Legacy Code

Cognition, an AI coding startup, is reportedly in discussions to raise capital at a $40 billion valuation—a significant jump from its $26 billion valuation just months ago during its previous $1 billion funding round. The acceleration reflects investor confidence in the company's technology and market traction as enterprises increasingly adopt AI-assisted software development.

Guitar company D'Addario has finally acknowledged using AI-generated music (created with Suno) in a recent promotional video after initially denying the allegations and allowing speculation to mount for nearly two weeks. The admission came after mounting evidence and public scrutiny, raising questions about the company's transparency and authenticity commitments to its audience.

Google has announced the Pixel Watch 5, integrating deeper AI capabilities (powered by Gemini) into health and fitness monitoring on the wearable. The new model maintains the $399 price point while enhancing processing power and adding new Gemini-powered health insights, predictive analytics, and personalized coaching. The watch continues to refine Google's ecosystem approach to AI-powered wearables.

OpenAI has begun testing advertisements within ChatGPT as a way to sustain free access to the platform while generating revenue. The company is implementing clear labeling for sponsored content, ensuring that ads do not influence the accuracy or independence of ChatGPT's responses. Users will have control over ad preferences, and OpenAI is committing to strong privacy protections so that ad targeting won't rely on personal chat data.

OpenAI and Amazon Web Services have made Daybreak, a specialized AI model for cybersecurity, available through Amazon Bedrock. Daybreak is designed to help security teams detect vulnerabilities, analyze threats, and respond to incidents more efficiently. By integrating with Bedrock, enterprise customers can now access Daybreak's capabilities alongside other AWS services without requiring separate infrastructure or licensing agreements.

OpenAI has published research examining how enterprises are deploying agentic AI in real workflows, moving beyond pilot projects to production systems. The research shows that leading companies are using ChatGPT and code-execution capabilities to automate routine business processes, with notable patterns around which use cases deliver measurable ROI. The findings reveal significant gaps between early adopters and mainstream enterprises in implementation maturity and results capture.

MIT Technology Review has published analysis on the critical bottleneck limiting AI agent adoption in enterprises: data quality and trustworthiness. As organizations scale AI agents from proof-of-concept to production, the quality of data flowing into agents becomes the primary determinant of ROI. The research identifies specific frameworks for ensuring agents have access to reliable, current, and contextually appropriate information.

Thrive Holdings, a platform backed by OpenAI for bringing generative AI to enterprise workflows, has raised $2 billion at a $12 billion valuation. The funding round included major institutional investors including SoftBank, D1 Capital Partners, and Altimeter Capital. Thrive is positioning itself as an implementation and integration layer, making it easier for large organizations to adopt and operationalize AI across departments.

Twitch has introduced an opt-out mechanism allowing creators to prevent their streams, VODs, clips, chats, and channel content from being used to train Amazon's generative AI models. The move comes as platforms face increasing scrutiny over use of creator content for model training without explicit consent or compensation. Streamers can toggle the setting in their channel preferences to block future training data collection.