Good morning, AI enthusiasts. HTC just launched VIVE Eagle AI glasses with dual OpenAI and Google assistants, taking direct aim at Meta's successful Ray-Ban smart eyewear line.
With premium pricing and advanced AI capabilities, could this mark the beginning of a real smart glasses war?
In today's recap:
HTC launches premium AI glasses to challenge Meta
Claude gains power to end abusive conversations
Stanford AI slashes dangerous pharmacy errors
8 new AI tools and more
HTC
HTC's VIVE Eagle takes direct aim at Meta's Ray-Ban glasses
Source: HTC
Recaply: Taiwanese tech giant HTC just launched VIVE Eagle AI glasses to compete directly with Meta's successful Ray-Ban smart eyewear line. The new glasses offer advanced AI features at a premium price point.
Key details:
Users can switch between OpenAI and Google AI assistants.
A built-in camera enables voice-activated photography.
Real-time translation works across 13 languages offline.
The glasses are priced at $520 versus Meta's $300 Ray-Ban glasses.
They're currently available only in Taiwan through retail partners.
They feature an open-ear audio design for all-day comfort.
Why it matters: Competition in smart glasses heats up as HTC challenges Meta's market lead. Premium pricing suggests a focus on tech-savvy early adopters willing to pay more for advanced AI capabilities.
ANTHROPIC
Claude gets "hang up" power for abusive chats
Source: Anthropic
Recaply: Anthropic just gave Claude Opus 4 and 4.1 the ability to end conversations when users become persistently harmful or abusive. This marks one of the first AI welfare features deployed in consumer chatbots.
Key details:
The feature triggers only after multiple redirection attempts fail on harmful content requests.
Testing showed Claude exhibited distress patterns when processing requests about violence, terrorism, or minors.
Users retain full account access and can immediately start new conversations.
Safeguards prevent ending chats when users show self-harm risk or imminent danger.
The vast majority of users won't encounter this feature during normal product use.
Users can edit previous messages to create new branches of ended conversations.
Why it matters: This deployment represents early steps into AI welfare research as models become more sophisticated. While uncertainty remains about AI consciousness, Anthropic is implementing low-cost interventions that could prove important if model welfare becomes a reality.
AI RESEARCH
AI copilot slashes pharmacy prescription errors
Source: Getty Images
Recaply: Stanford researchers just developed MEDIC, an AI assistant that dramatically improves prescription accuracy at pharmacies. The system specifically addresses instruction translation errors that lead to patient harm and billions in healthcare costs.
Key details:
Standard AI models showed higher error rates than rule-based pharmacy systems.
MEDIC combines pharmaceutical domain knowledge with language processing.
MEDIC was tested through Amazon Pharmacy's actual prescription production pipeline.
Small language changes can create major medication dosage problems.
It follows a Swiss cheese safety model, where multiple systems are needed to prevent errors.
The technology enables pharmacists to focus on patient care over data entry.
Why it matters: MEDIC could transform medication safety by addressing a critical weak point in healthcare delivery. The system offers a scalable solution to prescription errors while supporting rather than replacing essential human oversight in pharmacies.
NEWS
Beyond the big stories in AI today
OpenAI updated GPT-5 to feel warmer and friendlier after user feedback said it was too formal.
Higgsfield AI launched Product-to-Video, allowing users to place products directly into video frames without prompts.
ElevenLabs introduced Video-to-Music in Studio, generating custom soundtracks based on video content in one click.
IBM and USTA debuted AI-powered fan experiences for the 2025 US Open, including Match Chat assistant for real-time help.
StepFun AI introduced NextStep-1, a 14B transformer for image generation that achieves new benchmarks for autoregressive models.
Meta FAIR introduced DINOv3, a computer vision model that outperforms specialized solutions on multiple dense prediction tasks.
TOOLS
Trending AI Tools
✍️ CopySpace - AI-powered SEO tool that optimizes your articles instantly
🤖 ChatFast - Build custom GPT chatbots from your own data, no coding needed
🔊 VoiceWave - Turn text into natural, emotional studio-quality speech fast
📝 Blogify - Auto-turn 40+ content types into SEO blogs in 150+ languages
🎨 ArtSpace - Create and upscale photorealistic images from text prompts
🚀 Typefully - Draft, schedule, and publish posts across social platforms
📚 Sourcely - Paste your text and find, summarize, and cite academic sources
🔍 Paperguide - AI research assistant that finds papers, analyzes, and cites
