ChatGPT heads to Washington

PLUS: A Chinese open-source music generation model arrives

Good morning, AI enthusiasts. ChatGPT just got its federal security clearance — and the U.S. government may be getting its most productive civil servant yet.

With OpenAI’s new Gov platform deploying across thousands of agencies, bureaucracy's slow-moving wheels are about to get an efficiency upgrade for the AI age.

Speaking of ChatGPT, our next workshop is this Friday at 4 PM EST. In it, you’ll learn how to automate tasks with OpenAI’s Operator! RSVP here.

In today’s AI rundown:

  • OpenAI unveils ChatGPT Gov platform

  • Block releases open-source AI agent platform 'Goose'

  • Create personalized illustrations for your brand

  • YuE unlocks open-source AI music generation

  • 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

OPENAI

Image source: Ideogram / The Rundown

The Rundown: OpenAI just launched ChatGPT Gov, a specialized version of its flagship product designed specifically for U.S. agencies — allowing departments to securely take advantage of AI systems in their own private environment.

The details:

  • The platform allows agencies to deploy ChatGPT within Azure environments, enabling sensitive data processing and adherence to security protocols.

  • Gov users can access 4o, and Enterprise features like conversation sharing, custom GPTs, and admin controls for department-wide deployments.

  • The government has already actively used ChatGPT, with data showing that over 90k employees across 3,500 agencies have generated 18M messages since 2024.

Why it matters: With AI models becoming intertwined in nearly every sector, it's no surprise that the government is adopting the tech internally and needs a more secure deployment for sensitive data and materials. This new, tailored option could significantly spur even broader AI adoption across federal agencies.

TOGETHER WITH RUNNER H

The Rundown: 2025 is the year of AI Agents, and Runner H’s state-of-the-art agent delivers on the promise — transforming text inputs into real-world actions across your organization.

With Runner H, you can easily use natural language to:

  • Execute intricate tasks like marketing campaigns and code reviews with a single prompt

  • Automate workflows for instant, real-world business operations

  • Leverage AI across every aspect of your business

Join the waitlist now to be among the first to test Runner H.

BLOCK

Image source: Block

The Rundown: Twitter founder Jack Dorsey’s fintech company Block just launched Goose, an open-source AI agent framework that allows developers to build and deploy AI assistants across multiple platforms and systems.

The details:

  • The platform supports any LLM backend, including OpenAI, DeepSeek, and Anthropic, while maintaining data privacy and deployment control.

  • Goose integrates with Anthropic's MCP and APIs to enable a wide range of tool connections, with the ability to add new integrations mid-session.

  • Early implementations at Block show Goose automating complex tasks like code migrations, dependency management, and test generation.

  • The framework was released under the Apache 2.0 license, allowing unrestricted commercial and research use.

Why it matters: The open-source movement is accelerating across AI, and Jack Dorsey’s latest bird-themed release is now bringing it to the agentic layer. Given Block's track record with Square and Cash App, Goose could do what those apps did for AI agents for payments — taking complex tech and making it accessible to everyone.

AI TRAINING

The Rundown: Freepik lets you create your own branded illustration style by uploading reference images and getting consistent visuals across all your content.

Step-by-step:

  1. Access Freepik AI's Create tab and click the plus icon in the Style section.

  2. Upload your reference images (10-50 images recommended).

  3. Choose your quality setting (Ultra, High, or Medium).

  4. Generate illustrations by selecting your style and writing a prompt.

Pro tip: Check out our exclusive workshop on creating custom styles with Freepik here.

Additionally, Rundown University members enjoy a free monthly premium Freepik plan.

PRESENTED BY CONVEYOR

The Rundown: Conveyor just launched Sue, the first AI Agent for Customer Trust — capable of taking on all the tasks that infosec teams dread like sharing NDA-gated documents, answering security questionnaires and everything in between. Perfect for B2B infosec teams sick of manual work.

Sue can do all of this for you:

  • Manage and complete customer security requests while communicating across systems and teams

  • Complete, reject, or escalate questionnaires based on parameters you set

  • Handle any questionnaire format with no manual tagging — even portals

  • Personalize your Trust Center using insights from conversation intelligence tools

AI RESEARCH

Image source: HKUST

The Rundown: Hong Kong University of Science and Technology researchers just introduced YuE, an open-source AI music generation that transforms lyrics into complete songs, challenging closed commercial platforms like Suno and Udio.

The details:

  • YuE leverages two specialized models — one for vocals and music and another for production elements — enabling the creation of songs up to 5 minutes in length.

  • The system supports multiple languages and can handle complex vocal techniques like scatting and mixed-voice performances.

  • Users can fine-tune their creations with controls for genre, instruments, mood, and vocal characteristics while still maintaining a coherent musical structure.

Why it matters: With commercial AI music platforms like Suno and Udio facing growing legal challenges from record labels, an open-source system could seriously shake up the space. Whether from YuE or another entrant, broader access to AI music tools is coming fast – and the music industry is already reckoning with the changes.

QUICK HITS

  • 🎥 Pika 2.1 - New AI video model with enhanced realism, motion, and control

  • 🐋 Janus-Pro - DeepSeek’s new open-source AI image generation model

  • 🤖 Qwen2.5-VL - New family of vision-language models with agentic capabilities to interact with computers and phones

  • 📝 BulletPen - Transform spoken thoughts and rambles into polished writing

Alibaba’s Qwen introduced Qwen2.5-Max, a new large-scale MoE model that outperforms DeepSeek-V3, GPT-4o, and Claude 3.5 Sonnet across key benchmarks.

OpenAI CEO Sam Altman chimed in on DeepSeek’s R1 release, calling it an impressive model and saying it’s ‘invigorating’ to have a new competitor.

Figure AI launched a new Center for Advancement of Humanoid Safety to establish industry-wide testing standards and quarterly safety reports for workplace robots.

Ex-OpenAI safety researcher Steven Adler posted on X that he is ‘pretty terrified’ by the pace of AI development, saying no lab currently has a solution for alignment.

Convergence AI launched Proxy, a natural language agent it calls ‘Europe’s answer to OpenAI’s Operator.’

Hugging Face integrated four new serverless inference providers — fal, Replicate, Sambanova, and Together AI — enabling direct model deployment and faster inference.

COMMUNITY

Join our next workshop this Friday at 4 PM EST to learn about how to automate tasks with OpenAI’s Operator and create your own free personalized, offline assistant with Dr. Alvaro Cintas, The Rundown’s AI professor.

RSVP here. Not a member? Join The Rundown University on a 14-day free trial.

We’ll always keep this newsletter 100% free. To support our work, consider sharing The Rundown with your friends, and we’ll send you more free goodies.

That's it for today!

Before you go we’d love to know what you thought of today's newsletter to help us improve The Rundown experience for you.

Login or Subscribe to participate in polls.

See you soon,

Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team

Reply

or to participate.