CogLake Blog
Knowledge infrastructure, without the hand-waving.
Technical and operational writing about evidence, permissions, retrieval quality, organizational memory, and agent-ready systems.
Security & governanceAnthropic opens Singapore office: MOU on AI safety and enterprise adoption
Anthropic has opened an office in Singapore — its fifth in the Asia-Pacific region — and signed an MOU with the Ministry of Digital Development and Information covering AI safety, model evaluation and responsible adoption. What this means for enterprises.
Read articleSecurity & governanceOn-device scam detection: DeepMind and Singapore fine-tune Gemma
Google DeepMind and GovTech Singapore have fine-tuned a lightweight Gemma model that flags conversational scams directly on the phone — without chat content leaving the device. What is documented, how the two-stage pipeline works, and what enterprises can take away.
Read articleKnowledge infrastructureGemini 4 Argon: Google makes its new frontier model globally available
Google DeepMind has released Gemini 4 Argon and made it globally available via API: introductory pricing of $2 per million input and $10 per million output tokens. What is documented, what remains a vendor claim — and how enterprises should check availability.
Read articleSecurity & governanceAnthropic Cyber Mission: AI support for cyber defenders
Anthropic launched the Cyber Mission on October 8, 2026: the Critical Infrastructure Defense Program with eleven founding partners for OT security, plus OSS Scanner, a free opt-in scanning service for open-source projects. What this means for enterprises, fully sourced. As of October 13, 2026.
Read articleKnowledge infrastructureGPT-6 for everyone: OpenAI rolls out Sol and Luna
OpenAI switched on GPT-6 for everyone in ChatGPT on October 7, 2026, free and paid; Sol and Luna replace GPT-5.6. What the new system card says about safety, prompt injection and benchmarks — and what enterprises should check now. As of October 8, 2026.
Read articleKnowledge infrastructureClaude Haiku 5.5: Anthropic cuts prices for high-volume AI work
Anthropic has released Claude Haiku 5.5, a small model for high-volume, cost-sensitive work that it says costs about 75% less than Haiku 4.5 on average. A sourced look at pricing, benchmarks and availability for enterprises. As of October 8, 2026.
Read article
Retrieval engineeringEmbeddingGemma 2: Multimodal Embeddings for Local Enterprise Search
Google DeepMind released EmbeddingGemma 2: an open 740M-parameter model that maps text, code, images, video, and audio into one vector space. What it means for local enterprise search and RAG — with sources for every figure.
Read articleKnowledge infrastructureWhy enterprise AI needs evidence, not only answers
A useful answer is only the beginning. Enterprise teams also need provenance, permissions, and a path back to the original file.
Read articleRetrieval engineeringBeyond basic RAG: building a governed evidence engine
Exact terms, semantic meaning, graph relationships, and corrective retrieval solve different failure modes. The hard part is governing their combination.
Read article
Knowledge continuityWhat happens to company knowledge when people leave?
The most valuable knowledge is often not a document. It is the connection between a decision, a customer, a result, and the person who understood why.
Read article