1231 summarised stories about AI Agents & Workflows, each linking back to the original source. Browse all topics →
SiliconANGLE 1 month ago 51 ● 4 sources
Salesforce introduced AI agents to automate sales and technical support workflows across its cloud services. Hunter, the first sales agent, includes a long-horizon runtime that only it has at launch. Salesforce will roll out the rest of the features by year’s end and expand long-horizon runtime and multi-agent orchestration capabilities across more agents via Agentforce Coworker updates.
Simon Willison’s Weblog 1 month ago 27
OpenRouter routes requests to different backend providers behind a single API endpoint, which can make the same model call behave differently depending on the provider’s serving software and settings. One concrete issue raised is that some providers may not support vision for vision models. Users may need to use provider.only and the /endpoints list of available providers per model ID to keep behavior consistent.
MarkTechPost 1 month ago 33
HarnessDev tested whether LLMs can write and iteratively improve the agent “harness” (the runnable code with tools and execution loop) rather than just produce answers. Only 34 out of 64 harness changes moved held-out scores in the same direction as feedback, with execution gains described as small and noisy. The work concludes harness quality is executor-specific and that much generated state/memory code often never runs, so generalization across executors and revisions is limited.
MarkTechPost 1 month ago 36
Anthropic added a plugin evals workflow to Claude Code that runs a plugin against prompts, grades the output, and compares it to a no-plugin baseline. It runs twice per case by default, so the key metric is Δ and a $0.41 example took 74 seconds across 6 runs. Results now include per-grader verdicts and can be enforced in CI with a threshold and max-cost limit, revealing cases where skills don’t trigger on natural phrasing.
SiliconANGLE 1 month ago 48
Five9 is positioning its Humantic contact-center model around humans and AI agents working together instead of aiming for full automation. By year’s end, PODS is expected to route more than 100,000 calls through Five9’s Voice AI Agents, and Five9 reports a 44% containment rate that exceeded its target. The approach shifts deployments toward keeping complex, high-value, and vulnerable interactions with people while using AI for high-volume, low-stakes requests and emphasizing an open, orchestrated platform.
Amazon Web Services 1 month ago 43
Amazon Bedrock AgentCore now supports hosting Model Context Protocol apps that render interactive HTML widgets inside AI hosts like ChatGPT and Claude, demonstrated with a Unicorn Rentals MCP app.
SiliconANGLE 1 month ago 40 ● 2 sources
Contact centers are shifting their success metrics from containment—calls that avoid a human—to whether AI-assisted interactions actually resolve the customer’s request end to end. Zoom’s AI for Zoom CX team, led by product head Ram Rajagopalan, has spent the past year building this “conversation to completion” approach into its virtual agent platform. Enterprises are expected to pick smaller first use cases and ensure full call context carries over during virtual-to-human handoffs to achieve measurable resolution instead of just deflection.
Ben's Bites 1 month ago 12
Ben’s Bites built an AI-assisted “design words” style-picker tool that generates copy prompts for an agent by selecting layout styles and components. He generated 11 prototype versions and ended up with v8 as the front runner. After shifting the UI for faster visual comparison, he replaced a broken shuffle/preview flow with a better mechanism, then published the site via a Cloudflare Worker proxy while tracking 116 prompts over 2 days of work.
Product Hunt 1 month ago 14
Juggler launched as a desktop workbench that organizes AI coding agent tool calls and approvals in Finder-style columns, with conversations branching into sub-threads. It is free and open-source and ships as one Go binary. It changes how developers can operate and manage agent workflows by using JavaScript plugins and a structured thread view.
TLDR Dev 1 month ago 42 ● 3 sources
OpenAI introduced an Agents API that packages Codex’s long-running harness into a public-beta interface for managing context, tools, files, recovery, and subagent orchestration. It supports asynchronous sessions that can run for hours or days. As a result, teams can deploy these agent sessions in an OpenAI-hosted sandbox, on their own infrastructure, or through partner environments.
TheSequence 1 month ago 8
The Sequence Opinion’s Issue 931 argues that robotics is waiting for a ChatGPT-like moment where people can teach robots through conversation with minimal setup. Issue 931 is number 931. It says conversational models make new tasks easier to explore, but real-world progress still hinges on learning, control, and the economics of deploying robots in new environments.
Ben Hylak 1 month ago 31 ● 33 sources
OpenAI’s GPT-6 is presented as adding “Computer Use” as its next capability for AI agents. GPT-6 is tied to computer use that builds on an earlier shift during October 2025 to February 2026 when OpenAI “missed” a more unhobbling style of tools. As a result, the article argues agents can handle tasks that require using real computers, shrinking the need for gap-fillers and speeding workflows like software development and personal assistance that lack APIs.
X 1 month ago 27 ● 3 sources
OpenAI paused new $200 Pro signups so existing users can keep access to Astra under heavy demand. The pause applies to the $200 Pro tier. Other plans and the API will remain open.
Product Hunt 1 month ago 1
Weave Engineering Intelligence launched Weave Router 2.0 to route coding agent requests across Claude and GPT models based on what’s available. On Terminal-Bench 4.0 and SWE-Atlas, it matched GPT-6 Astra at half the cost. It now uses a task-complexity and quota-aware classifier to switch models only when estimated savings exceed the rebuild cost.
Trending Topics 1 month ago 5
Chris Schnabl, a 25-year-old Austrian investing partner at Andreessen Horowitz, says he invests in early-stage AI application startups based on his technical background and his path from robotics competitions to Cambridge and UC Berkeley research stints. He focused on seed-to-Series A AI applications, including voice and agentic systems. As a result, his firm’s early-stage AI bets increasingly emphasize custom models built by application teams inside enterprise workflows while shifting from token-maxing toward cost and continual learning during use.
SiliconANGLE 1 month ago 41
Extreme Networks launched Agent ONE Coworker to replace network troubleshooting dashboards with an AI agent that provides answers in Platform ONE. In the first eight days, it recorded about a 900% increase in interactions. The release uses a context layer, “nudge” notifications, and honesty guardrails to cut resolution time from hours to minutes while preparing the later, more autonomous Operator mode.
Product Hunt 1 month ago 41
izzit launched as a mobile game that gives one word a day and uses a vision model to guess what you made. It lets you try for free, with $0.99 to unlock it permanently. Players can earn Gold (first try) or Diamond (extra-creative attempt) and keep the Mystery card if the model can’t identify it.
SiliconANGLE 1 month ago 2 ● 4 sources
Salesforce introduced Enterprise AI Harness and an AI Control Plane to help customers build and manage AI agents. Enterprise AI Harness includes six feature collections, and its rollout is set for early in Salesforce’s 2028 fiscal year starting next February. The tools add reliability, data-quality, cybersecurity controls, intelligent model routing, and centralized monitoring of agent performance and inference costs.
Together AI 1 month ago 46
Together AI expanded Together Fine-Tuning to add support for more open-weight models plus live experiment tracking and additional training and data controls. The release also drops LoRA training prices by 30% to 70% on selected models. It adds features like metrics at every step, dataset tokenization previews, per-example weighting, early stopping, adjustable batch sizing, and finer control over sequence packing.
The daily briefing
TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.
TLDRocket
24 October 2026
Local deployment took the spotlight today, with NVIDIA making small enough models feel practical again. Its Nemotron 3.5 Bolt E4B targets a 4GB VRAM sweet spot for on-device voice agents, game NPCs, and robot interfaces, packing 8B parameters on paper (4B effective) and a 64K context window. The twist isn’t just the size; it’s the packaging. NVIDIA is offering an open-weight 4GB-environment build in Q4_K_M, alongside API and integration paths aimed at tool-calling agent behavior—meaning developers can stitch together speech, actions, and domain tools without needing a data center’s worth of compute.
Read the full briefing →