- AI Valley
- Posts
- Meta’s facial recognition Ray Ban glass
Meta’s facial recognition Ray Ban glass
PLUS: OpenAI unveiled 4 key updates at DevDay 2024
Howdy! It’s Barsee again.
Happy Thursday, AI family, and welcome back to AI Valley.
In today’s edition:
🔮 OpenAI's ‘DevDay 2024’ Big Updates
👀 Someone Put Facial Recognition Tech onto Meta's Smart Glasses
🧠 Microsoft's Copilot Can Now See, Speak, and Think Deeply
🤖 Plus Trending AI Tools, Guides, and Resources
Ready, set, go…
TOGETHER WITH WRITER
Writer RAG tool: build production-ready RAG apps in minutes
RAG in just a few lines of code? We’ve launched a predefined RAG tool on our developer platform, making it easy to bring your data into a Knowledge Graph and interact with it with AI. With a single API call, writer LLMs will intelligently call the RAG tool to chat with your data.
Integrated into Writer’s full-stack platform, it eliminates the need for complex vendor RAG setups, making it quick to build scalable, highly accurate AI workflows just by passing a graph ID of your data as a parameter to your RAG tool.
OPEN AI
🔮 OpenAI's ‘DevDay 2024’ Big Updates
OpenAI announced a slew of updates to its API services at a 'DevDay 2024' event on Wednesday in San Francisco.
What are the major updates?
Real-time API:
Enables direct audio interaction with AI models, allowing developers to create applications that process voice input and output without linking multiple systems.
It supports function calling (for tasks like ordering pizza or making appointments) and will eventually support multimodal experiences, including video. Here’s one example.
Vision Fine-Tuning:
Developers can now fine-tune GPT-4o with images and text, allowing for a stronger understanding of visual content and enhancing object detection, visual search, and more accurate image analysis.
Model distillation:
A technique that enhances the performance of smaller models (like GPT-4o mini) by allowing them to learn from larger models (such as GPT-o1).
Prompt Caching:
Allows developers to reuse prompts without paying full price each time.
Maintaining the same context across multiple API calls reduces costs and latency, previously seen in Anthropic’s and Google’s models.
SMART GLASS
👀 Someone Put Facial Recognition Tech onto Meta's Smart Glasses
Two Harvard students created I-XRAY, smart glasses that use facial recognition to identify faces and gather personal information like addresses and phone numbers—something big tech has avoided due to safety concerns.
How easily can our identities be compromised?
The aim is to highlight the potential dangers of such technology.
The demo video shows them testing the glasses on unsuspecting individuals in public, demonstrating how easily someone's identity and personal details can be accessed.
What if your face was your ID card?
Using Meta’s Ray-Ban smart glasses, the technology allows users to identify people by simply walking by them.
The system connects to facial recognition services, like Pimeyes, to retrieve and display information, including names and contact details, on a connected phone.
Why does it matter?
The glasses show how readily available personal information can be misused, prompting a conversation about data protection and privacy in our increasingly interconnected world.
MICROSOFT AI
🧠 Microsoft's Copilot Can Now See, Speak, and Think Deeply
Microsoft has just upgraded its Copilot personal AI assistant with exciting voice, vision, and reasoning features.
What are the upgrades?
Copilot Voice:
This upgraded voice assistant allows for multi-turn conversations and can be interrupted mid-sentence.
It’s designed to respond to your emotions and offers seamless communication with Copilot AI.
Users can choose from four unique voice options, all built on OpenAI models but specifically customized by Microsoft.
Copilot Vision:
Now integrated into Microsoft Edge, this feature enables Copilot to understand the content of web pages—whether text or video—and respond to your questions about them.
Think Deeper:
This new reasoning capability allows Copilot to tackle complex queries, providing detailed, step-by-step answers for math, logic, and coding tasks.
Where to access?
The revamped interface is now live on web, iOS, Android, and Windows, and users can also access Copilot through WhatsApp instead of Meta's AI assistant.
QUICK HITS
OpenAI raises $6.6B in the largest VC round ever, reaching a post-money valuation of $157B. Microsoft, NVIDIA, and SoftBank were among the participating investors. (link)
AI companies are opting you in by default. (link)
OpenAI also asked investors to avoid backing five rival AI startups, including Anthropic and Elon Musk's xAI. (link)
8 reasons why WhatsApp was able to support 50 billion messages a day with only 32 engineers. (link)
Men stole over $1 Million from DoorDash delivery drivers by impersonating them to customer service. (link)
Google is working on AI with human-like reasoning capabilities. (link)
NVIDIA released NVLM 1.0, a powerful open-source multimodal AI model to rival OpenAI's GPT-4. (link)
VCs are eager to meet with the former OpenAI CTO Mira Murati, expecting her to launch a new company soon. (link)
USEFUL AI LINKS
Trending Tools
Flow > a Mac dictation app that writes 3x faster than typing, perfect for AI prompts, with auto-edits and 100+ languages. (link) *
Ledger Up > AI bookkeeper for startups. (link)
Video SDK 3.0 > Build and integrate real-time multimodal AI characters. (link)
Semblian 2.0 > Outsource your time-consuming tasks to AI. (link)
Buzzabout > Get real-time audience insights from billions of discussions on social media. (link)
Pika 1.5 > An image-to-video tool, now features Pika Effects, allowing you to explode, melt, crush, inflate, and "cake-ify" anything in your videos. (link)
Resources / Guides
How to use Cursor AI. From one of the makers of Cursor. (link)
DAILY DOSE OF CONTENTS
1/ A drone coded and controlled with OpenAI's o1 live on stage.
A drone programmed from scratch on stage at OpenAI demo day
— Nick Dobos (@NickADobos)
5:27 PM • Oct 1, 2024
2/ Generative AI is truly amazing: Creating 10 podcast episodes from scratch and publishing them on Spotify in 2 hours.
Over the last ~2 hours I curated a new Podcast of 10 episodes called "Histories of Mysteries". Find it up on Spotify here:
open.spotify.com/show/3K4LRyMCP…10 episodes of this season are:
Ep 1: The Lost City of Atlantis
Ep 2: Baghdad battery
Ep 3: The Roanoke Colony
Ep 4: The Antikythera… x.com/i/web/status/1…— Andrej Karpathy (@karpathy)
9:40 PM • Oct 2, 2024
THAT’S ALL FOR TODAY
That’s all for today’s issue, folks.

💡 Help me get better and suggest new ideas at [email protected] or @heyBarsee
👍️ Like what you see? Subscribe here
Thanks for being here.
HOW WAS TODAY'S NEWSLETTER |
REACH 100K+ READERS
Acquire new customers and drive revenue by partnering with us
Sponsor AI Valley and reach over 100,000+ entrepreneurs, founders, software engineers, investors, etc.
If you’re interested in sponsoring us, email [email protected] with the subject “AI Valley Ads”.