The Hallway Track
Every signal tracked from the world's AI conferences — 4,665 talks, quotes & announcements, newest first.
Choosing AMD EPYC-powered Azure infrastructure can improve PostgreSQL performance by over 60%.
“the 'Ea' series are often most recommended because the extra memory gives you more room to scale and PostgreSQL absolutely benefits from having more memory per vCPU”
Microsoft trains 200+ field engineers on Postgres using modular, persona-mapped content rather than one-size-fits-all curriculum.
“Postgres is the database that never changes phone number.”
PostgreSQL and SQL Server differ fundamentally in their security models, especially around roles, ownership, and privileges.
“Every login, every user, every group that we know in SQL Server, everything is a role in Postgres.”
NVIDIA outlines building a transaction foundation model to replace brittle hand-engineered features in financial data.
“Transaction data is one of the richest signals an enterprise owns.”
AWS launches container image caching for SageMaker AI inference, cutting scale-out latency up to 2x for generative AI models.
“This speeds up end-to-end latency by up to 2x for generative AI models during scale-out events.”
Continental is applying agentic AI on AWS to transform global tire supply chain planning across a million-plus network intersections.
“So, looking at that simple question, it becomes a complex problem to answer that. It is requiring a huge manual effort.”
Databricks introduces governed vibe coding to bring context and control to enterprise coding agents.
“Vibe coding needs context and control to work in the enterprise”
AWS invented and open-sourced P-EAGLE, parallelizing speculative decoding for up to 1.69x faster LLM inference on SageMaker.
“P-EAGLE instead fills positions 2–4 with learnable placeholders and predicts all four tokens at once”
Databricks introduces OpenSharing, an evolution of Delta Sharing built for the agentic AI era.
Databricks launches Apps on its Marketplace, letting organizations distribute and discover data intelligence applications.
Databricks introduces OpenSharing SecureConnect for sharing live data across organizations.
NVIDIA launched the ACE Game Agent SDK and Unreal Engine 5 plugins for building on-device AI companions.
Microsoft demos LLM-planned robot control via Semantic Kernel skills written in Python, with novel robotics skills not yet public.
“It's a while loop for the robot.”
Nadella says true ambition with AI means building new conceptual models for 'meta cognition' work, not automating old tasks.
“True ambition is about making the impossible possible.”
Georgi Gerganov confirms Qwen3.6-27B is a capable local model for daily coding tasks on consumer hardware.
“I can 100% attest to the fact that Qwen3.6-27B is a very capable local model for coding tasks.”
Optimizing transformers for low-precision training cuts GPU hours and speeds up experimentation and model scaling.
“Accelerating transformers is therefore not just a performance optimization, but directly affects how quickly teams can experiment and how large a model they can afford to train.”
General Robotics exposes a unified quadruped robot API and AI inferencing via an MCP server so Copilot agents can program robots in natural language.
“I very infrequently have to write code myself. I'm using, you know, models from Cloud Model, from OpenAI through Copilot to kind of help me write all this code.”
Anthropic's new interpretability research uses round-trip translation to reliably decode Claude's internal activations into text.
“It is what is missing from the formula.”
NVIDIA Blackwell delivered a clean sweep in MLPerf Training v6.0, leading on scale and per-accelerator performance.
“NVIDIA delivered a clean sweep in MLPerf Training v6.0, the latest edition of industry-standard AI training benchmarks developed by the MLCommons consortium.”
Sciene built an autonomous Customer Success platform on Databricks to scale relationship-centered work.
Jensen Huang says AI is not mysterious—it is understood software, and fears of existential doom are nonsense.
“You may or may not lose a job to an AI. But you will absolutely lose a job to someone who uses AI.”
2026 frontier post-training has shifted to Multi-Teacher On-Policy Distillation (MOPD), merging many specialist models into one.
“The shape of a post-training recipe has changed more in the last year than in the prior three.”
Databricks expands Agent Bricks one year after launch at Data + AI Summit 2026.
Databricks announces Unity Catalog updates at Data + AI Summit 2026 to govern enterprise AI agents.
“The agentic era is here. Hundreds of thousands of agents are now acting on enterprise”