Hey friends,
Welcome back to cmd+growth.
This week felt like a preview of where the next few years are heading.
AI companies want their own devices. Meta is still copying competitors. LinkedIn keeps rewarding deeper engagement over vanity metrics. And AI agents are getting increasingly capable of handling work without constant supervision.
Let’s get into it.
Today’s topics
- 📱 OpenAI’s hardware ambitions get more serious
- 🤖 AI agents keep getting more autonomous
- 📸 Meta launches another Snapchat competitor
- 🎯 Why marketers may be targeting the wrong variable
- 📊 LinkedIn engagement keeps evolving
Top Story
📱 OpenAI’s device plans are becoming harder to ignore
For years, AI companies have mostly lived inside websites and apps.
That may not be true for much longer.
OpenAI continues to move toward hardware following its acquisition of Jony Ive’s io team, and multiple reports suggest the company plans to unveil its first device in late 2026. :contentReference[oaicite:0]{index=0}
The interesting part is not the gadget itself.
The interesting part is what happens if AI becomes the primary interface between users and the internet.
For the past fifteen years, marketers optimized for search engines, social feeds and app stores.
An AI-first device creates an entirely new gatekeeper.
Recent reporting suggests OpenAI has explored a family of AI devices beyond traditional phones and laptops, while CEO Sam Altman has previously indicated the product will be a new category rather than a smartphone. :contentReference[oaicite:1]{index=1}
If users start asking an AI assistant to research products, compare vendors, book services and make purchases, distribution changes completely.
The real question is no longer whether AI can answer questions.
It is whether AI becomes the layer that sits between users and everything else.
Read more:
🤖 AI agents keep getting more capable
One of the most useful AI metrics right now is not benchmark scores.
It is autonomy.
The question is simple: how long can an AI system work independently before it gets stuck?
The answer keeps increasing.
That matters because AI agents are no longer competing with search engines.
They are competing with workflows.
The longer they can operate independently, the more realistic it becomes for them to handle research, planning, analysis and execution without constant supervision.
The next phase of AI will probably look less like asking questions and more like assigning projects.
📸 Meta launched another Snapchat clone
Meta has officially rolled out Instants, a disappearing-photo product that looks very familiar to anyone who has used Snapchat or BeReal. The feature allows users to send photos that can only be viewed once and disappear after 24 hours. :contentReference[oaicite:2]{index=2}
Meta initially tested Instants as a standalone app in Spain and Italy before expanding the feature. :contentReference[oaicite:3]{index=3}
Will it work?
Maybe.
But the more interesting story is that social platforms continue moving toward private sharing.
Public feeds still matter, but much of the internet’s activity has shifted into DMs, groups and smaller circles.
For marketers, that creates a challenge.
You cannot optimize for conversations you cannot see.
Yet those conversations increasingly influence discovery and purchasing decisions.
Read more:
🎯 Marketers may still be targeting the wrong variable
One trend that keeps showing up across advertising research is that life-stage signals often outperform demographics.
Most campaigns still begin with age ranges.
25-34.
35-44.
45-54.
But buying behavior is often driven by events rather than age.
Moving house.
Starting college.
Getting married.
Having children.
Those moments often predict purchase intent better than demographics alone.
The lesson is not new.
What is surprising is how many campaigns are still built primarily around age buckets.
📊 LinkedIn keeps rewarding deeper engagement
LinkedIn continues moving away from vanity metrics.
Likes still matter, but increasingly the platform appears to care more about behaviors such as clicks, dwell time, carousel interactions and meaningful engagement.
That mirrors what we are seeing across social media generally.
Platforms are becoming better at measuring whether people actually care about content instead of simply reacting to it.
The result is that engagement bait gets weaker over time.
Useful content keeps getting stronger.
⚡ Quick hits
- OpenAI continues expanding its hardware ambitions beyond software alone. :contentReference[oaicite:4]{index=4}
- Meta’s Instants rollout shows social platforms are still searching for the next private-sharing format. :contentReference[oaicite:5]{index=5}
- AI agents are steadily becoming more autonomous and workflow-oriented.
- The race to become the default AI interface is accelerating.
My take
The story this week is not really AI.
It is distribution.
OpenAI wants to become the interface.
Meta wants to own more private communication.
AI agents want to handle tasks instead of simply answering questions.
Different products.
Same battle.
The companies that control discovery keep getting stronger.
The companies that sit between users and action keep becoming more valuable.
And for marketers, that means one thing:
Distribution is becoming just as important as creation.
See you next week.
Diana Yefremova is an independent SEO consultant and the creator of Cmd+Growth. She writes about SEO, AI search, PPC and digital marketing, with a focus on the updates and changes marketers actually need to know.
Each week, she curates the most useful stories, platform updates and industry conversations so readers can keep up without spending hours digging through everything themselves.
