PLUS: GM reboots its workforce for AI and the coming end of Python's dominance
Happy reading
OpenAI is making a massive $4 billion push to get its AI models out of the lab and into the core of business operations. The company is launching a new venture that will embed specialized engineers directly into client organizations to speed up adoption.
The move marks a significant strategic pivot, shifting focus from pure research to hands-on enterprise integration. Does this new, high-touch deployment model signal a future where AI's value is unlocked not just by powerful models, but by expert human implementation?
In today’s Next in AI:
OpenAI's $4B enterprise deployment company
Thinking Machines' real-time interaction models
GM's major AI workforce reboot
The approaching end of Python's dominance
OpenAI's $4B Enterprise Play
Next in AI: OpenAI is launching the OpenAI Deployment Company, a new company backed by over $4 billion to embed specialized engineers directly into businesses. The goal is to help organizations move AI projects from concept to core operations.
Explained:
The new venture launches with over $4 billion in capital from a partnership with 19 leading global investment firms, including TPG, Advent, and Bain Capital.
To hit the ground running, OpenAI acquired Tomoro, an applied AI firm that will bring approximately 150 experienced deployment specialists to the team from day one.
The core strategy involves embedding Forward Deployed Engineers inside businesses to connect OpenAI models with company-specific data and redesign critical workflows.
Why It Matters: This signals a major strategic shift for OpenAI from a research lab to an enterprise integration powerhouse focused on real-world application. For businesses, this offers a direct line to expert implementation, significantly reducing the friction in deploying complex AI solutions at scale.
Beyond the Chatbot
Next in AI: Thinking Machines has released a research preview of 'interaction models,' a new AI architecture designed for continuous, real-time collaboration. This approach moves beyond today's turn-based systems, allowing the AI to process audio, video, and text simultaneously for a more fluid partnership.
Explained:
Instead of bolting on interactivity, this design builds it directly into the model's core. The creators argue this follows the bitter lesson of AI development, ensuring that a model's collaborative ability scales directly with its intelligence.
This native interactivity unlocks new capabilities that feel more natural. The model can make verbal interjections, handle simultaneous speech for tasks like live translation, and run tools or searches in the background without pausing the conversation.
The system uses a multi-stream, micro-turn architecture that grounds the interaction in time. When a task requires deep reasoning, the interaction model delegates it to a background model while remaining present and responsive to the user.
Why It Matters: This represents a fundamental shift from prompting a tool to collaborating with a partner. If widely adopted, this technology could create AI interfaces that are far more intuitive and effective for complex, dynamic work.
GM's AI Workforce Reboot
Next in AI: General Motors is replacing hundreds of its IT workers with new talent skilled in AI-native development, showcasing a major shift in how legacy companies are restructuring their workforce for an AI-first world.
Explained:
GM laid off over 600 salaried IT employees in a deliberate skills swap, clearing out traditional roles to make way for a new wave of AI experts as part of a larger strategic overhaul.
The company is actively hiring for roles centered on AI-native development, data engineering, prompt engineering, and agent/model development, making AI fluency a new test for its tech workforce.
This move follows a series of leadership changes under Chief Product Officer Sterling Anderson, including hiring AI leads from Apple and Cruise, signaling a top-down commitment to embedding deep AI capabilities.
Why It Matters: This signals how large enterprises are moving beyond simply adopting AI tools to fundamentally rebuilding teams around AI expertise. For tech professionals, it's a clear indicator that foundational AI skills are quickly becoming essential for career growth and relevance.
The Post-Python Era
Next in AI: AI agents can now write code in high-performance languages like Rust and Go so proficiently that the long-held advantages of developer-friendly languages like Python are eroding, signaling a major shift in software development.
Explained:
Microsoft is leading the charge, rewriting its TypeScript compiler in Go. The newly released TypeScript 7.0 beta is roughly 10x faster than the previous version, proving the massive performance gains available when AI assists with difficult porting tasks.
AI agents are tackling exceptionally complex projects. In a landmark experiment, Anthropic used 16 parallel Claude agents to build a C compiler in Rust from scratch, successfully booting Linux and compiling major open-source projects.
The infrastructure layer is rapidly consolidating around performance. AI labs see this as critical, with OpenAI recently acquiring Astral, whose Rust-based tools save its Codex model millions of compute minutes weekly.
Why It Matters: The steep learning curve for systems languages, once a major barrier, is quickly disappearing. This shift enables developers to prioritize runtime performance and efficiency from the start, rather than treating it as an afterthought.
AI Pulse
Cerebras increased its IPO price range, now seeking a valuation of up to $48.8 billion as the AI chipmaker prepares to go public this week.
Google announced its new Gemini-powered AI Health Coach, a $9.99/month subscription service launching May 19 that provides personalized fitness and wellness plans and rebrands the Fitbit app to Google Health.
Anthropic claimed its new Claude Mythos model is too powerful for public release due to its advanced ability to find cybersecurity vulnerabilities, limiting access to select partners under its Project Glasswing initiative.
A new study found significant gender bias in the perception of AI use, with reviewers labeling an AI-generated résumé for a woman as “weak” while giving an identical one for a man a 97% approval rating.