March just rewrote the rules. Did you catch it?

Hey there!

March 2026 has been, without exaggeration, one of the most packed months in AI history. Three frontier models dropped in under four weeks. A major AI product quietly shut down. And an infrastructure protocol hit a milestone that signals something much bigger than most people realize.

If you blinked, you missed a lot. Here's what actually matters this week:

3. Signals from the industry you need to internalize

2. Myths vs. what we're actually seeing

1. Something fresh from us

Three signals from the industry this week :

✦  MCP just hit 97 million monthly SDK downloads. Agents are no longer a demo.

Up from 2 million at launch in November 2024, MCP is now adopted by every major AI provider OpenAI, Google, Microsoft with 10,000+ servers live. It's becoming the invisible plumbing of the agentic layer, the same way HTTP was for the web. If your product doesn't speak agent yet, this number is your deadline. 

✦ OpenAI shut down Sora and the real story is messier than the headlines.

$15M/day in compute costs. $2.1M in lifetime revenue. Downloads down 66% from peak. Add deepfake scandals, a collapsed $1B Disney deal, and a strategic pivot to AGI and the shutdown makes sense. The lesson: not every capability that can exist at scale, should. At least not yet.

✦ Only 15% of pages ChatGPT retrieves actually get cited. SEO just changed shape.

85% of webpages pulled during a ChatGPT query never make it into the final answer. Ranking well isn't enough, your content needs to be the kind AI systems actually select. That's a completely different game, and most teams haven't started playing it.

Two myths vs. what we’re actually seeing

Myth #1: A bigger, more powerful model always means better output for your use case.

Reality: This month, Claude Sonnet 4.6 scored higher than Opus 4.6 on real-world office work benchmarks despite being the smaller, cheaper model. Efficiency is the new intelligence. The smartest teams aren't chasing the largest model. They're chasing the right one for the job.

Myth #2: AI is replacing people wholesale, layoffs equal automation.

Reality: Atlassian cut 1,600 people (10% of its workforce) while appointing two new AI-focused CTOs on the same day. The CEO's words: "AI has fundamentally changed the mix of skills we need." It's not elimination, it's redefinition. And the gap between teams that retool now versus later is compounding fast.

One fresh update from us:

Ads to Landing Page Personalization

Your ads already know who clicked. Now your landing page does too.

Fibr's Ads to Landing Page Personalization reads the signal behind every ad click and adapts your landing experience automatically — to the exact intent that brought each visitor there. No manual pages. No dev sprint. Build and deploy personalized pages across all your campaigns.

The last mile of performance marketing, finally closed.