What’s New in AI?
Let’s be real: keeping up with AI tools is a full-time job now. Blink and you miss a breakthrough. Or a breakdown. Whether you’re chasing creative flow, shaving hours off your workflow, or just trying to stay relevant—it moves fast out here.
But you’re in the right spot.
This week, the biggest players leveled up. We’re talking copilots with memory, brand-genie web apps, and video models reacting to your voice in real-time. Plus, a few surprises. Let’s dig in.
Microsoft Copilot finally gets fun (and smart)
Used to be, you opened Copilot because your company set it as the default. Now? You might open it because it’s actually helpful.
Here’s what’s new in the refreshed desktop and mobile Copilot app:
- Memory — Remembers your previous chats for smoother, more contextual replies.
- Connectors + Actions — Link to external data and tell Copilot to do things, not just suggest them.
- Groups — Chat with teammates inside the same convo and co-pilot your AI.
- Image morphing — Create and remix visuals directly in the chat.
- “Mo” the Avatar — A friendly interface that wink-wink transforms back into Clippy if you tap enough.
The takeaway? Copilot is stepping out of its Microsoft suit and into assistant-mode—with nuance, memory, and just enough personality.
Google’s Pomelli is Canva with a brain (and your brand)
Imagine if a brand-savvy designer watched your website, learned your style, and then whipped up a cross-platform marketing campaign in under 30 seconds.
That’s Pomelli—quietly launched via Google Labs.
Here’s what it pulls off:
- Scans your current public site.
- Extracts your unique fonts, colors, and tone.
- Instantly generates social posts, product descriptions, and visuals all on-brand.
Rough edges? Yep. Week one saw server strain and image-output bugs. But when it clicks, it’s a beast for bootstrapped teams and solo founders.
ChatGPT Pulse: The AI that nudges, not waits
You already talk to ChatGPT. Now, it talks back—with summaries and personalized nudges.
If you’re a ChatGPT Pro or Enterprise user, you’ll spot Pulse on the left sidebar on desktop. Here’s what it does:
- Weekly recaps — “Looks like you’ve been focused on building that pitch deck.”
- 1-minute resets — quick exercises to shift mental gears.
- Goal tracking — Remember when you said “learn Spanish”? Pulse does too, and it brings it up at the right time.
It’s early—but this is a glimpse of proactive AI. Less “Ask me anything,” more “Here’s what you might need.”
Adobe’s Firefly Studio wants to be your creative HQ
Forget juggling tabs between image, audio, and video tools. Adobe’s new Firefly Studio brings everything under one sleek roof.
Highlights from Adobe MAX:
- Integrated image, video, and text tools.
- Plug-and-play models inside Photoshop—including third-party ones like Runway and Pika.
- One-click “Quick Actions” like de-noising, captioning, and auto-clipping.
- Trained only on copyright-cleared data — fewer legal headaches down the line.
If you’re a creative bouncing between tools, this just cut your tab count in half.
Claude for Excel = The AI spreadsheet whisperer
Anthropic is testing a dedicated Claude add-in for Excel, and insiders are buzzing. It’s still wait-list only, but early users already swear by it:
- Explain complex data with natural prompts.
- Clean and classify messy rows in seconds.
- Generate formulas across multiple sheets—no syntax gymnastics required.
Claude might just skip past every built-in cell assistant. Watch this space.
Grokipedia: The AI encyclopedia you can’t edit
Say hello to Grokipedia, a vault of 900,000 machine-written reference entries generated by xAI’s Grok model.
Yep—no editors, no rewrites, no edit wars. Just clean data, locked in.
Here’s how it stacks up:
- Closely mirrors Wikipedia’s outlines, minus the human bias.
- Better coverage on sensitive topics (no moderation filter fluff).
- No images yet—but the text is surprisingly readable.
Whether it replaces Wikis or just supplements them, it’s a sign of where reference content is heading.
Orisa 2 reacts to your imagination… live
“A penguin hosting a talk show in an underwater jazz club.”
Type it—or say it out loud. Orisa 2 listens and transforms your prompt into a video…in real-time. No waiting, no renders. And yes, it updates on the fly: “Now add audience applause and neon jellyfish.”
It’s goofy. It’s magic. It’s a whole glimpse into live, voice-guided virtual worlds.
Fair warning: the servers are twitchy. Get in while it’s hot.
Music makers: Udio just locked your files
Here’s one for the “local backup or bust” folder.
AI music tool Udio quietly changed its download policy. Now, unless your track includes stems from Universal Music’s licensed catalog, you can’t export it.
Bottom line: If you’re composing with AI, export everything. Cloud’s not your friend when terms change overnight.
Quick Hits You Should Know
- Seed3D 1.0 launched with sharper AI text-to-3D support. You’ll need a Chinese phone number to sign up. Yeah, kind of a bummer.
- Netflix is staffing up AI ad-tech teams. Automated trailer cuts inbound?
- A crew of ex-Meta engineers just joined OpenAI. Expect smarter ad personalization via ChatGPT Memories.
- Google Earth Engine + AI now lets you talk to climate data—literally. Try a query like “Show U.S. coastal regions with ≥ 20% flood risk in 2030.”
- HighLevel 2.3 brings better motion and 4 free videos/day to dancing-text-to-video creators.
The Bottom Line
From spreadsheets to storyboards, AI is evolving from assistant to co-creator. The real win? You’ve now got tools that actually save time and sharpen your output—not just demo fodder.
Want to not just follow the trend but master it? Learn how to use these tools without burning out. Start exploring AI the beginner-friendly way with guided lessons over at tixu.ai.
Ready when you are.
Leave a Reply