PLUS: Anthropic's COBOL killer tanks IBM stock and ChatGPT solves a major physics puzzle
Happy reading
OpenAI is making a major push into the enterprise, teaming up with global consulting giants like McKinsey and Accenture to accelerate the adoption of its new ‘Frontier’ AI agent platform.
The move is a clear strategy to embed autonomous agents into the core workflows of large businesses. But does this direct challenge to established SaaS vendors signal a major disruption for the entire enterprise software market?
In today’s Next in AI:
OpenAI's new consulting army
Anthropic's COBOL killer tanks IBM
ChatGPT solves a major physics puzzle
The US launches an AI ‘Tech Corps’
OpenAI's Consulting Army
Next in AI: OpenAI is partnering with global consulting giants like McKinsey, BCG, and Accenture to accelerate enterprise adoption of its new 'Frontier' AI agent platform. The partnerships, detailed in an official announcement, signal a major push to integrate AI agents into large-scale business operations.
Explained:
Frontier is a platform that lets businesses build, deploy, and govern AI agents that can operate across their entire technology stack, including CRM systems, HR platforms, and internal tools.
The new Frontier Alliances position BCG and McKinsey as strategy partners for leadership, while Accenture and Capgemini will handle the technical, end-to-end systems integration.
This move directly challenges established SaaS vendors like Salesforce and Microsoft and intensifies OpenAI's competition with its main rival, Anthropic.
Why It Matters: This strategy makes it easier for large enterprises to move beyond AI experiments and embed autonomous agents into their core workflows. It also positions OpenAI's platform as a potential replacement for traditional enterprise software, threatening to disrupt the multi-billion dollar SaaS market.
Anthropic's COBOL Killer
Next in AI: Anthropic unveiled a new AI tool that translates the decades-old COBOL programming language, sending IBM's stock tumbling as investors weighed the threat to its lucrative mainframe business.
Explained:
COBOL is a language from the 1950s that still powers immense parts of our economy, including an estimated 95% of all ATM transactions in the U.S.
Anthropic’s tool automates the painstaking analysis of legacy code, mapping dependencies and documenting workflows that would otherwise take analysts months to complete.
The announcement immediately shook the market, causing IBM shares to plunge over 13% and spooking investors in other tech consulting firms.
Why It Matters: This development signals a major shift where AI tools are targeting the costly, complex work of modernizing legacy systems that run global infrastructure. The technology's value now extends beyond creating new applications to making it economically viable to upgrade critical, aging codebases.
ChatGPT's Physics Discovery
Next in AI: In a stunning human-AI collaboration, physicists used OpenAI's ChatGPT to solve a decades-old puzzle in particle physics, proving a long-doubted particle interaction is theoretically possible.
Explained:
The challenge involved an incredibly complex formula for gluon interactions that had stumped researchers for over a year. After months of manual calculations hit a wall, ChatGPT-5.2 Pro was prompted to simplify the expression, reducing dozens of terms to a single elegant line in minutes.
The team then used an internal OpenAI model, nicknamed "SuperChat," which took the AI-generalized formula and generated a formal proof in just 12 hours. This two-step process—generalizing with one model and proving with another—showcases a powerful new workflow for scientific research.
While the underlying physics ideas weren't new, the physics community was shocked that a machine could perform such a task. This marks a potential new paradigm, where AI acts as a powerful assistant to find hidden patterns and accelerate research in fundamental science.
Why It Matters: This breakthrough signals AI's evolution from a data processor to a genuine partner in core scientific discovery. It provides a powerful blueprint for tackling other intractable problems in fields like quantum gravity, materials science, and drug discovery.
The AI 'Tech Corps'
Next in AI: The White House is launching the 'Tech Corps,' a new initiative modeled after the Peace Corps to deploy American tech talent abroad. The program aims to promote U.S. AI solutions and counter China's growing global influence.
Explained:
The program will recruit and train STEM professionals to provide on-the-ground support for implementing American AI solutions that address grassroots problems in agriculture, health, and education.
This initiative is a key part of the broader American AI Exports Program, a strategy designed to maintain U.S. tech dominance as countries like India are expected to participate in U.S.-led tech alliances.
Volunteers will serve for 12 to 27 months with benefits including housing and a stipend, and the Peace Corps is now accepting applications for deployments expected to begin in fall 2026.
Why It Matters: This initiative marks a strategic shift toward using AI as a tool for diplomatic soft power. By exporting technology along with talent, the U.S. aims to build lasting technological alliances and shape global AI standards.
AI Pulse
Anthropic published its AI Fluency Index, revealing that while users often iterate with AI, they are less likely to question a model's reasoning when it produces polished outputs like code or documents.
Opper ran a simple "car wash" reasoning test through its LLM gateway, finding that only 5 out of 53 top AI models could consistently answer that you should drive your car to the car wash.
Ladybird ported its JavaScript engine to Rust with assistance from AI tools, successfully translating 25,000 lines of C++ code in just two weeks with zero regressions.
xAI delayed a new Grok model release after engineers were reportedly redirected to improve the chatbot's knowledge of the video game Baldur's Gate.