Welcome, AI Innovators!
The skills gap is now a product. OpenAI will certify 10M Americans inside ChatGPT and route them into its own jobs marketplace. DeepSeek readies a self-improving agent, and DeepMind puts on-device embeddings, so your files stay private while search gets smarter. Decide what you build with, what you hire for, and what you keep off the cloud.
📌 In today’s Generative AI Newsletter:
OpenAI jobs marketplace and certifications
DeepSeek self-improving agent in Q4
DeepMind offline AI, 100+ languages
💼 OpenAI Turns from Job Killer to Job Broker with New Hiring Platform
Image: OpenAI
OpenAI’s Fidji Simo, CEO of Applications, announced plans for an OpenAI Jobs Platform to connect businesses with AI-skilled workers, paired with a certification program that teaches “AI fluency” directly inside ChatGPT.
Inside the initiative:
Jobs marketplace: A platform to match employers with AI-trained candidates, including tracks for small businesses and local governments.
Certification push: Walmart and other employers helped design tiered programs for measuring AI skills, building on the existing OpenAI Academy.
National target: The company aims to certify 10M Americans by 2030, making AI literacy a baseline credential across industries.
Policy alignment: The rollout parallels White House AI education efforts, with tech leaders meeting in Washington this week on workforce development.
OpenAI is presenting itself as both the architect of disruption and the builder of the safety net. It also creates a collision course with LinkedIn, Microsoft’s job platform, sharpening tensions between the two uneasy partners.
🐳 DeepSeek Prepares ‘Self-Improving’ AI Agent for Q4 Launch
Illustration: Allie Carl/Axios
Chinese startup DeepSeek is developing a new AI agent with the ability to execute multi-step tasks autonomously and refine itself over time, according to Bloomberg. The release is targeted for the fourth quarter of this year, marking the company’s most ambitious step since the R1 model that jolted the AI landscape last year.
What we know so far:
Autonomous workflows: The system is designed to handle complex tasks with minimal human input, adapting based on prior performance.
Timeline: Founder Liang Wenfeng says the agent is on track for year-end release, even as the awaited R1 successor faces delays.
Industry context: The project follows a wave of agentic products, from OpenAI’s ChatGPT Agent to Anthropic’s Claude for Chrome.
Competitive landscape: While rivals Alibaba and Tencent push rapid release cycles, DeepSeek has kept quiet since R1’s breakout moment.
Agentic AI has yet to prove it can reliably live up to the hype. If DeepSeek delivers on the promise of self-improvement at scale, it could trigger another shockwave across the sector and potentially change what it means to compete in the “year of the AI agent.”
📂 DeepMind’s New Model Handles Files Offline
Image source: Google
Google DeepMind has introduced EmbeddingGemma, an open-source model designed to run directly on consumer devices. It supports over 100 languages, works in real time, and consumes less memory than a photo app, making it light enough for smartphones and laptops.
Key features at a glance:
On-device privacy: Handles personal files, messages, and emails without sending data to the cloud.
Flexible precision: Developers can tune the model for either higher accuracy or faster speeds depending on the use case.
Broad integration: Already compatible with popular dev tools and runs directly inside web browsers.
Real-world efficiency: Capable of powering offline search and text understanding in everyday applications.
DeepMind is positioning EmbeddingGemma as infrastructure for the next wave of AI agents and assistants, where privacy-first, offline performance will matter as much as speed or scale. With Apple’s Siri ambitions looming, Google’s move keeps it in the spotlight for the on-device AI race.
🚀 Boost your business with us. Advertise where 13M+ AI leaders engage!
🌟 Sign up for the first AI Hub in the world.
