What OpenAI and Jony Ive are constructing

Good morning, { AI enthusiasts }. We have known OpenAI and ex-Apple design guru Jony Ive have been constructing AI hardware since last May’s $6.5B deal. What no person knew was what the primary device would actually be.

With recent reporting revealing an upcoming smart speaker that may see, listen, and make purchases, the long-awaited collaboration is finally coming into focus — and it’s heading straight for Amazon, Apple, and Google’s turf.

In today’s AI rundown:

LATEST DEVELOPMENTS

The Rundown: OpenAI-Jony Ive’s first hardware product will reportedly be a $200-$300 smart speaker with a built-in camera and facial recognition for purchases, in line with The Information — backed by a 200+ team aiming to ship it by early 2027.

The main points:

Why it matters: OAI has never shipped a physical product, however the mystique surrounding Jony Ive has made its hardware a hotly anticipated launch. With Apple ramping up AI device plans and Amazon already rolling with Alexa+, OAI’s window to define the category is shrinking fast — making the speaker a vital first swing.

The Rundown You wouldn’t expect a brand new worker to know all the pieces without onboarding, would you? The identical concept goes for AI agents—and metadata is the important thing.

On this ebook, you’ll learn:

Get the Ebook.

The Rundown: The Rundown Roundtable is a weekly feature where we poll members of The Rundown staff about how we use AI in our work and each day lives.

Rowan, Founder & CEO: As a fast-moving startup, a lot of our team’s best ideas come from random Slack threads, but wander away and never fully hashed out. As an alternative of spending hours a day manually adding tasks to our databases, we used Notion’s recent Agents feature (rolling out soon for GA) and built an “AI Project Manager” that monitors Slack messages each day and logs tasks autonomously.

Shubham, Editor: I used Gemini to research my blood work covering some 100 parameters. It immediately extracted and categorized all—from cholesterol to iron levels—into a transparent, structured table, identified subtle deviations (corresponding to barely elevated eosinophils) and explained the clinical significance of every in plain language.

Finally, it translated these data points into a personalised wellness roadmap, suggesting natural lifestyle adjustments to optimize health markers.

Jennifer, Tech & Robotics Author: Over the past yr, I’ve relied heavily on LLMs to support my French citizenship application— including drafting required letters and tracking paperwork. Although I’m already fluent, I used ChatGPT to prep for the required language exam to get aware of the test format for 100% confidence.

The Rundown: On this guide, you’ll learn to establish your individual n8n automation server — helping you run 1000’s of automations per thirty days on n8n with only a $5 virtual server. Better of all, you possibly can set it up in under 10 minutes.

Step-by-step:

Pro tip: You possibly can invite users via email to your n8n server. This makes it great for client- or team-specific projects. You possibly can even save API keys within the server for others to make use of.

The Rundown: Individual AI productivity gains hit a ceiling fast — without orchestration, they do not scale across your org, and leadership has no approach to measure impact or implement security standards. Oz is the platform built to alter that.

Oz’s recent report breaks down:

Download the free report.

The Rundown: AI chip startup Taalas just emerged with HC1, a custom chip built to run a single AI model and nothing else — delivering responses roughly 100x faster than today’s standard hardware and 10x the SOTA for extreme speed in outputs.

The main points:

Why it matters: The model baked into the primary chip is much from competitive, however the tech itself is the story. The speed must be seen to be comprehended (demo here) — and if the approach scales to frontier models, it could change what’s possible in areas like physical AI or agentic workflows where every millisecond matters.

QUICK HITS

Sam Altman called concerns about ChatGPT’s water usage “totally fake”, arguing that creating AI may already be more energy-efficient than raising and ‘training’ a human.

Anthropic opened early access to Claude Code Security, a brand new tool that uses AI to detect hidden software vulnerabilities and suggest patches for human review.

Zyphra released ZUNA, an open-source AI trained on brain wave data that may clean up and reconstruct brain signals, an early step toward thought-to-text without surgery.

Pika Labs launched AI Selves, a brand new product that lets users create persistent AI clones that may post on social media, send messages, and interact across platforms.

Amazon’s Kiro AI coding agent reportedly caused a 13-hour AWS outage in December after autonomously deciding to delete and recreate an environment.

OpenAI’s Head of Codex posted he’s ‘beyond excited’ for the approaching weeks, and that current coding agents can be seen as “so primitive that it can be funny as compared.”

COMMUNITY

Every newsletter, we showcase how a reader is using AI to work smarter, save time, or make life easier.

Today’s workflow comes from reader Gina T. in Minnesota:

“During a recent snowstorm, a giant drift formed right behind my garage stall, and I could not get my automotive out. My husband was out of town, and I had never run the snowblower before. I went to the garage and took an image of the back where all of the controls were and loaded it into ChatGPT, and asked, “How do I start my snowblower?”

It identified the make and model, walked me through the steps on easy methods to start the machine, and just a few questions later, I used to be clearing my driveway!”

How do you employ AI? Tell us here.

See you soon,