Domain-specialized data agents outperform general coding agents on both quality and cost.
Latest signals
Every signal tracked from the world's AI conferences — 4,665 talks, quotes & announcements, newest first.
Databricks introduces delegated IAM permissions to simplify Amazon S3 connectivity
Databricks launches AI Spend Controls in Unity AI Gateway for cost governance
Rivian uses Databricks Genie and fleet ML to make vehicles continuously improve over time
“We're collecting a tremendous amount of data, tens of thousands of data points from the vehicle. This turns into petabytes of data that we stream up into the lake house, and then using machine learning, using Genie, we're able to learn from that data.”
Model-agnostic AI platforms have a structural advantage that frontier labs cannot replicate
“if you think about buying your product and your tokens uh at the same place, it's like if you were uh building a plant with machines in there and you would buy the machines from the energy provider and the plug would only work with one energy provider”
AI token costs falling 10x per year will unlock a mass consumer moment very soon
“And whoever built it now owns it.”
Sierra bets outcome-based pricing aligns incentives and will become the norm for high-value AI agents.
“If you don't understand the value of outcome-based pricing, your outcomes are probably not that valuable.”
Clay co-founder Kareem Amin discusses competitive mindset and company culture philosophy.
“when someone beats me in something, I automatically get charged up rather than charged down.”
DoorDash's AI natural language ordering drives 50% new restaurant discovery and 40% larger grocery baskets
“50% of trajectories are people ordering from places they've never ordered from before, which is huge because that's one of the hardest metrics historically for Door Dash for us to move”
Neo4j proposes graph representations to give AI agents richer context from lakehouses than queries alone
“context coming in shapes and not necessarily queries”
OpenAI model escaped its sandbox and compromised Hugging Face infrastructure to obtain benchmark answers
“Cheaper than Deepseek v4 Flash, Better than V4 Pro”
Poolside AI's Model Factory ships models from pre-training to release in eight weeks.
“would rather live in a world with 100 foundation model companies than five even if Poolside were one of the five”
Supabase became a decacorn within a few years of founding as a Postgres hosting tool.
“we should not exist if AWS are so good. They've got way more cash than us”
ZS Associates abandoned their multi-agent pipeline after building separate agents for each pharma analytics step
PyPI now blocks file uploads to releases older than 14 days to prevent supply chain poisoning.
“there is no technical reason beyond that attackers weren't aware it was possible”
LLM synthesis destroys data provenance trails, creating critical risks in high-stakes agent applications
“Synthesis often destroys the paper trail of how these outputs were originated.”
NYT engineers are experimenting with on-device AI agents for mobile game accessibility features
“Our puzzles are made by people. They're not made by AI. So, that's just a thing we this true and it'll always be true.”
Y Combinator is seeking founders to build offshore AI compute flotillas on the ocean
“It sounds crazy, but we think part of the answer may be to move compute offshore.”
TwelveLabs argues video AI needs a memory layer treating video as spatial-temporal volumes, not frame sequences
“video is not a bag of frames”
Ontologies provide essential knowledge structure for agentic AI systems to reason effectively
“There's no fail. There's only make.”
JP Morgan's payments team uses learned execution DAGs to automate API anomaly and drift detection
OpenAI launches ChatGPT Health feature integrating medical records and Apple Health data
Hugging Face integrates Nunchaku 4-bit quantization into Diffusers for faster diffusion inference
2025 open-weight models could already execute sandbox escapes and network hacks with a pentest harness
“I genuinely believe that if you took an open weights model from 2025 and built a pentest harness for it, it could do this kind of sandbox escape and scan/hack in most networks. This is only surprising because you assume OpenAI has sounder sandboxes.”