AI DUNN Right Weekly - Issue #19
- Feb 2
- 6 min read
Practical AI insights for business growth
Hey AI Innovators! 👋
This week showed us what happens when established companies stop watching AI from the sidelines and start betting their future on it.
Airtable launched its first new product in 13 years. Chrome put Gemini in your sidebar. Google gave everyone the power to build interactive worlds. And Musk made bold predictions that got everyone talking about AGI timelines.
Here's what matters for your business right now.
Big Story
When a $4 billion company bets everything on multi-agent AI
Airtable just did something most mature companies won't.
They launched Superagent, their first standalone product in 13 years. Not a feature add-on. Not an upgrade. A completely new product.
Here's what I think...
This is a company that's been around since 2013, serves 80% of the Fortune 100, and still has $700 million in cash deciding that multi-agent coordination is important enough to build a separate product line.
CEO Howie Liu isn't hedging his bets either. He's on record saying Superagent could become bigger than Airtable itself.
Here's how it works:
You ask one question.
The system builds a research plan, deploys multiple specialized agents working in parallel, then delivers an interactive report you can explore.
Ask about expanding your business into a new market. Superagent identifies what needs investigating, then deploys agents simultaneously:
Agent 1 handles financials
Agent 2 analyzes competition
Agent 3 reviews demographics
Agent 4 maps timelines
Agent 5 digs into regulatory issues
The output is an interactive analysis with data you can filter, explore, and drill into.
What makes this different:
Most "agents" today are just workflows with AI calls mixed in. They follow predetermined steps and can't adapt when they hit something unexpected.
Superagent uses premium data sources like FactSet, Crunchbase, SEC filings, and earnings transcripts. It's built for serious business research, not casual questions.
The business context:
Airtable's valuation dropped from $11.7 billion in 2021 to around $4 billion on secondary markets. But they're still profitable and have significant cash reserves.
They hired OpenAI's former engineering lead for ChatGPT's business products as CTO last fall. They acquired DeepSky (an AI agent startup) at the same time.
This is what strategic repositioning looks like when you still have resources to execute.
Pricing follows the emerging pattern:
$20/month for entry tier
Up to $200 for power users
Generous inference credits included
Bottom line is...
When a mature company with hundreds of millions in the bank decides multi-agent AI is worth building a separate product line, that tells you where enterprise software is headed.
Single-agent processing is already being positioned as outdated. Multi-agent coordination is becoming the standard for complex work.
What's New This Week
Google's Project Genie lets anyone build interactive worlds
Google launched Project Genie for AI Ultra subscribers in the U.S.
You describe what you want in text. The AI generates an explorable, interactive environment that builds itself in real-time as you move through it.
What this means:
Creating interactive 3D environments used to require game engines, modeling skills, and coding knowledge. Project Genie removes those barriers completely.
Describe a historical setting. Walk through it. Need a training simulation. Build it with a prompt. Want a virtual showroom. Generate it without technical expertise.
The technology is early. 60-second generation limits and inconsistent physics show it's not production-ready yet.
But the direction is clear... by 2027-2028, creating explorable worlds will be as common as generating images.
Chrome puts Gemini in a persistent sidebar
Chrome integrated Gemini into a sidebar that stays open while you browse.
You can ask about the current website, compare tabs, and get answers based on your Gmail, Search, YouTube, and Google Photos data.
The ambitious feature is auto-browse. AI agents that shop for you, find coupons, and complete tasks across websites.
Reality check:
Browser-based agents are finicky. They often fail to complete tasks. Google's demos involved shopping and travel planning, like most agent demos.
But Chrome has the largest browser market share globally. When they integrate AI features, millions get them by default.
AI browser startups built sidebar assistants and automated tasks as differentiators. Google just matched those features in the browser everyone already uses.
xAI closed $20 billion Series E
Elon Musk's AI company completed a massive funding round, cementing xAI as a major competitor to OpenAI and Anthropic.
At Davos, Musk doubled down on his timeline: AI will surpass individual human intelligence by end of 2026, and collective human intelligence by 2031.
What "smarter than humanity" actually means:
Musk conflates three different milestones:
Narrow superhuman performance (already here for specific tasks)
General human-level intelligence across any task (the 2026 claim)
Superintelligence that surpasses collective human knowledge (2031)
The technical bottlenecks he ignores:
Energy efficiency gap (current models use 1000x more power than human brains)
Test-time compute requirements (seconds per decision vs. real-time needs)
Data saturation (running out of internet-scale training data)
What 2026 will actually deliver:
Specialized superhuman agents handling specific factory tasks autonomously. Enterprise systems handling 80% of knowledge work without human intervention. Multimodal reasoning that rivals human spatial understanding.
Not: A single model reasoning across every domain without retraining.
Zuckerberg positions AI as the new social media
Meta's earnings call revealed Zuckerberg's vision for AI as the next evolution of social platforms.
He's teasing agentic commerce tools and major AI rollout in 2026. The idea... AI becomes the interface layer for how people interact with businesses and each other.
WhatsApp will charge AI chatbots to operate in Italy, signaling monetization strategy for AI integrations.
Tool of the Week
Chrome with Gemini sidebar (free for AI Pro/Ultra subscribers)
If you're already using Chrome, this changes how you work with AI.
The sidebar stays persistent while you browse. Ask questions about websites, compare information across tabs, get answers based on your Gmail and calendar data.
The auto-browse feature:
AI agents handle tasks for you by navigating websites, finding coupons, filling forms, and completing purchases.
It asks for intervention when performing sensitive tasks like logging in or final checkout.
Available now:
Gemini sidebar support rolling out today to AI Pro and Ultra subscribers in the U.S.
Auto-browse agentic features coming to same tier.
Why this matters:
You don't need to download a separate AI browser. The features integrate into the browser you already use on billions of devices.
For AI browser startups, this is Google leveraging distribution advantage. For users, it means AI assistance becomes built into your existing workflow.
Quick Hits Worth Your Time
→ OpenAI's annualized revenue surpassed $20 billion in 2025, operating 1.9 gigawatts of compute capacity with record weekly active users.
→ Anthropic extended Claude into healthcare with HIPAA-compliant tools and clinical modules for regulated domains.
→ EU formally delayed high-risk AI Act implementation from August 2026 to August 2027-2028, expanding regulatory sandboxes for SMEs.
→ Kenya's AI-powered land registry system cut fraud by 90% in pilot regions, showing practical governance applications.
→ Merge Labs secured $250 million seed funding led by OpenAI for brain-computer interface technology.
→ Hugging Face launched "Inference at the Edge" for on-device model deployment on smartphones and micro-devices.
Prompt of the Week
The Multi-Agent Research Planner
Stop doing research manually when AI can coordinate multiple specialists for you. This prompt breaks complex research into parallel workstreams:
Act as a strategic research coordinator. I need comprehensive analysis on [RESEARCH TOPIC] for [BUSINESS DECISION].
Essential Details:
• Topic: [SPECIFIC SUBJECT]
• Business Context: [WHY THIS MATTERS]
• Decision Timeline: [WHEN NEEDED]
• Key Questions: [WHAT YOU NEED TO KNOW]
• Data Sources Available: [WHAT YOU HAVE ACCESS TO]
Create one research coordination plan including:
1. Core research questions (what needs answering)
2. Parallel research workstreams (what can happen simultaneously)
3. Data sources for each workstream (where to look)
4. Key metrics to track (what matters most)
5. Red flags to surface (what could derail the decision)
6. Synthesis framework (how to combine findings)
Coordinate research like Superagent does... multiple specialists working in parallel, synthesized into actionable intelligence.
Why this works:
This prompt forces you to think like a multi-agent system. What can happen in parallel? What needs sequential investigation? Where are the dependencies?
You get better research faster because you're coordinating workstreams instead of doing everything serially.
Best use case:
Market analysis, competitive intelligence, expansion decisions, vendor evaluations, strategic planning, investment research.
Anything where comprehensive analysis requires looking at multiple dimensions simultaneously.
My Take
The biggest shift happening right now is understanding that AI is moving from reactive tools to operational systems.
Airtable bet their company on it. Chrome integrated it into your daily workflow. Google made world creation accessible to everyone.
These are strategic bets from established companies with real resources and real customers.
For business owners, the question shifts from "should we use AI" to "how do we design systems where AI handles entire workflows based on rules we set once."
That's a completely different way of working.
Most people are still manually prompting AI for every single task. The operators who get ahead will be designing systems where AI handles judgment calls, coordinates multiple specialists, and reduces manual decision-making over time.
Start thinking in responsibilities instead of tasks.
That's where this is headed.
That's it for Issue #19!
This week proved mature companies are betting big on multi-agent AI, browsers are becoming AI platforms, and world creation is moving from technical expertise to simple prompts.
The businesses that win will be the ones who know how to design AI systems that run operations without constant supervision.
Stay innovative, Jackie @ AI DUNN Right
P.S. - Airtable launching Superagent as a separate product tells you something important: multi-agent coordination is significant enough to build an entirely new business line. When a mature company makes that move, pay attention.
Comments