Cohere

Cohere's Aidan Gomez on Why Enterprise AI Beats AGI Theater — and What's Actually Hard in 2025

Aidan Gomez, co-author of the seminal "Attention Is All You Need" paper and CEO of Cohere, argues that enterprise-focused AI deployment is more impactful than AGI-chasing, and that the transformer architecture's dominance persists not because alternatives are absent, but because the ecosystem lock-in (custom silicon, tooling, infrastructure) makes switching costs prohibitively high. Reasoning/test-time compute, long anticipated within the research community, has delivered outsized intelligence gains at surprisingly low cost relative to pre-training. Cohere's strategic differentiation centers on private, on-premise/VPC deployment and multilingual enterprise models — a wedge that enables access to sensitive data that API-based competitors cannot touch.

Cohere Partners with SAP to Bring Agentic AI to Enterprises

Cohere and SAP have partnered to integrate Cohere's enterprise-grade AI models into SAP Business Suite and make them available through SAP AI Core. This collaboration aims to accelerate AI adoption at scale for SAP customers, offering advanced agentic AI capabilities for various business functions. The focus is on providing secure, multilingual, and domain-optimized AI solutions for regulated industries.

Cohere’s Command A Model: High Performance, Low Compute for Enterprise AI

Cohere has released Command A, a generative AI model optimized for demanding enterprise tasks. The model demonstrates performance comparable to or exceeding larger competitors like GPT-4o and DeepSeek-V3, particularly in agentic, multilingual, and RAG scenarios. A key advantage is its efficiency, requiring significantly less computational resources for deployment, making it an attractive option for private and on-premise solutions.

Cohere Co-founder Ivan Zhang on the Origins of Forai and the Future of AI Research

Ivan Zhang, co-founder of Cohere, discusses the origins of Forai, a community-led AI research initiative started in 2017. He highlights the importance of curiosity, persistence, and interdisciplinary collaboration in the early days of AI. Zhang also emphasizes the need for better evaluation metrics (evals) in AI research and the crucial role of community-led initiatives in shaping AI policy, especially for developing nations.

Cohere’s Enterprise AI Strategy: Balancing Innovation and Practical Application

Cohere and its co-founder, Aiden Gomez, focus on enabling enterprises to adopt AI language models to enhance productivity and transform services. They prioritize a hybrid approach, combining generalist tools with custom-built, domain-specific solutions. Cohere emphasizes the importance of robust models, strong customer support, reliability, and security, especially for sensitive data and regulated industries.

Aidan Gomez: What No One Understands About Foundation Models | E1191

Navigating AI/ML Careers: Essential Advice from Tech Talent Leaders

Securing a role in AI/ML requires a strategic approach focused on foundational skills, continuous learning, and impactful project showcases. Candidates should prioritize networking and engagement within relevant communities, carefully tailoring their applications to demonstrate business impact and technical proficiency. Recruiters spend a limited time on resumes, emphasizing the need for clear, concise, and skill-centric presentations.

The Transformer Architecture: Simplicity, Scale, and the Future of AI

The Transformer architecture, introduced by the paper "Attention Is All You Need," is foundational for large language models. Its success is attributed to its simplicity, scalability for massive compute, and its ability to handle sequence data through attention mechanisms. The future of AI models may involve state-space models to overcome Transformer limitations and a shift towards data-centric approaches and human feedback for model improvement.

Cohere: Democratizing Large Language Models and Advancing AI Research

Aiden Gomez, co-founder of Cohere, highlights the transformative impact of large language models (LLMs) and Cohere's role in making this technology accessible. The discussion covers the architecture and training of LLMs, their rapid adoption, and future applications including enhanced personal assistants and browser control. Cohere also supports open AI research through its non-profit arm, Cohere For AI, aiming to foster broader participation in the field.

Making Large Language Models Accessible | Scale AI Fireside chat with Bill MacCartney (Cohere)

Rethinking ML Research: Efficiency, Openness, and Societal Impact

Sammy Bengio, a veteran in deep learning research, advocates for a shift in machine learning research towards efficiency and understanding "why" models work, rather than solely focusing on scaling up. He emphasizes the importance of open-source contributions, diverse teams, and cultivating research environments that prioritize exploration and risk-taking. Bengio also cautions against the disconnect between academic research and real-world societal needs, advocating for researchers to stay grounded and address contemporary challenges.

Cohere AI: Democratizing Large Language Models through API Access

Cohere AI aims to democratize access to large language models (LLMs) by providing them as an API service, abstracting away the immense computational and development costs for businesses. This "power plant" model allows companies to leverage state-of-the-art NLP capabilities without building and maintaining their own massive Transformer neural networks. They offer services for generating, classifying, and embedding text, targeting a broad range of developers from novices to experts.