🤓 This Week in AI
PLUS: AI Search Wars, ChatGPT’s Must-Know Updates, and Google’s Stunning Video Tech
Happy Friday—and Happy Holidays!🎄
Even though it’s the last working week of the year, AI companies clearly didn’t get the memo. Between launches and big announcements, they’ve got your girl on her toes—but I wouldn’t have it any other way.
As wild as it is, it’s truly an exciting—and yes, a little scary—time to be alive.
Quick heads up—this is my last edition of 2024, but I’ll be back in your inbox on Friday, January 10th.
That said, AI moves fast, and important updates are bound to happen before then. I’ll be sharing the latest on LinkedIn, Threads and Bluesky, so make sure you’re following me there to stay in the loop.
If you’ve missed some editions over the past few months, the holidays are a great time to catch up. Pour a glass of wine, grab a cozy blanket, and dive in.
If you're short on time, focus on the last 4-6 weeks—they're packed with context and insights that will help you stay informed and feel empowered without the overwhelm (or maybe a little overwhelm if you read them all in a row). 😁
On a more personal note, pretty sure I’ve found my dream job 👇. I’d absolutely kill it—two years deep in generative AI have prepped me for exactly this.
Thank you so much for reading, sharing, and supporting this newsletter—especially my paid subscribers. Your support means everything and keeps me going.
May your holidays be restful, fun, and as peaceful as the moment after everyone finally agrees on which movie to watch.
Ok, let’s dive in.
Here’s what you need to know about AI this week (clickable links appear in orange in emails and underlined in the Substack app):
Week two of OpenAI’s 12-day announcement spree didn’t disappoint. Here are the highlights:
🎁 ChatGPT Search Gets Updates and Is Now Free to All Users
This move puts ChatGPT directly in competition with Google.
All you need is a ChatGPT account. Here’s what’s new:
Mobile Search That Feels Familiar: Search for local businesses, and you’ll get results with images, ratings, hours, and even maps with directions—just like a traditional search engine.
Faster Results with Direct Links: · Searching for specific sites? ChatGPT now surfaces clickable links first, before generating longer responses.
Search with Your Voice: This is a big one—instead of typing your questions or requests, you can now talk to ChatGPT hands-free to get real-time answers. Imagine how this will change our expectations and behavior.
I HIGHLY recommend watching this live demo to see how it all works.
AI search engines are the new gatekeepers. They are quickly becoming the primary interface for how we find and interact with information, content, products, and services.
If you’re in marketing or publicity and haven’t been closely following these shifts, it’s time to get moving.
AI search is reshaping online discovery and visibility—and your strategy needs to adapt. Fast.
Here’s a real-world example of what this means: Some major outlets, like The New York Times, have opted out of AI crawlers to protect their IP. Others, like The Wall Street Journal, have partnered with OpenAI. This means WSJ content is crawled, can surface in ChatGPT’s search results, and—importantly—can be cited and linked to.
What does this look like in practice? I recently asked ChatGPT about the new AI production company Promise. It pulled multiple sources, but guess what link showed up first? The Wall Street Journal.
Yes, this is partly because they did a major piece on the startup’s launch, but it’s also because their partnership with OpenAI ensures their content is prioritized, discoverable, and cited within the search results.
Now imagine if that same story had been published in The New York Times. It wouldn’t appear at all.
This doesn’t mean that there aren’t strategic reasons to prioritize The New York Times. The point is to make these decisions with all the factors in mind.
These shifts are creating big questions—and even bigger opportunities.
But it’s a lot. New updates are rolling out weekly, and sorting through what actually matters feels like a full-time job.
Trust me, even I have to work to keep up, and I live and breathe this stuff.
The important thing is to start. If you need a hand, just reach out—I’ve got you.
—
🎁🎁 OpenAI Introduces “Projects” – Smart Folders for ChatGPT
Projects lets you organize related chats, files, and custom instructions (or master prompts) in one place for easier access.
If you’re using ChatGPT for work (and if you’re not 👀), this is going to make life so much easier.
Let’s say you’re working on a new Instagram campaign for your business. Here’s how you could use Projects:
Step 1: Set the Foundation. Create a project for your campaign and upload your brand strategy and guidelines—everything from brand voice to audience insights and examples of your top-performing posts.
Step 2: Add Custom Instructions. Write a detailed master prompt with additional context to guide ChatGPT on how to craft posts that match your tone and objectives.
Step 3: Work Smarter, Not Harder. Now, each time you need a new post or a series of posts, simply start a chat within your project, share specifics for the post, and let ChatGPT reference your knowledge base and follow instructions from your master prompt automatically.
If this sounds familiar, you’re not wrong—Claude has had a similar feature for a few months now, and I’ve been using it regularly. It’s been a great way to keep things organized, but here’s the catch: I hit the usage caps on Claude’s paid plan a lot sooner. So, if you’re working on larger projects and have high-volume needs, ChatGPT’s Projects might offer a little more breathing room.
You can find detailed instructions on how to use Projects here, and watch this demo to see how it works.
Projects is available now to paid users and coming “soon” to free users.
—
🎁🎁🎁 You can now call ChatGPT to talk to it—even on a rotary phone.
This is going to be especially clutch when Wi-Fi isn’t so good.
It’s free, no account needed, but only available in the U.S.
It’s also genius.
1-800-CHATGPT ➡️ 1-800-242-8478
OpenAI also rolled out big updates and price reductions for developers—and teased a major announcement coming later today, so stay tuned…
YouTube is now letting creators decide whether third-party companies—like OpenAI, Anthropic, Perplexity, and Microsoft—can train their AI models on their videos, but only with explicit permission.
Google is also quick to point out that “unauthorized scraping remains prohibited.”
This move comes after reports that some of these same companies (like OpenAI and Anthropic) have already trained their AI models on content scraped from YouTube.
But what creator is going to opt in by choice? Ummmm, no one. 🙃
So, what is this really about? It’s Google’s way of making unauthorized scraping less appealing and harder to ignore or justify.
Also, as my friend Morgan pointed out during our discussion this week, this makes scraping feel “more personal” for creators—who, as we know, are a loud bunch. Maybe this new approach will actually work.
And yet, creators still can’t stop Google from training its AI on their videos.
Google has introduced Veo 2, an AI video generator it claims outperforms OpenAI’s Sora in audience tests. But, in classic Google fashion, it’s locked behind a waitlist on Google Labs’ VideoFX platform—so most of us can’t access it yet.
From what we've seen in early access previews, the quality is genuinely impressive. Its standout feature is more realistic physics. That means objects fall, bounce, and reflect light, the way they would in the real world. For example, a ball rolling off a table doesn’t float in the air—it drops naturally and rolls as expected.
AI video models often struggle with this level of realism, producing movements or interactions that feel “off.” While Veo 2 isn’t flawless, it’s a clear step up from many of its competitors, including Sora.
Combine this with the ability to create videos up to 2 minutes long at 4K resolution, and it’s clear Veo 2 is raising the bar.
One of my favorite clips so far is this one, created with the prompt “a sitcom TV show about potatoes”.
Google also launched its advanced image generation model, Imagen 3, and it’s good. Real good. It follows prompts exceptionally well, creating realistic, high-quality and detailed images that rival Midjourney and Flux. I’ve been testing it, and I’m blown away. You can access it for free via Google’s ImageFX tool, but you’ll need a Google account to log in
Speaking of new video generation models….
Pika labs just dropped an updated video generation model that lets you add images of people or characters (including yourself), objects, clothing, and environments, and give AI instructions to create a fully customized scene.
So, if you upload a photo of someone and a picture of a cat, then prompt it with something like "a person petting a cat," it will generate a video of that person petting the cat. Here’s an example of AI filmmaker Dave Clark inserting himself into a few scenes.
There are endless valuable use cases here, especially for creators and brands looking to create customizable, shareable social media videos and ads—with more creative control.
But here’s the thing: it won't matter who has the “best” image or video generation model this year. The tech is advancing fast. Soon, most tools will be able to generate high-quality, hyper-realistic videos with plenty of controls—and at a fraction of the cost.
When that happens and the tech is a commodity, which companies will be defensible? Well, that’s the trillion-dollar question I’ve been thinking about, and I have some initial thoughts, but that’s a whole other newsletter. 😉
Wired created this 👇 visualization of every AI copyright lawsuit in the U.S.
OpenAI CEO Sam Altman, Jeff Bezos, and Mark Zuckerberg each contributed $1 million to Trump’s inauguration, signaling tech’s willingness to kiss the ring for potential influence and lighter regulations.
Runway just launched Talent Network, a media platform to connect brands, studios, and agencies with the next wave of talented storytellers.
This is both smart business and awesome.
The future of love is getting an AI upgrade. 💕 Dating apps like Hinge and Bumble are rolling out smarter matching tools and AI dating coaches to help you pick better profile photos, craft flirty replies, suggest icebreakers, and even plan dates.
Honestly, I’m down for anything that will improve my Bumble matches 😁.
Meta’s Ray-Ban smart glasses just got smarter. Here are the upgrades:
Live Translation: The glasses now translate conversations in real-time between English, Spanish, French, and Italian. You’ll hear the translation through the speakers or see it as text on your phone.
AI Visual Assistance: With a built-in camera, the glasses can “see” your surroundings and answer questions about what you’re looking at.
The v11 update is available for Early Access users starting this week.
HarperCollins CEO says AI is shaking up publishing. It’s helping make audiobooks for smaller markets and translating books into more languages—both of which could open up new audiences.
Though he also acknowledges that gen AI will likely flood the market with low-quality content, leaving publishers to fight for attention and forcing them to double down on quality.
I’ve been banging on about this for the past year: comprehensive AI training is key to surviving the AI shift, and Mastercard is finally showing how it’s done.
The company offers its workforce three levels training: broad foundational courses, job-specific upskilling, and a reskilling program to prepare those whose roles may become obsolete.
Their bet? Employees who understand and know how to use AI will outperform those who don’t.
YouTube is partnering with CAA to help celebrities identify AI-generated content using their likenesses, and submit removal requests.
A first-of-its-kind experiment in South West England will use AI-powered cameras to detect drivers who could be under the influence of alcohol or drugs.
In case you missed last week’s edition, you can find it 👇:
That's all for this week.
I’ll see you next Friday. Thoughts, feedback and questions are always welcome and much appreciated. Shoot me a note at avi@joinsavvyavi.com.
Stay curious,
Avi
💙💙💙 P.S. A huge thank you to my paid subscribers and those of you who share this newsletter with curious friends and coworkers. It takes me about 12+ hours each week to research, curate, simplify the complex, and write this newsletter. So, your support means the world to me, as it helps me make this process sustainable.