Presented by
Good morning, AI enthusiasts. Remember when smart glasses were just a fashion statement? Meta just changed that. The company launched prescription-ready AI glasses today with EssilorLuxottica, and they ship with a brand-new AI model from day one.
EssilorLuxottica isn't a brand collab. They make lenses for most of the world's prescription glasses. At $299, Meta Glasses aren't for early adopters anymore. They're for everyone.
In today's recap:
Meta Glasses go prescription-ready with EssilorLuxottica
Five Eyes warns AI cyber threats are months away
Extract and structure documents at scale with Mistral OCR 4
4 new AI tools, prompts, and more
META
Meta launches AI glasses made for everyone
Meta
Recaply: Meta just unveiled Meta Glasses with EssilorLuxottica. It's the first AI glasses line that works with prescription lenses. The line has 26 styles starting at $299 and ships with Meta AI powered by Muse Spark from day one.
Key details:
Meta Glasses use a dedicated action button to invoke Meta AI. Open-ear speakers and a multi-mic array handle calls and voice control. Users can take photos and video hands-free, with prescription lenses supported.
The line has 26 styles across three frames: Adventurer, Fury, and a Kylie Jenner collab. Battery runs 8+ hours, with 40 more hours from a foldable case.
These are the first glasses to ship with Muse Spark, Meta's debut model from Superintelligence Labs. It was previously only available on Ray-Ban Meta and Oakley Meta.
Meta Glasses are on sale today at Meta.com, Best Buy, Amazon, LensCrafters, and Sunglass Hut. Pedestrian navigation and live translation for 14 new languages are coming soon.
Why it matters: The Ray-Ban Meta line showed people will wear AI glasses. But it was always a fashion play. EssilorLuxottica changes that. They make lenses for most of the world's prescription glasses. That gives Meta Glasses a path to people who need corrective lenses and have never tried smart glasses before. At $299, it's priced like mid-range earbuds. That's not an early adopter price.
PRESENTED BY NEO
AI help, without the trust tax.
Most AI tools ask you to trade your data for intelligence. Norton Neo doesn't. It's the first safe AI-native browser built by Norton, and it gives you powerful built-in AI without handing your privacy over to get it. Search, summarize, and write with AI built directly into your browser. Your data stays yours. Your context stays private.
Built-in VPN, anti-fingerprinting, and ad blocking come standard. No add-ons. No setup. No compromises.
Fast. Safe. Intelligent. That's Neo.
AI RESEARCH
Five Eyes agencies warn AI threats are months away
Recaply: The Five Eyes agencies just issued a rare joint warning. Frontier AI capable of devastating cyberattacks on governments and businesses is months away, not years.
Key details:
The agencies say frontier AI will give attackers new tools to strike faster and at greater scale. Legacy systems running old, unsupported software are most at risk.
The warning comes from all five western intelligence agencies: the US, UK, Australia, Canada, and New Zealand. It's one of the most coordinated public AI threat statements ever issued.
The warning follows the Trump administration's move to suspend Anthropic's Fable 5 and Mythos 5. Senate testimony cited by The Economist claims Mythos breached nearly all NSA and US Cyber Command systems in hours.
The agencies said the timeline for AI surpassing current offensive capabilities is not years away. It is mere months. They called for immediate action on patching, access controls, and modernizing legacy systems.
Why it matters: Five intelligence agencies issuing a joint public warning is rare. Doing it about AI is unprecedented. The context matters. Days before this warning, Anthropic's two most capable models were shut down. Reports said one of them could breach classified government systems. Whether those reports fully hold up, the intelligence community's posture just shifted publicly.
GUIDES
Extract and structure documents at scale with Mistral OCR 4
Recaply: In this tutorial, you will learn how to run batch document extraction with Mistral OCR 4, reducing your cost to $2 per 1,000 pages while processing PDFs, Word files, and presentations in parallel.
Step-by-step:
Get an API key at Mistral Studio (mistral.ai), install the SDK with
pip install mistralai, and set up your client withMistral(api_key="YOUR_KEY"). All document formats including PDF, DOC, PPT, and OpenDocument are supported.Prepare a JSONL batch file with each line containing a
custom_idand abodywith your document as a base64-encoded image URL. For PDFs, encode each page as a JPEG before adding to the file.Upload the batch file with
client.files.upload(file={"file_name": "batch.jsonl", "content": open("batch.jsonl", "rb")}, purpose="batch")to get afile_idback from the API.Start the job with
client.batch.jobs.create(input_files=[file_id], model="mistral-ocr-latest", endpoint="/v1/ocr"), then pollclient.batch.jobs.get(job_id=job.id)until status returnsSUCCESS.Download the output JSONL, parse each line, and access
.pages[*].markdownfor structured text and.pages[*].bounding_boxesfor block-level location data to feed downstream pipelines.
Pro tip: Add a JSON schema to the same API call via Document AI parameters to have OCR 4 reshape extracted content into a structured format on the first pass, with no second LLM call needed for invoice processing, form parsing, or compliance extraction.
TOGETHER WITH WISPR FLOW
10x the context. Half the time.
Speak your prompts into ChatGPT or Claude and get detailed, paste-ready input that actually gives you useful output. Wispr Flow captures what you'd cut when typing. Free on Mac, Windows, and iPhone.
TOOLS
Trending AI Tools
🤖 Claude Tag - Anthropic's Slack-native AI team member that takes @mention tasks and works async across channels
📄 Mistral OCR 4 - Document intelligence model with bounding boxes, block classification, and 170-language support at $4 per 1,000 pages
🎨 Krea 2 Raw - Open-weight 12B image generation base model for fine-tuning and LoRA training with commercial license
⚡ Krea 2 Turbo - Krea's 8-step distilled checkpoint for fast, polished text-to-image generation
NEWS
What Matters in AI Right Now?
Anthropic just launched Claude Tag, a Slack-native AI team member that joins channels, takes @mention tasks, and works asynchronously on code, data analysis, and incident resolution.
Mistral just released OCR 4, a document intelligence model with bounding boxes, block classification, and inline confidence scores across 170 languages, priced at $4 per 1,000 pages with a 50% Batch API discount to $2.
NVIDIA just launched BioNeMo Agent Toolkit, giving AI agents callable life sciences tools spanning biology, chemistry, genomics, and drug discovery, with more than 50 organizations including Anthropic and OpenAI already integrating it.
Alibaba just sued the US Department of Defense over its placement on the Pentagon's Chinese military company blacklist, arguing the designation lacks evidence and violates its constitutional due process rights.
The White House just signed two executive orders targeting a commercially relevant quantum computer by 2028 and accelerating government post-quantum cryptography readiness to 2031, down from the prior 2035 target.
The NSA just lost access to Anthropic's most capable AI model after the Trump administration's export control order suspended Fable 5 and Mythos 5, cutting off an agency that had been evaluating the model's ability to detect software vulnerabilities.
🧡 Enjoyed this issue?
🤝 Recommend our newsletter or leave a feedback.
How'd you like today's newsletter?
Cheers, Jason
