Videos

Strata local AI runs on a single 12GB 3060 and works surprisingly well, averaging about 49 tokens/sec decode once...

The best LLMs only follow complex enterprise policy instructions 36.2% of the time under strict grading, according...

Howie Liu's central claim is that the next AI product form factor is the 'claw': an always-on agent that wakes...

Langfuse's workshop demonstrates a practical workflow for continuously improving AI agents by combining online...

Arize AI is pushing toward fully autonomous self-healing agents where observability data is consumed...

Strands Agents, an open-source agent harness from AWS, lets you build and deploy custom agents with any LLM, and...

The presenter builds a complete meeting notes system in Notion that uses AI to automate prep, transcription, action...

GitHub Copilot's agent system is designed around three tiers: local agents in the CLI/IDE, shared agents distributed...

Training a 4B Qwen model with RL on expert-validated synthetic financial QA data beat its 235B variant (60% vs 51%...

Airbyte's agent-facing CLI and MCP interfaces reveal that designing for agents requires fundamentally different...

Factory's software factory concept proposes that engineers shift from writing code to governing AI agents that...

Holo4 is a model trained to choose between clicking, coding, and calling tools, unifying three agentic lanes that...

Agent sandboxes should contain zero credentials to limit the blast radius of autonomous actions, according to Docker...

The most significant claim here is that stale and duplicate source documents quietly poison RAG: DZ Labs says that...

Most teams ship AI features on vibes—testing a few queries manually and shipping—but this breaks as...

Strand Agents, an open-source framework from AWS, demonstrates agents that write, fix, and extend their own tools...

MCP's main problem—context bloat from loading all tools upfront—is a flaw in agent harness design, not the protocol...

OpenAI's computer use is now faster than average humans at most tasks, with the next frontier being superhuman...

The Chief AI Officer role is still a moving target, varying by company type and AI maturity, but at WSO2 it breaks...

For builders, OpenAI DevDay's real news is pricing: GPT-6.1 Soul is claimed to deliver near-Astra quality for...

GEPA is a reflective optimization technique that achieves large performance gains from very few samples by using...

Weights & Biases' Arya agent uses its own production traces and offline evaluations to generate new tasks and prompt...

Building agents on document intelligence requires decomposing the parsing problem: use traditional computer vision...

Open source AI coding agents with local models can handle routine changes like adding a CSV export, but the lack of...

AI underwriting is emerging as a new risk layer for AI adoption, with AIU raising $40 million to build insurance...

Reinforcement learning can make search agents roughly 20 times faster and about 100 times cheaper than frontier...

Pinecone's Nexus is a persistent, specialized knowledge layer for enterprise agents, designed to solve the problem...

Grokbot can be wired to its own email inbox via Agent Mail and webhooks, letting it wake up and act on incoming...

Agent orchestration code is being replaced by files: markdown instructions, YAML configurations, and bash skills are...

Vercel's shift from a multi-agent architecture to a single file-system-based agent, inspired by Claude Code and Opus...

Restate is an open-source durable execution framework that treats agents as persistent, stateful entities rather...

Mike Chambers argues that production agents need a 'harness'—the scaffolding around the model that handles memory,...

OpenAI's internal data shows agent runtime surpassed human work in June and hit a 3.14:1 ratio by mid-August, but...

The talk argues that agents have a measurement problem because token spend is not clearly tied to outcomes, making...

tldraw's spatial harness for agents turns the infinite canvas into a medium where LLMs can understand 2D space, set...

Replicating the IFScale benchmark shows the old 200-instruction ceiling is obsolete: today's frontier models hold...

Recursive Language Models (RLMs) treat context as an object in a symbolic environment (like a Python REPL) rather...

Two Sigma runs all cloud agents as the user's own identity, not a separate machine identity, which solves permission...

Anthropic's new Certified Claude Architect exam is a proctored, 60-question, 2-hour test that assesses judgment in...

Strands Agents is an open-source framework from AWS that adds an LLM-powered agent layer to robots, enabling natural...

Building reliable multi-agent systems before mastering a single agentic loop is a recipe for failure, mirroring the...

Existing agent frameworks that require editing prompts in code are counterproductive for building reliable...

Building agents has become trivial with modern frameworks like Flu and Cloudflare, but the real challenge is...

Niels Rogge automated his job of convincing researchers to upload model weights to Hugging Face by building a...

The new MCP Tasks Extension replaces the experimental 1120 spec with a simpler, stateless design that removes...

The Model Context Protocol (MCP) has gone stateless, removing the need for sticky sessions and session stores when...

This week four Chinese labs shipped frontier open-weight models, and Hugging Face's report shows the open-source...

Deno CEO Ryan Dahl argues that AI agents need a hard, external security boundary rather than relying on the agent...

Anthropic's own 186-page report reveals that its biological safety filters were off for 11 months, affecting 50,000...

Boris Cherny, creator of Claude Code, lays out a four-step ladder from basic AI chat to full AI-native operations...