AI Newsletter

Microsoft and Nvidia are breaking Intel/AMD's dominance in the PC market by launching Windows PCs with Nvidia chips as primary processors and equipping them with locally running AI agents. This enables more performant, privacy-friendly AI applications directly on the end device and could reshape the entire PC architecture and software ecosystem. The move addresses the growing AI PC market and could redistribute significant market share from the established x86 duopoly.

Cybercriminals are abusing the public chat-sharing features of ChatGPT and Claude as vehicles for malware distribution, as these are hosted on trusted domains and thus bypass security tools. This jeopardizes trust in AI platforms, forces OpenAI and Anthropic to implement faster security measures, and could amplify regulatory demands for stricter content controls.

OpenAI is expanding its Codex app with "Computer Use" – AI agents can now autonomously control, test, and debug Windows programs. The feature significantly reduces manual developer and QA work and strengthens OpenAI's position in the enterprise market against competitors like Anthropic and Microsoft.

Salesforce has deployed Claude Code company-wide for its developers and reports extreme productivity increases (79% more pull requests, 5% fewer bugs). This signals a strategic shift to Anthropic as the preferred AI coding partner and could build competitive pressure on Microsoft/GitHub Copilot – though without independent validation of the figures.

OpenAI is optimizing its product portfolio through a GPT-5.5 Instant improvement and simplification of chat functionality, while phasing out older models. This signals an aggressive product pace and focus on user value while reducing complexity. The move targets higher conversion and API adoption through improved performance and unified developer environment.