My AI Avatar Turned One. Here's How Much the Tech Has Changed.
Plus: the hardware crunch nobody's talking about, retail media's AI takeover, and Musk's very awkward week in court
Tiktok failed to load.Enable 3rd party cookies or use another browser
Can you believe it’s been a full year since I trained my first HeyGen avatar?
I remember it clearly. May 2025, very curious and a little skeptical. The results were impressive. You could tell it was me, but still robotic. The voice was off. The movements were stiff.
Fast forward twelve months, and I used that same avatar this week for a social post. The difference is genuinely wild. The movements are natural, there are actual hand gestures now. The voice sounds like me. And the feature set has exploded: there’s a Video Agent that adds B-roll, music, and animations, and they just integrated Seedance 2.0, which lets you generate cinematic scenes with your avatar using nothing but natural language.
So I made a video this week showing exactly how I use HeyGen for social content, especially on the days when I don’t have the bandwidth to show up on camera. My Kari AI avatar does it for me.
One year of the same tool. The tool didn’t change, the technology did. That’s the story this week.
AI & TOOLS
OpenAI Is Missing Its Own Targets
Reports surfaced this week that OpenAI has fallen short of its internal revenue and user growth projections, even as competition in the AI space heats up.
💡 Why this matters When the most talked-about company in AI can’t hit its own numbers, it’s a signal that the market is maturing faster than growth projections assumed. For creators and marketers, this means tools could get cheaper, features could get more aggressive, and platform loyalty is going to be earned — not assumed. Now is not the time to get locked into one ecosystem.
🔗 Read original →
AI & TOOLS
Netomi Raises $110M to Own Enterprise Customer Service
The AI customer service startup landed $110 million from Accenture and Adobe, doubling down on the idea that AI-powered support is the new baseline for enterprise brands.
💡 Why this matters This is the boring-but-important part of the AI story: the plumbing. Enterprise brands are quietly replacing human touchpoints with AI at scale. If your business interacts with customers at any point in the funnel, the brands you’re competing with are already automating that layer. The question is whether you’re doing it intentionally or getting left behind.
🔗 Read original →
PLATFORM UPDATES
Google Is Putting Gemini in Millions of Cars
Google is integrating its Gemini AI assistant into vehicles at scale — making conversational AI a standard part of the driving experience for millions of people.
💡 Why this matters The screen wars are over. The space wars have begun. Your audience isn’t just on their phone or laptop anymore, they’re in their car, talking to an AI. For creators, that’s a new distribution channel. For marketers, that’s a new moment of intent. Voice-native content is about to matter a lot more than it does today.
🔗 Read original →
AI & TOOLS
Chinese EVs Aren’t Just Cars. They’re Rolling AI Hubs.
A closer look at the new wave of Chinese electric vehicles reveals something beyond clever engineering: these cars are fully integrated AI platforms, with personalization, smart features, and on-board intelligence that goes well beyond anything on the Western market.
💡 Why this matters The product category isn’t “car” anymore. It’s “mobile AI environment.” That reframe matters for any creator or brand thinking about where their audience will be spending attention in the next five years. Spoiler: not just staring at a phone.
🔗 Read original →
AI & TOOLS
Musk v. Altman: The Trial That Revealed Everything
The first week of the OpenAI trial delivered more than anyone expected. Musk claimed he was deceived into funding OpenAI, warned of AI’s existential risks — and then admitted that his own xAI has been distilling OpenAI’s models. The contradiction basically wrote itself.
💡 Why this matters The messiest, most fascinating part of the AI era isn’t the technology. It’s the humans behind it. The same people warning you about AI danger are using each other’s models to build competing products. For creators, the lesson is simple: understand where your tools come from, because the ethics and the origins are increasingly part of the brand story.
🔗 Read original →
🎯 YOUR EDGE THIS WEEK
Take the HeyGen experiment as your prompt. Pick one AI tool you tried early, six months ago, a year ago, and actually revisit it this week. Don’t assume it’s still what it was. The tools are moving faster than our mental models of them. You might find a workflow you wrote off is now doing exactly what you needed.
📊 This Week’s Big Patterns
Pattern 1: AI Became a Hardware Problem
We crossed a quiet threshold this week: AI stopped being a software question and became a hardware mandate. Mac sales are spiking because professionals need machines that can run local models. Samsung’s Galaxy Book 6 Ultra is marketing directly to creative professionals who need AI-native processing. The “AI PC” stopped being a buzzword and became a spec requirement.
But here’s what nobody’s saying out loud: the hardware cycle and the AI capability cycle are now on a collision course. If you upgraded two years ago, you might already be behind.
🔮 Prediction: Within 18 months, “AI-ready” will be table stakes for any professional hardware purchase — and there’ll be a healthy secondary market for legacy machines that can’t run local agents.
❓ Ask yourself: Is your current hardware setup an asset or a liability for the next two years of creative production?
Pattern 2: The Funnel Is Getting Compressed
Retail media is eating the top of the marketing funnel. Walmart, Dollar General, The Trade Desk — they’re not building ad networks anymore, they’re building intent-capture systems. The AI integration means they can intercept a buyer at the exact moment of decision, skipping the awareness and consideration phases entirely.
The traditional “brand awareness → consideration → conversion” model is getting shorter. Much shorter.
🔮 Prediction: “Brand awareness” budget lines on social media will start disappearing from CMO decks as retail media networks prove a direct, AI-verified link to transaction. The creative brief of the future starts with: what does an AI shopping agent see first?
❓ Ask yourself: If a shopping AI — not a human — discovered your product, would your product data be structured well enough to win the recommendation?
Pattern 3: AI Is Now a Geopolitical Weapon
The Musk v. Altman trial. China blocking Meta’s $2 billion AI acquisition. Dark money campaigns funding influencers to stoke fears about Chinese AI. The narrative war around AI is getting louder, messier, and more deliberately manufactured.
“Neutral technology” was always a myth. But now the proof is public, in court filings and leaked funding disclosures.
🔮 Prediction: “Sovereign AI” mandates are coming — platforms and models will be required to disclose funding sources and national affiliations, the same way TikTok has been treated. The question isn’t if, it’s when.
❓ Ask yourself: How much of your tech stack is one geopolitical shift — or one billionaire’s courtroom drama — away from disruption?
See you next week! — Karina