PLUS: Signal's creator builds a private AI and Meta's metaverse retreat
Good morning
OpenAI's Sam Altman and former Apple designer Jony Ive are reportedly collaborating on their first piece of AI hardware. New details suggest an earpiece-style gadget designed to launch a new category of personal, AI-native devices.
The device may operate as a standalone gadget, aiming to replace common phone actions with voice commands. The biggest question is whether a new form factor can successfully challenge the convenience of smartphones and traditional earbuds.
In today’s Next in AI:
Jony Ive and Sam Altman's AI hardware plan
A new private AI from Signal's founder
Meta retreats from its metaverse ambitions
Microsoft’s new data center energy plan
The 'AirPod Killer' From Jony Ive and Sam Altman
Next in AI: OpenAI's Sam Altman and former Apple designer Jony Ive are reportedly developing their first AI hardware, an earpiece-style device codenamed 'Sweetpea,' based on new leaked details. This collaboration aims to create a new category of personal AI-native gadgets.
Decoded:
The device reportedly features a unique behind-the-ear design with two pill-shaped earpieces that are stored in an egg-shaped charging case.
It may be powered by a 2nm smartphone-class chip, hinting that it could operate as a standalone device rather than just a phone accessory.
Powered by a ChatGPT-like assistant, the gadget's goal is to replace phone actions by using voice commands, positioning it as a direct competitor to Siri and existing earbuds.
Why It Matters: This collaboration signals a major effort to define the next wave of personal computing hardware beyond the smartphone. The biggest test will be convincing users to embrace a new form factor that challenges the convenience of phones and traditional earbuds.
Signal's Creator Builds a Private AI
Next in AI: Moxie Marlinspike, the mind behind Signal, has launched Confer, a new open-source AI assistant designed with end-to-end encryption to keep your conversations completely private. This initiative brings the principles of secure messaging to the world of AI.
Decoded:
It tackles the problem of major AI platforms becoming a “data lake” for sensitive user information, where even deleted chats can be subject to legal requests. Marlinspike argues users need a space to interact with AI without confessing their secrets to a corporate database.
Confer uses a combination of on-device encryption via passkeys and Trusted Execution Environments (TEEs) on its servers. This ensures that no one—not even Confer's own administrators—can access the content of your prompts or the AI’s responses.
As a fully open-source project, its code is available for public inspection, offering cryptographic proof that its privacy promises are being kept. This transparency allows users to verify the security measures themselves.
Why It Matters: Confer provides a crucial alternative in an AI landscape dominated by data-hungry models. This focus on privacy could empower professionals to use AI for sensitive tasks they previously couldn't.
Meta's Metaverse Retreat
Next in AI: Meta is laying off over 1,000 employees in its Reality Labs division, signaling a major strategic pivot away from its costly metaverse ambitions. The company is now redirecting its resources toward developing AI-powered wearables and mobile features.
Decoded:
The decision comes after the Reality Labs division reported staggering losses of over $70 billion since the start of 2021, pushing the company to find a more sustainable path forward.
Meta is now doubling down on AI-powered smart glasses, discussing plans with partner EssilorLuxottica to significantly increase production of its Ray-Ban and Oakley lines.
The metaverse concept itself isn't dead but is moving to mobile, as the Horizon team shifts to bring virtual experiences and AI creator tools directly to smartphones to reach a much larger user base.
Why It Matters: This move signals a broader industry shift from speculative, long-term virtual worlds to practical, near-term AI applications. The new focus is on augmenting the reality we have today rather than trying to replace it entirely.
Microsoft's AI Power Play
Next in AI: Microsoft is tackling the growing concern over AI's massive energy footprint by vowing to reject local subsidies and pay the full utility costs for its expanding data center network.
Decoded:
The move is a direct response to growing hostility in local communities concerned about new data centers straining power grids and water supplies.
As part of a new five-point plan, the company will pay higher utility rates and explicitly promises it won’t ask for local property tax subsidies to fund its AI infrastructure.
Microsoft isn't just changing its own policy; President Brad Smith is meeting with federal lawmakers to champion this "pay our own way" approach for the entire industry.
Why It Matters: This is a significant shift in how Big Tech finances its infrastructure, setting a new standard for corporate responsibility in the AI era. It's a calculated strategy to earn social license and accelerate AI development by addressing community concerns head-on.
AI Pulse
Amazon detailed its vision for a more personal and conversational Alexa+ that uses memory and context to better compete with assistants like ChatGPT and Gemini.
NVIDIA announced a first-of-its-kind AI co-innovation lab with pharmaceutical giant Lilly to accelerate drug discovery using large-scale biomedical foundation models.
The Pentagon announced that it will integrate xAI's Grok into military networks for both unclassified and classified systems later this month as part of a new AI acceleration strategy.