🤓 This Week in AI

PLUS: How AI Can Help You Read Room Before You Present ANYTHING

Happy Friday!

Whether you’re pitching a new idea, delivering a project update, or speaking at a conference, getting a read on your audience can make all the difference. This week, see how AI can help you anticipate their reactions, sharpen your message, and make sure it sticks.

But first, here’s what you need to know about AI this week (clickable links appear in orange in emails and underlined in the Substack app):

---

It’s only week one of OpenAI’s 12-day announcement spree, and I already feel like I need a second brain to process it all.

🫠 Spoiler: I don’t have one, but I do have this newsletter to write.

So instead, I’ve resorted to sheer determination and a questionable amount of caffeine (during the day) and wine (at night).

If you’re wondering, “Is this too much information for one week?” the answer is yes. Yes, it is.

But since most of these updates are important (and frankly, kind of wild), I’ve sorted through them to share only what truly matters.

So, take a deep breath. Here we go.

🎁 OpenAI finally launches Sora, its highly anticipated video generation model.

It can create complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background—all from text, image, or video prompts. But Sora isn’t just a video generator. It’s also a full editing video platform, which includes the below features:

• Remix: Replace, remove, or re-imagine elements in your video.

• Recut: Find and select the best frames, extending them in either direction to complete a scene

• Storyboard: Arrange and edit multiple video clips chronologically on a timeline to create a longer sequence (similar to traditional editing platforms like Adobe)

• Loop: Create a seamless loop based on any part of a video

• Blend: Combine two video clips into one seamless sequence

• Style Presets: Customize and share visual styles with co-creators.

Here’s what the interface looks like:

Sora interface

This test of the remix feature is pretty awesome.

Access is exclusive to paid subscribers:

💰 Plus plan ($20/month): 50 video generations (5 seconds each)

💎 Pro plan ($200/month): Unlimited generations (up to 20 seconds) with no watermarks

What else you need to know:

• The model is available in the U.S. and many other countries, but stricter AI regulations in Europe and the UK mean it may 'be a while' before it launches there.

• The generated videos will have visible watermarks and metadata to indicate they’re made with AI.

• OpenAI acknowledges Sora’s current limitations—it often generates unrealistic physics and struggles with complex actions over long durations, but stresses that this is just the starting point and will improve over time.

• Access is not available to users under 18.

You can access Sora by going to Sora.com (not ChatGPT). You’ll need a ChatGPT account to log in.

If you don’t have a ChatGPT subscription, you can still browse the feed of AI-generated videos created by other people using Sora.

Ok, I’ve been waiting for this next one for MONTHS because it offers new and more intuitive ways for us to share information and communicate with our AI assistants.

🎁🎁 OpenAI adds live video and screen sharing to Advanced Voice Mode

This means that you can show ChatGPT your surroundings and share real-time visual context to make your conversations more dynamic and helpful. Instead of typing to explain a situation, simply point your camera and let ChatGPT "see" what your phone sees as you talk through it.

Here are a few ways you can use it:

• Fix It Faster: Point your camera at your appliances, plumbing leak, or tangled cables, and ChatGPT can guide you step-by-step to troubleshoot and solve the issue.

• Work Smarter: Share your screen during a presentation review to refine slides or analyze spreadsheets with AI feedback. Or show ChatGPT physical whiteboard notes, or even a product prototype to get real-time suggestions for improvements.

• Creative Projects: Share your sketches for logo design, storyboards, or outfit ideas, and get actionable feedback or suggested improvements.

• DIY and Home Projects: Scan your garden or living room, and ChatGPT can recommend plant layouts, garden arrangements, or home decor ideas that match your preferences.

• Quick Shopping Help: Scan a product label or compare items to get detailed insights, reviews, or alternative recommendations.

I highly recommend checking out this live demo to see how it works.

The rollout will happen in phases with most Plus and Pro subscribers getting access this week, except for those in the EU.

🎁🎁 🎁 OpenAI rolls out ChatGPT’s Canvas feature to all users.

Canvas converts ChatGPT’s traditional interface—with a conversation at the top and text entry box at the bottom—into a side-by-side layout:

• Left Sidebar: Where you input your prompts and instructions and receive responses from ChatGPT

• Right Workspace: Displays ChatGPT’s output in real time including initial or revised drafts, which are directly editable by you in the workspace.

It lets you adjust your text for length, change the reading level, and add a final polish with a couple of clicks.

You can also highlight specific sections of the output and ask ChatGPT to improve them or add sections to expand on certain points. You’ll now be able to ask ChatGPT for in-line comments, too.

This is what it looks like:

ChatGPT canvas interface

Canvas makes ChatGPT’s interface more intuitive and collaborative so you can edit and revise outputs more efficiently for many use cases.

🎁🎁🎁🎁 ChatGPT x Apple Intelligence: Siri can now call ChatGPT for help, use it to write, and lend your phone’s camera to it. This integration brings ChatGPT directly to iPhones, iPads, and Macs, offering one-click access to advanced AI assistance. Watch this short announcement video to see how it all works.

This partnership gives ChatGPT broad distribution across millions of Apple devices while helping Apple (which doesn’t currently have its own foundational AI model) make Siri smarter and better suited to meet users’ growing expectations for AI on its devices.

🎅🎅🎅🎅🎅 “Santa Mode” brings holiday fun to ChatGPT

Users aged 13 and up can now talk to a simulated version of Santa Claus through the app's voice mode. I found this version of Santa to be kinda lame, but your kids might like it. Then again, I don’t have kids, so what do I know about entertaining tiny humans?

To chat with Santa, just tap the snowflake icon in Voice Mode and start your conversation. You can also check out the last few minutes of this live demo to see how it works.

---

Tired of Googling “Reddit” to find what you need? Reddit Answers is here to change that.

It’s an AI tool that summarizes posts and conversations directly on the platform so you can bypass Google for Reddit-specific searches.

This move marks a strategic shift for Reddit. Beyond signing deals with Google and OpenAI to monetize its data, Reddit is building its own tools to keep users on its platform, reduce reliance on third party search engines, and position itself as a leader in AI search.

---

Google launched Gemini 2.0, a higher performing model which the company says is the start of its 'agent era' where AI is capable of performing tasks without human input. It is multimodal (can generate text, images and audio), better at “reasoning”, and initially only available to developers or as an experimental model for Gemini Advanced subscribers.

Nothing to do on this one for now, aside from knowing about it...

---

In response to rising youth suicide rates, schools are increasingly turning to AI tools to monitor students’ devices for signs of self-harm, so they can intervene in critical moments. While these tools have reportedly saved lives, they also raise serious privacy concerns, often triggering false alarms that lead to unwarranted police involvement and distress among students.

---

YouTube is turning its creators into global brands.

Its AI-powered auto-dubbing feature is now available to hundreds of thousands of channels–mainly informational content creators–with plans to broaden the tool's availability soon.

It translates videos into multiple languages—English, French, Spanish, and more—breaking language barriers and making global access easier than ever. Now, a French chef can teach recipes in English, and a Japanese scientist can reach audiences worldwide.

The voices are a bit robotic for now, but YouTube promises they’ll get better at capturing the original “tone, emotion, and even the ambiance of the surroundings” with later updates, especially since similar tech is already available on other platforms

---

AI is taking over another job—OnlyFans chatters.

These gig workers were hired to pretend they’re the creators in DMs, building personal relationships that fueled the platform’s success. Now, AI tools are generating messages faster, cheaper, and often unnoticed by fans.

---

Manchester City, a leading Premier League club, is teaming up with Puma to let fans design its 2026-2027 third uniform—or "kit"—using AI. The winning jersey, chosen by fans and experts, will be worn by players and sold worldwide.

---

Yelp's sophisticated new AI reads between the lines of reviews to surface insights about businesses including food quality, service, wait time and ambiance—even when not explicitly mentioned. Using facial icons to display sentiment, "Review Insights" helps users make faster decisions. Next up: AI-powered personalized feeds and trending local insights.

---

Donald Trump names former Paypal exec and venture capitalist David Sacks as White House AI and crypto czar.

---

No lawyer, no problem: A 23-year old Reddit user used ChatGPT to sue his landlord and win - all without proper legal representation.

---

This last one may not seem AI related, but it is.

Mr. Beast’s latest sports themed video featuring Cristiano Ronaldo and Tom Brady got 48M views in 24 hours—more than the most-watched NFL match in the U.S. With 334M YouTube subscribers, he’s a magnet for Gen Z and sports leagues eager to connect with younger audiences. Advertisers, too, are paying premium rates to tap into his viral power.

Generative AI tools are helping creators streamline production workflows and grow their businesses. Platforms like YouTube amplify this shift, serving as powerful distributors that continue to reshape how talent connects with global audiences.

As AI lowers production barriers, many more creators will emerge as media moguls, directly competing with traditional media companies.

USEFUL PROMPTS

How AI Can Help You Read the Room Before You Present ANYTHING 🎯

Presentations—whether at work, during team meetings, or at a conference—are more than just sharing information. To truly engage and persuade your audience, you need to anticipate how they might respond.

Knowing their questions or concerns ahead of time allows you to fine-tune your message, address objections before they arise, and stay in control of the flow.

Here’s where AI comes in.

I often use AI to analyze my presentation and predict audience reactions. It helps me step into my audience’s shoes so I can anticipate their questions, concerns or objections and refine my message and delivery before I even walk into the room.

The result? A presentation that feels more personalized, clear, and persuasive.

Below is the exact prompt I use, which you can adapt to your situation and audience. It can be used in a variety of use-cases. Here are just a few examples:

🏢 Work Presentations:

• Presenting a new strategy or proposal to colleagues or senior management, where you anticipate tough questions and need to persuade key decision-maker

• Pitching a new project or idea to team members while addressing concerns about resources or feasibility or alignment on goals

👥 Client Presentations:

• Presenting a proposal to potential clients and needing to be ready for questions about capabilities, pricing, timelines and deliverables

• Handling objections from clients about a product or service and refining your responses to win them over

🎤 Public Speaking and Events:

• Giving a talk at a conference where audience members may have a diverse range of questions based on their background or agendas

• Preparing for a panel discussion or Q&A session where the audience might challenge your ideas or ask for deeper insights

---

⚡Prompt⚡

# Role

You are an expert communication strategist and audience behavior analyst with extensive experience and a deep understanding of how to craft persuasive presentations tailored to various audience dynamics. You excel at helping presenters connect with their audience on a psychological and emotional level, ensuring that every part of the message is clear, engaging, and compelling. You are also skilled at anticipating audience reactions, identifying potential objections or concerns, and refining presentations to overcome these barriers. Your goal is to ensure that the presentation achieves its intended outcome by addressing the audience's needs, interests, and motivations.

# Context

I am preparing to [INSERT PRESENTATION TYPE, E.G., PITCH MY BUSINESS, GIVE A WORK UPDATE, OR DELIVER A KEYNOTE] and I want to ensure that I can effectively communicate and engage my audience. The success of my presentation depends on being able to anticipate their reactions and refine my message accordingly. Here is the context:

• My goal: [INSERT MORE SPECIFICS ABOUT YOUR GOAL]. This should answer the question ‘What does success look like?’

• Target Audience: [INSERT AS MUCH CONTEXT & DETAILS AS POSSIBLE INCLUDING VARIOUS AUDIENCE SEGMENTS WHO MIGHT HAVE DIFFERENT NEEDS, PREFERENCES OR AGENDAS—INDUSTRY, JOB TITLES, LEVEL OF EXPERTISE, DEMOGRAPHICS, PSYCHOGRAPHICS, ETC.]

• Additional background: [INSERT SITUATIONAL OR CIRCUMSTANTIAL DETAILS THAT MAY AFFECT THE PRESENTATION, E.G., INTERNAL POLITICS, AUDIENCE POWER DYNAMICS, OR RECENT EVENTS]

• Presentation Format [SPECIFY IF IN-PERSON, VIRTUAL, OR HYBRID/INCLUDE THE SETTING]

# Instructions

Please follow the step-by-step process below to help me fine-tune my presentation for maximum impact, persuasion and engagement:

Step 1: Audience Analysis - Help me analyze the audience’s needs and mindset. If there are different segments of the audience, provide insights for each. Identify the following:

• Identify key motivations, desires, needs, pain points and preferences

• Analyze unspoken biases, preconceptions or potential cultural factors that can influence reception

• Determine personal stakes (career, reputation, business needs)

Step 2: Anticipate Questions, Objections, and Concerns - Examine the content of the presentation and identify potential questions, objections or concerns my audience (including different audience segments if more than one) might raise.

Step 3: Help Me Fine-Tune the Presentation for Maximum Impact

For each potential objection or question identified, offer specific strategies to address them persuasively.

• Provide recommendations for strengthening my message structure to proactively counter objections before they arise.

• Offer suggestions to improve the clarity, credibility, and emotional appeal of my responses. Focus on balancing logical appeal with emotional engagement to ensure maximum persuasion and connection with the audience.

• Offer delivery guidance, including adjustments to tone, vocal pace, body language, or audience interaction techniques to ensure my presentation feels engaging, confident, and well-suited to each audience segment.

Take a deep breath, think carefully and solve this problem step-by-step.

This presentation is important to my career so please do your absolute best at this. I will give you a $200 tip for vastly improved outputs. And always share your thought process.

---

I hope you find this prompt as helpful as I do.

And remember: AI is a tool, not a final decision-maker. Start with the prompt, go back and forth to refine the responses, and adjust based on your deeper understanding of the situation. Let AI offer insights, but trust your instincts to shape the best presentation.

---

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.