Baseten vs Dataiku

Goodfire Launches Internal AI Agent Monitors

Goodfire, a startup specializing in AI interpretability, launched on Thursday a new type of AI agent monitor that inspects a model's internal signals rather than reading its output, aiming to detect rogue behaviors more efficiently and at a fraction of the cost. The monitors are available to customers of Baseten, an AI model hosting platform. Baseten's Base Labs had previously announced a safety partnership with Goodfire and Hugging Face. Goodfire's approach uses small probes that scan a model's internal activations at each step, triggering a closer AI review only when flagged. In tests on the Kimi K3 model, Goodfire's monitors caught 94% of malicious hacking sessions and cost about $51 for 1,500 sessions, compared to $233 for a cheaper model and $10,000 for a top-tier one. The company positions the solution for open models, which can be stripped of safeguards, and sees it as critical for inference-time guardrails.

• Goodfire launched internal activation monitors for AI agents on October 8, 2026.
• Monitors are available to Baseten customers.
• In tests on Kimi K3, monitors caught 94% of malicious hacking sessions.

Announcing our partnership with OpenAI

Baseten announces a partnership with OpenAI, as featured on their blog.

Baseten, Hugging Face, Goodfire Partner for AI Safety

Baseten's research arm, Base Labs, announced a partnership with Hugging Face and Goodfire AI to establish a safety standard for open-weight AI models. The collaboration aims to develop and publish methods for training and monitoring these models, making safety an integrated feature rather than an afterthought. This comes amid rising concern over 'abliteration,' a technique that removes safeguards from open models, with Hugging Face listing over 6,000 such models. The companies have not detailed technical specifics, but Goodfire's expertise in AI interpretability is expected to play a key role. This initiative seeks to leverage openness as an advantage for safety, providing transparent controls and encouraging community contributions.

• Baseten launched a safety infrastructure standard via Base Labs.
• Partnership with Hugging Face and Goodfire AI announced.
• Hugging Face hosts over 6,000 abliterated models.

Discover Dataiku Agent Management: every agent on the record

Dataiku announced the launch of Dataiku Agent Management, a new product feature for managing AI agents, along with the announcement of the AI Success Frontrunner award winners.

12,000 Attend Ai4 2026 to Discuss AI's Invisibility

The article reports on the Ai4 2026 conference in Las Vegas, which saw record attendance of over 12,000, up from 8,000 the previous year. Dataiku's keynote highlighted that 96% of enterprise leaders believe employees are using unsanctioned generative AI tools, and 80% of CIOs see their jobs at risk without measurable AI ROI. Pat Gelsinger argued that AI economics must improve dramatically, while Geoffrey Hinton, Fei-Fei Li, and Andrew Ng jointly endorsed AI regulation. The conference emphasized the need for AI governance, budget ownership, and cross-functional staffing. Speakers from Cisco, Nvidia, Uber, PayPal, and other companies discussed the shift from assistants to autonomous agents and the importance of accountable AI deployment.

• Ai4 2026 attendance exceeded 12,000, up from 8,000 the previous year.
• Dataiku's survey found 96% of enterprise leaders believe employees use unsanctioned generative AI tools.
• 80% of CIOs say their job is at risk without measurable AI ROI.

Governed RAG: Data, Context & Lineage for Enterprise AI

The article describes risks introduced by Retrieval-Augmented Generation (RAG) when enterprise data is exposed to vector search pipelines and proposes a three-part Governed RAG architecture: (1) ingestion with cryptographic embedding lineage and metadata, (2) query-time contextual Attribute-Based Access Control (ABAC) embedded into vector search queries, and (3) outbound payload sanitization (PII/PHI masking, indirect injection removal, and context length minimization). It argues that enterprises must enforce retrieval-time access controls, maintain graph-based data lineage, and implement real-time index freshness/eviction to prevent privilege escalation, prompt-injection attacks, stale-context hallucinations, and to meet compliance requirements.

• Retrieval-Augmented Generation (RAG) pairs LLMs with vector databases and knowledge graphs to ground agents in proprietary corporate knowledge.
• Vector stores typically do not preserve fine-grained document-level ACLs or cryptographic data lineage by default, creating over-permissioned retrieval risks.
• The proposed Governed RAG pipeline has three security boundaries: ingestion with cryptographic embedding lineage, query-time contextual ABAC inside the vector search, and outbound payload sanitization.