's holiday accessories (like Santa hats and reindeer antlers) and could recall them later when quizzed.

It also watched and provided real-time feedback on the coffee-making technique, suggesting improvements to “pour speed” and water distribution.

For screen sharing, you can show ChatGPT anything on your device—messages, documents, websites—and get help. Oh, and because it's the holidays, OpenAI added a fun Easter egg: You can now chat with "Santa" through December. Just look for the snowflake icon on the home screen.

Because this is OpenAI (and we can’t trust they’ll deliver on their promises), here are the rollout* details (with a heavy ASTERISK):

Starting today for Teams users and most Plus/Pro subscribers.
Enterprise and edu plans get access in early 2025.
European Plus/Pro users will get access “as soon as possible” (and you know what that means…).
Works on the mobile app, desktop apps, and chat.com.

Our take: This finally feels like a genuine leap forward for AI interaction. Instead of describing what you're looking at or what you need help with, you can just show it. For tasks like troubleshooting tech issues, learning new skills, or getting real-time feedback, this could be a game-changer. Plus, the fact that ChatGPT can maintain context across video, voice, and text in the same conversation is seriously impressive.

Now, how well it works in practice? Remains to be seen. We’re gonna spend the whole weekend testing it out…

FROM OUR PARTNERS

Discover the future of AI with Augmented Worker Intelligence

Imagine if generative AI had access to all the current and historical information from every one of your meetings.

Now, imagine it could respond to your questions with intelligent insights and generate complete and accurate document artifacts in its responses!

That's precisely what Semblian 2.0 delivers.

Generate complete documents using information from your meetings.
Get personalized tactical recommendations after every meeting.
Chat with our AI agent about your entire meeting library.

Be the first to experience the next groundbreaking evolution in AI-driven work automation, enabling turn-key Augmented Worker Intelligence.

Treats To Try.

*Want to create amazing video effects with AI? Learn how to enhance your videos using Dell and NVIDIA’s latest tools + pro tips from industry experts. Register here.
Twelve Labs lets you search your videos by describing what you're looking for ("dogs sneaking into a supermarket," "someone walking on a treadmill") and get the exact moments you need(try the playground here).
Cartesia uses a breakthrough memory-efficient architecture to turn your text into lifelike speech and clone voices in just 90ms - faster than a blink - making it run smoothly even on mobile devices—perfect for gaming, dubbing, and real-time translations—try it here (docs for the devs).
Suno V4, the AI music generator, is now available on Android and iOS. There’s also Antiplaylist, which provides you with randomly curated music to listen if you’re tired of listening to the same old redundant playlists.
TimeMap lets you explore historical events, people, and places through an interactive world map where you can slide through different time periods and zoom into 500k+ detailed street-level maps.
Surf is an assistant, browser, and file manager all in one that makes storing all the cool stuff you find online (like from us!) easier and more intuitive.
Designify creates designs on top of images you share with it, turning your still photos into social posts for free (paid version w/ batch features as well).

See our top 51 AI Tools for Business here!

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

Around the Horn.

Grok’s new image generator in action.

Meta shared a ton of new tools at Meta FAIR:
- Video Seal for watermarking genAI video.
- Motivo for realistic digital agent movement in the Metaverse (try it here).
- The Large Concept Model for improved language processing (code),
- Plus advances in memory scaling, vision-language encoding, and social intelligence (lots to dive into here for Sunday!).
Google announced limited testing with select users for its Project Astra AR glasses—they’ll run on the new Android XR operating system, and feature real-time multimodal capabilities for translation, text reading, and visual processing.
Microsoft launched Phi-4, a smaller and faster AI model that excels at math problems while using high-quality synthetic datasets to improve performance
Check out Anybotics, who just raised $60M and their inspection bot Anymal—it’s like a wonky alt version of the Boston Dynamics bot (read more here).

FROM OUR PARTNERS

The Must-Have Software for Every YouTube Channel

Spotter Studio helps you generate better-performing video ideas, titles, and thumbnails for

YouTube. It’s like a creative partner personalized to your YT channel.

Creators are already celebrating their success after using Spotter Studio. If you're a content creator, entrepreneur, or professional YouTuber, don’t miss out.

Get a 30-day free trial of Spotter Studio

Intelligent Insights

Check out this deep analysis of the AI scaling laws that flips the “scaling is dead” narrative on its head—it turns out companies are discovering entirely new dimensions of scaling through synthetic data, post-training, and compute-heavy inference.
Here’s a practical guide to the implementation challenges you'll face when building open-source LLM applications, complete with battle-tested solutions for everything from handling wonky outputs to managing context windows.
Fascinating profile on Fei-Fei Li of World Labs covering her critical role in the AI boom and where World Labs goes next.
Here’s a list of what software engineers predict will happen in 2025 (take these nerds with a grain of salt!).
1. An AI personality will become "household famous" (though there was debate about whether Alexa already qualifies)
2. Web traffic will drop "at least 10% across the board" due to AI search results, LLMs, and improved phone assistants
3. AI investment will bottom out, with small 512 MB phone models being "enough for most people.”
4. Gaming's use of AI (like NVIDIA's DLSS) will become so popular and effective that it'll cause another shortage of high-end graphics cards and drive prices through the roof.