Skip to content
Unfold Byte UB markUBUnfoldByte
Model watch

Apple links AI mentions to device upgrades in record fiscal year results

Apple links AI mentions to device upgrades in record fiscal year results
Model watch

Will £20bn Drax AI hub deliver 1GW capacity and direct power by 2031?

A £20bn deal aims to build a major AI facility at the Drax site, with operations potentially starting in early 2028.

Will £20bn Drax AI hub deliver 1GW capacity and direct power by 2031?
Model watch

China Plans Comprehensive AI Safety Standards for Domestic Model Makers

Beijing intends to introduce broad regulations covering the full lifecycle of artificial intelligence systems, following recent global security concerns.

China Plans Comprehensive AI Safety Standards for Domestic Model Makers
Model watch

How AI wearables process and store conversation data

Amazon’s Bee Frontier deletes raw audio after real-time processing, keeping only encrypted summaries.

How AI wearables process and store conversation data
Model watch

OpenAI reports Heimstaden deploys ChatGPT Enterprise for property management

A customer story by OpenAI details how Heimstaden integrates AI into European residential real estate operations to reduce costs and support staff.

OpenAI reports Heimstaden deploys ChatGPT Enterprise for property management
Model watch

How does Sequoia-backed Trajectory enable non-technical users to optimise AI models?

Sequoia leads $45M Series A for Trajectory at $300M valuation to help non-technical users optimise AI models.

How does Sequoia-backed Trajectory enable non-technical users to optimise AI models?
Model watch

Does LingLong Insight 1.0 Validate Industrial Supply Chains Using Local AI?

ZKH and Intel launch a four-billion-parameter multimodal model for verifying industrial goods receipts using on-device processing.

Does LingLong Insight 1.0 Validate Industrial Supply Chains Using Local AI?
Model watch

Can AI systems genuinely understand human emotions or just mimic them?

Analysis suggests emotional intelligence may become more critical as teams navigate AI outputs that simulate empathy without true comprehension.

Can AI systems genuinely understand human emotions or just mimic them?
Model watch

Anthropic discloses $895.4bn compute commitments in amended SEC filing

The AI lab’s updated S-1 shows nearly all capital is tied to infrastructure leases and hardware purchase agreements.

Anthropic discloses $895.4bn compute commitments in amended SEC filing
Model watch

Scott Alexander argues safety labs should continue advancing AI models rather than stopping

LessWrong post distinguishes transition plans and argues against unilateral cessation of AI development by safety-focused laboratories.

Scott Alexander argues safety labs should continue advancing AI models rather than stopping
Model watch

Google suspends engineer over LaMDA sentience claims

Blake Lemoine’s assertion that the chatbot is sentient led to his suspension and a firm denial from Google’s internal review teams.

Google suspends engineer over LaMDA sentience claims

Also on the desk

  1. Model watch

    How does OpenAI’s five-value framework align advanced AI reasoning with human intent?

  2. Model watch

    Can a $50 Waveshare pendant replicate Meta’s Muse Charm functionality?

  3. Model watch

    Hugging Face tests Parakeet and Qwen3-ASR in Swift speech APIs