OpenAI and Broadcom unveiled Jalapeño, a custom AI chip built for LLM inference.
Latest signals
Every signal tracked from the world's AI conferences — 4,665 talks, quotes & announcements, newest first.
Databricks claims the largest data and AI conference in the world with 32k attendees as CEO touts AI co-workers fed full enterprise context.
“We're going to have more software generated in the next year than in the history of humanity.”
Hugging Face launches the FFASR Leaderboard to benchmark automatic speech recognition on real-world audio.
Databricks unveils LTAP, combining Lake Base (managed Postgres) and a transactional/analytical engine for real-time agentic data workflows.
“So Lake Base is open source Postgres fully managed by Databricks with separated storage and compute and it's really easy to get started.”
Databricks introduces LTAP architecture to unify OLTP and analytics without fragile CDC pipelines or failed HTAP systems.
“CDC doesn't stand for change data capture. It really stands for continuous data corruption.”
Databricks unveils Lakehouse//RT and Reyden to enable millisecond-latency serving workloads without copying data out of the Lakehouse.
“the warehouses or the Lakehouses hit a wall at about a second”
Mastercard standardized on Databricks Lakebase to power its agentic platform consolidating 80+ services into one architecture.
“as one agent takes action or generates insights, it becomes instantly available for the next one”
Databricks pitches Lakeflow as a unified platform to simplify fragmented data engineering for AI and agents.
“All these apps, agents, AI, everything needs data, and you might be thinking, "My job got a little bit harder."”
Databricks launches Genie One, an AI coworker that connects to enterprise data and apps to deliver accurate insights and automate actions.
“Genie 1 is an AI co-worker that can connect to all of your data and all of your apps so that everyone in your organization can get insights and even automate action.”
Databricks ships Iceberg v3 unified data layer so Delta and Iceberg tables share data files without rewrites.
“No one should have to care about formats.”
LangChain launched Engine, an agent that autonomously investigates traces and drafts PRs to improve other agents.
“We're working towards a future where agents improve themselves.”
Apple releases updated Figma and Sketch design kits for iOS, iPadOS, and macOS 27 with Liquid Glass updates.
PepsiCo consolidated 60+ data lakes into one Databricks lakehouse, with 90% of its data now on the platform powering AI.
“for us, what we truly believe is an AI transformation is a data transformation”
Databricks promotes Genie, an AI coworker positioned to understand your business data.
LangChain shows how to convert an existing LangGraph agent into a voice agent using the Pipecat framework.
“Pipecat is going to handle all of the glue required to take the input audio, convert it to text, run it through our LangGraph LLM layer, then convert that back to speech in order to send it back to the end user.”
Simon Willison built a browser playground testing whether OPFS can let Pyodide edit persistent local SQLite files.
“I've been pondering if Datasette Lite - the Python Datasette application run entirely in the browser using Pyodide and WebAssembly - might be able to edit persistent SQLite files stored on the user's computer.”
Databricks and NVIDIA launch Genesis Workbench, a blueprint for GPU-accelerated AI in life sciences.
GPT-5 Pro helped immunologist Derya Unutmaz solve a 3-year-old T cell immunology mystery.
Midjourney launched Midjourney Medical to build a cheap, fast ultrasonic CT imaging device using its image-reconstruction expertise.
“because Midjourney is bootstrapped and profitable, they can afford to focus on something no other frontier AI lab would touch with a 10-ft pole, and that is actually improving the human experience”
NVIDIA frames agentic AI and open biology models as moving scientific discovery from human speed to computational speed.
“Discovery no longer waits on us.”
AWS shows how to build a conversational protein research copilot using Amazon Bedrock AgentCore and Strands Agents SDK.
Power can reach 40% of AI factory operating expenses, making performance per watt a critical efficiency metric.
“Power can account for 40% of the operating expenses (OpEx) to run an AI factory.”
Meta engineered 7mm-wide steel-can batteries to power all-day AI features in smart glasses.
“Smart glasses need a battery that can claim every micron of space – something rigid, precise, and shaped to the product rather than the other way around.”
SQL Server 2025 adds native vector search enabling semantic, meaning-based queries beyond keyword matching.
“But vector search is completely different uh in the sense uh that it really is a search for meaning.”