AI Tidbits – From Thinking to Doing: AI’s Next Leap into Everyday Work

AI Tidbits – From Thinking to Doing: AI’s Next Leap into Everyday Work

If 2024 was about using AI and 2025 was about trusting it, then 2026 is shaping up to be the year we delegate to it.

This week’s updates show AI moving beyond single answers into councils, browsers that act on your behalf, and tools that collapse weeks of work into minutes. The common thread? AI is no longer waiting for instructions — it’s starting to execute. Let’s get into it. 👇

 

🥊 Punch Below the Belt? Anthropic Takes a Shot

The ad wars are officially here.

Anthropic has doubled down on its position that Claude will remain ad-free, calling it “a space to think” — no sponsored links, no advertiser influence, no product placements.

To make the point, Anthropic released a four-part satirical video series (including a Super Bowl spot), showing AI assistants awkwardly switching from helpful advice to blatant product pitches — all set to Dr. Dre’s “What’s the Difference.”

The timing is no coincidence.
This comes just weeks after OpenAI confirmed it would begin testing ads inside ChatGPT.

Sam, apparently, is not amused. 😄
Watch all four ads — they’re sharp, funny, and pointed.

 

🧠 Perplexity Launches Model Council

Perplexity has introduced Model Council, a powerful new way to compare AI answers without guesswork.

Instead of relying on one model, Model Council:

• Runs your query across three frontier models simultaneously (e.g. Claude Opus 4.5, GPT-5.2, Gemini 3 Pro)
• Uses a synthesizer model to review all outputs
• Resolves conflicts and delivers one consolidated, higher-confidence answer

This is especially useful when accuracy matters and models disagree — think research, planning, or high-stakes decisions. Limited availability for now but will soon be global.

🌐 Google Supercharges Chrome with Gemini

Google is pushing Gemini deeper into everyday browsing with a major Chrome update.

New Gemini-powered features include:

• 🧭 A cross-tab assistant that works across open tabs and Google apps
• 🤖 Auto Browse, an agentic tool that:
  • Navigates websites
  • Logs into accounts
  • Completes multi-step tasks on your behalf
  • Pauses before purchases for safety

This isn’t just search anymore — it’s assisted execution.

📦 Available now for AI Pro and Ultra subscribers
🎬 Watch 75 seconds of the updates here:
https://www.youtube.com/watch?v=56b9uHAcHYc

 

🔎 Research Anything with Claude in Chrome

Claude is also stepping into the browser.

How it works:

• Install the Claude in Chrome extension
• Open any site (Bayut, Amazon, research portals, etc.)
• Click the Claude icon and type your request in plain English

Prompt:

I’m looking for a 1-bedroom house in Dubai Downtown under AED 900K, with a parking and at least 900sqft. Can you search Bayut and show me the top 5 options?

 

Claude can:

• See the page you’re on
• Click, scan listings, and gather data
• Return structured results in seconds

Search is no longer just about links — it’s about delegation.

Life is changing.
So is search. 😊

 

⚡ Vibecode Goes Full-Stack Vibe Coding

Vibecode just rolled out what it claims is the first full-stack vibe coding platform.

What’s new:

• One-click web deployment
• Mobile-first platform with 5,000+ apps already published
• Powered by Claude Code
• Transforms plain-English prompts into:
  • Frontend
  • Backend
  • Database
  • Responsive layouts

This is not a toy — it’s a production pipeline in disguise.
Take it for a spin here.

 

📸 Turn Selfies into Professional Headshots (Nano Banana Pro)

You can now generate polished, studio-quality headshots using Nano Banana Pro.

Quick setup tips:

• Neutral background
• Direct light or window light
• Chest-up framing, lens slightly above eye level

Steps:

• Go to Google AI Studio
• Click Playground
• Add an API key with Gemini + Nano Banana Pro enabled
• Upload your photo and use this prompt:

Prompt:

“Generate: Post-processing enhancement, professional color grading, balanced studio lighting, remove noise and grain, sharpen focus, upscale to 8k, skin texture refinement, subtle dodging and burning, clear and crisp details, maintain original facial structure and clothing, photorealistic”

💡 Pro Tip:
Once you like a result, regenerate in 4K, or ask Gemini to create four headshot concepts using the same principles — clearly stating you want to edit photos.

 

🎬 Kling 3.0 Is Worth a Look

Kling has released Kling 3.0, consolidating:

• Text-to-video
• Image-to-video
• Native audio generation

All in one multimodal model.

Highlights:

• Longer outputs (up to 15 seconds)
• Improved character & scene consistency
• Storyboarding + control tools built in

This is another step toward real production workflows — not just demos.

🔗 https://x.com/Kling_ai/status/2019064918960668819?s=20

 

🖼 Restore Old Photos with ChatGPT

Bring old memories back to life with this prompt.

Prompt:

“[upload old photo] Restore this old photo into a professional portrait of DLSR – quality colour and detail, using an advanced upscaling algorithm comparable to the results from a Canon EOS R6 II. Ensure the restored image looks natural, retains exact facial features, has great clarity……”

 

Prophecy or Coincidence??

 

 

Closing Thoughts

Across models, browsers, coding tools, and creative workflows, one thing is clear: AI is shifting from answering questions to taking action.

The skill gap in 2026 won’t be about who knows the tools —
It’ll be about who knows when to let them work for you.

See you next Tuesday. 👋