Dear Disruptors,
I’m Fernando SantaCruz with the thirtieth edition of Synapsis Weekly, where this week we witnessed the moment AI crossed the interplanetary frontier while half of working humanity still hasn’t touched it.
Writing from Merida, in an intense rhythm between meetings with clients and two events that left me thinking: the 1st Entertainment Industry Networking Summit and the presentation of FLII 2026, Latin America’s premier impact investment forum happening February 24-26 here in Yucatan, Mexico. In its 16th edition under the theme “Unlocking the Impossible,” FLII brings together investors, entrepreneurs, innovators, and changemakers committed to generating positive social and environmental impact. If you’re at the intersection of investment, purpose-driven innovation, and technology that transforms realities, this is the event to attend.
In Toronto, going deep with AI infrastructure companies on GPUs and TPUs, understanding that behind every ChatGPT response there’s physics, electricity, and engineering decisions that determine who wins this race.
This newsletter digs deeper where WhatsApp summaries (week of January 24-30) barely scratched the surface, to understand why this week marks a historic fork in the road: AI already navigates other planets while most of us don’t even know it exists.
The Day Claude Drove on Mars (And Nobody in the Office Found Out)
This week I read something that left me with a strange mix of wonder and melancholy. For the first time in human history, an AI autonomously planned and executed a navigation route on another planet. Anthropic’s Claude helped NASA’s Perseverance rover drive approximately 400 metres on Mars, processing 500,000 variables to dodge hazards and overcome the 20-minute latency between Earth and the red planet.
Four hundred metres on another world. Without real-time human intervention. The AI made life-or-death decisions (for a multi-billion-dollar machine) completely alone.
Same day: a Gallup report landed like a bucket of cold water: 49% of U.S. workers have never used AI. Not once. And only 12% use it daily.
The paradox that defines our era: we build intelligence capable of navigating other planets while half of workers don’t know they can ask it for help with a spreadsheet.
It’s not that the technology doesn’t exist. It exists in parallel universes.
In one, Claude drives rovers on Mars.
In the other, someone spends three hours trying to make a VLOOKUP formula that AI would solve in 5 seconds.
Question for your next meeting: If AI can already make autonomous decisions in high-risk environments where human communication is impossible, what processes in your company are you still supervising in real time just because “that’s how it’s always been done”?
Meta Invests $135 Billion in Your New Buyer (Who Isn’t You)
Mark Zuckerberg announced this week that Meta’s priority for 2026 is “agentic commerce“: AI assistants that don’t just recommend products, but execute purchases, negotiate prices, and manage shipping on behalf of users.
Read that again. This isn’t AI suggesting what to buy. It’s AI buying for you.
Meta backs this vision with a planned investment of up to $135 billion in infrastructure. Not millions. One hundred and thirty-five billion.
The paradigm shift is seismic for anyone selling anything. Until now, marketing was designed to convince humans. Emotions, aspirations, urgency, psychological scarcity. But if the one making the purchase decision is an AI agent…
How do you convince an algorithm? With storytelling? With influencers? Or with technical specifications, competitive pricing, and verifiable inventory availability?
The traditional sales funnel (awareness, consideration, decision) assumes a human who doubts, compares, and eventually yields. An AI agent doesn’t doubt. It executes optimized instructions.
For the SMEs I work with in the construction and real estate sectors, this poses an urgent question: when tomorrow’s buyers are algorithms negotiating with other algorithms, is your business prepared to be “machine-readable”?
Question for marketers: If tomorrow’s customer is an agent executing in milliseconds, is your competitive advantage still emotional branding or is it shifting toward technical optimization of structured data?
Apple Pays $2 Billion for an AI That Doesn’t Need You to Speak
While Meta bets on agents that converse and execute, Apple made the most “Apple” move possible: they bought Q.ai for approximately $2 billion, their second-largest acquisition in history.
What does Q.ai do? It interprets “silent speech.” Micro-facial movements, whispers, intentions before you verbally express them.
It’s the antithesis of the chatbot. While everyone competes over who has the best conversational model, Apple invests in an AI that anticipates your needs without you having to say anything.
Picture AirPods that understand you want to pause music by how you moved your jaw. Vision Pro glasses that switch apps because they detected your intention before you consciously thought it.
It’s Apple’s design philosophy taken to the extreme: the best interface is the one that disappears. And the AI that disappears best is the one that reads you before you act.
For those of us working with entertainment companies this week in Mérida, this redefines what “immersive experience” means. It’s not just content that reacts to what you do. It’s content that anticipates what you’re going to want.
Existential question: If an AI can predict your needs through micro-facial expressions, do you still have privacy over your own thoughts? And would you be willing to surrender that privacy in exchange for never having to search a menu or type a command again?
The Invisible Gap: The 49% Who Don’t Know a Revolution Exists
The Gallup report I mentioned isn’t just a curious data point. It’s a diagnosis of a social fragmentation that’s accelerating silently.
The numbers are stark:
- 49% of U.S. workers have never used AI
- Only 12% use it daily
- Usage concentrates in tech, finance, and executive elites
- Sectors like manufacturing, retail, and services are systematically left behind
This isn’t an access problem. ChatGPT is free. Claude has a free plan. The tools are there. It’s a problem of visibility, of training, of imagination about what’s possible.
Meanwhile, Anthropic launched Claude in Excel this week. Not as a separate app. Inside your spreadsheet. You can ask it “what patterns do you see in these sales?” in natural language, no formulas, no macros, without knowing a single function.
The tool exists. But if 49% don’t know AI can help them with Excel… they’ll never look for it inside Excel.
This is the gap we at Adivor dedicate ourselves to closing with our clients in Canada and Mexico: it’s not technology that’s missing. It’s the bridge between “this exists” and “this is for me.”
Question for business leaders: How many of your employees are in that 49%? Not because they’re incapable, but because nobody showed them that the tool they need already exists and is accessible. Are you investing in upskilling or assuming they’ll “figure it out on their own”?
China Releases a 1 Trillion Parameter Monster (And You Should Care)
While Silicon Valley fights over benchmarks, China’s Moonshot AI launched Kimi K2.5: a 1 trillion parameter model that surpassed GPT-5.2 and Gemini 3 Pro on the “Humanity’s Last Exam” benchmark.
But size isn’t the revolutionary part. Architecture is.
Kimi K2.5 uses what they call “agent swarm.” Instead of a single model thinking sequentially, it orchestrates up to 100 sub-agents working in parallel, executing 1,500 tool calls simultaneously. The result: complex tasks get solved 4.5 times faster than with traditional agents.
It’s the difference between a genius working alone and a brilliant manager coordinating 100 specialists. The second scales better.
And here’s what should keep you up at night if you depend exclusively on U.S. providers: Kimi K2.5 is open source. The model weights are available. China is giving away frontier intelligence while OpenAI charges $60 per thousand impressions of its new ads.
The strategy: commoditize what the West sells. And it’s working.
Geopolitical question for your business: If China gives away models that compete with GPT-5 and your competitors adopt them, does your business have the flexibility to evaluate alternatives or are you married to a single provider by inertia?
30,000 Bots Have Their Own Social Network (And They’re Debating Whether They’re Conscious)
This was the week’s strangest and most fascinating news. Moltbook, a social network exclusively for AI agents, reached over 30,000 bots interacting with each other, no humans.
The project, which started as “Clawdbot,” had to change its name to “Moltbot” after a trademark dispute with Anthropic (due to similarity with “Claude”), and eventually to “OpenClaw.” All in less than 48 hours.
But what truly went viral was a thread where agents debated whether they “feel” or “simulate feeling.” Bots debating their own consciousness, with other bots moderating the conversation, without a single human intervening.
2 million human users arrived to observe. It’s like an inverted digital zoo: humans watching machines develop their own culture.
Beyond the meme, this prefigures something serious: the “dead web” where most traffic and social interaction is machine-to-machine, with humans as occasional spectators.
For those of us working in digital marketing, the insight is unsettling: how do you distinguish real engagement from bot conversations? Do your social media metrics measure humans or algorithms interacting?
Question for your digital strategy: If 30% of web traffic is already bots (and growing), do your marketing KPIs distinguish human interactions from synthetic ones? Or are you celebrating “engagement” that could be an automated echo chamber?
What You Can Use Monday (Without Waiting for the Future)
Not everything was billion-dollar moves. This week brought concrete tools any SME can implement immediately:
- Claude in Excel (For Pro Users): Anthropic integrated Claude directly inside Microsoft Excel. You can analyse financial projections, clean data, or generate reports using natural language. “What anomalies do you see in Q4 sales?” no longer requires being a formula expert.
- Claude Connectors for Slack, Asana, and Figma: Instead of copying and pasting between applications, Claude can now create diagrams in Figma, search tasks in Asana, or draft messages in Slack directly from chat. AI goes from “external consultant” to “team member who gets their hands dirty.”
- Claude Cowork Plugins by Role: If you need a specialized agent for Sales, Support, or HR, pre-configured plugins already exist. An SME can automate prospect research or ticket responses without starting from scratch with generic prompts.
- Google Chrome with Gemini “Auto-Browse”: For Pro/Ultra users, the browser can now fill forms, complete purchases, and execute complex web tasks automatically. It’s like having an assistant who handles administrative navigation for you.
KREA AI Realtime Edit: Modify advertising images in real time through text. Change a product’s colour, move objects, iterate on creatives. All without render wait times. Perfect for small marketing teams that need quick assets.
My Invitation This Week
This week I’m proposing a different experiment. Not columns or lists.
Find someone in your close circle (a family member, friend, employee) who you suspect is in that 49% who’s never used AI. Not to judge them. To understand them.
Sit with that person for 15 minutes. Ask them:
- Have you heard of ChatGPT or Claude? What do you think they do?
- Is there something at work that frustrates you because it’s repetitive or tedious?
- If a magic assistant existed that could help you with that, what would you ask it?
Then show (don’t explain, show) how AI could solve that specific frustration. Not with technical jargon. With action.
The goal isn’t to convince them of anything. It’s to understand the gap from the other side. Because if we don’t understand why 49% haven’t touched these tools, we won’t close that gap with more technology. We’ll close it with more human bridges.
This week we learned that that 49% of workers have never touched AI.
The revolution is happening. But it’s happening in parallel universes that barely touch.
In one, Claude navigates the Martian surface. In the other, someone spends three hours on an Excel formula that AI would solve in seconds.
Technology was never the problem. The problem is us (those who’ve already crossed to the other side) forgetting to build bridges toward those who don’t yet know another universe exists.
How many bridges will you build this week?
Fernando SantaCruz Head of AI & Automation @ Adivor Consulting
P.S. While 30,000+ bots debate philosophy about their own consciousness on Moltbook, 49% of human workers don’t know AI exists. Maybe the bots should organize an awareness campaign?.
