惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

U
Unit 42
博客园 - Franky
T
Tailwind CSS Blog
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
月光博客
月光博客
人人都是产品经理
人人都是产品经理
雷峰网
雷峰网
Hugging Face - Blog
Hugging Face - Blog
有赞技术团队
有赞技术团队
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
阮一峰的网络日志
阮一峰的网络日志
C
Check Point Blog
爱范儿
爱范儿
T
The Blog of Author Tim Ferriss
aimingoo的专栏
aimingoo的专栏
Stack Overflow Blog
Stack Overflow Blog
博客园 - 聂微东
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
L
LangChain Blog
云风的 BLOG
云风的 BLOG
MyScale Blog
MyScale Blog
Microsoft Security Blog
Microsoft Security Blog
The Cloudflare Blog
博客园 - 三生石上(FineUI控件)

Hacker News - Newest: "AI"

AI can't read an investor deck AI as an attorney? Student uses ChatGPT, Gemini to sue UW over alleged racial discrimination Hacking MCP Servers in AI Systems – The Rug Pull: Tool Changes After Approval GitHub - MeepCastana/KubeezCut: Free Web based video editor Can AI judge journalism? A Thiel-backed startup says yes, even if it risks chilling whistleblowers Coming soon: 10 Things That Matter in AI Right Now DARPA built an AI to fact-check enemy weapons claims What explains heterogeneity in AI adoption? When AI Meets Muscle: Context-Aware Electrical Stimulation Promises a New Way to Guide Human Movements - Department of Computer Science AI Changed How We Build. It Did Not Change What Matters. Linux rules on using AI-generated code - Copilot is OK, but humans must take 'full responsibility for the… Meta spins up AI version of Mark Zuckerberg to engage with employees Code Mode: Let Your AI Write Programs, Not Just Call Tools | TanStack Blog GitHub - Delavalom/graft: Go framework for building AI agents. Type-safe tools, multi-provider (OpenAI, Anthropic, Gemini, Bedrock), zero vendor SDKs. India's TCS tops estimates, says new AI models did not dent services demand Gen Z's fading AI hype Strong feeling: we are in a folded AI reality GitHub - machinarii/total-recall-catalog: A reference catalog of latest knowledge retrieval, memory & RAG systems GitHub - mensfeld/code-on-incus: Give each AI agent its own isolated machine with root, Docker, and systemd. Active defense detects and stops threats automatically.. Quantization, LoRA, and the 8% Problem: Benchmarking Local LLMs for Production AI Iran war: We spoke to the man making Lego-style AI videos that experts say are powerful propaganda Powell, Bessent discussed Anthropic's Mythos AI cyber threat with major U.S. banks GitHub - immartian/bellamem: Persistent belief-graph memory for AI agents. Retrieves decisive context by importance — not recency, not RAG, not /compact. recursive-mode: The Repo-Native Operating System for AI Engineering After the attack on Sam Altman's home, will AI CEO's go on the offensive? The biggest advance in AI since the LLM Opus 4.6 vs GPT 5.4 One Prompt Unity World Generation Test “AI polls” are fake polls Client Challenge Can AI be a 'child of God'? Inside Anthropic's meeting with Christian leaders
Dust raises $40M Series B to scale multiplayer AI for hum...
doppp · 2026-05-21 · via Hacker News - Newest: "AI"

Three years ago, we started Dust with a core conviction: as AI models improve at an unprecedented pace, companies will need a new type of system that turns raw model capability into coordinated, compounding work across teams. 

Today, Dust is used by more than 3,000 organizations globally, from high-growth AI-native companies to established enterprises. Across those companies, the people closest to the work have deployed more than 300,000 agents. At Clay, Dust serves as foundational knowledge infrastructure for a rapidly growing GTM team. At Profound, it has become the source of truth for customer intelligence and post-sales. At Persona, teams across 11 departments have deployed more than 300 Dust agents. And at Doctolib, Dust is central to a company-wide AI strategy reaching 3,000 employees.

Across all of this, one thing has become clear: organizations don’t need standalone agents rolled out employee by employee. They need a shared system where humans and agents can work together in parallel, with the same context, tools, and aligned goals.

We call this multiplayer AI. Today, we’re announcing a $40M Series B with Abstract, Sequoia, Snowflake Ventures, and Datadog to accelerate building it.

The single-player AI bottleneck

Most companies today are stuck in single-player AI mode. Everyone gets access to their own agent, which can do work on their behalf, like researching a prospective account or creating a presentation, with a limited set of knowledge and tools. Each person might get some of their work done better or faster, but productivity stays trapped with individual people, and within their instance of a tool. The benefits don’t compound across the team. 

What this looks like in practice is the sales rep who saves time by using AI to research an account before the call, but the SE running technical discovery the next day starts from scratch. It’s the product marketer who uses AI to draft the one-pager, the content writer who writes the blog from their own version of the brief, and the sales enablement manager who builds the battlecard from an entirely different deck.

As people within organizations delegate more of their workload to AI, the bottleneck shifts from generation to coordination. The question is no longer simply whether AI can do the work; it’s whether teams can coordinate work across many people and many agents at once, with seamless reviews, approvals, handoffs, and shared visibility as work moves in parallel. Enabling that degree of human-agent collaboration is what unlocks compounding productivity gains across the entire organization.

Human-agent collaboration

Multiplayer AI can’t be built as a thin layer on top of a chatbot. The hard part is no longer prompting AI to produce an output; it’s designing the system that allows humans and agents to share context, use the right tools securely, and facilitate reviews and handoffs to keep work moving forward. That’s why we built Dust as a system for operational collaboration from the beginning, not just an interface for interacting with a model.

Shared workspaces

In Dust, that collaboration happens inside persistent, shared workspaces established around a team, initiative, or workflow. Instead of operating in isolated sessions, agents work alongside people with access to the same context, tools, skills, and work that’s already been done.

Context layer

Many tools can connect to other systems, but there’s a difference between connecting and understanding. Dust goes further: it helps agents make sense of information across sources, synthesize it accurately, and take action. It does this through a hybrid approach to enterprise knowledge: semantic search for deep ongoing contextual understanding and high-quality synthesis, plus MCP connections for simple queries and actions across tools.

Self-improvement

Most agents are static, meaning they perform the way they were initially configured until someone manually updates them. Dust creates a continuous improvement loop around organizational work where every interaction generates signal. Teams refine workflows over time, agents accumulate memory and context, and Dust identifies patterns across conversations to suggest improvements to agent instructions and skills.

Observability and governance

Admins get complete visibility and control over every agent interaction. That means granular permissions, cost and usage monitoring, a full audit trail, and agent analytics, all in one place. Dust is also

SOC 2 Type II

certified, GDPR compliant with EU/US data residency, and guarantees zero model training on your data.

Multiplayer AI in practice

If you think about the most complex work that happens at companies today, it’s not being done by just one person in isolation. Launching a product, onboarding a customer, conducting deal reviews, and running competitive teardowns can each involve dozens of humans and agents working in parallel sometimes over weeks and months — researching, synthesizing information, generating deliverables, coordinating follow-ups, and taking action from the same live context.

To coordinate this very announcement, Stan and I worked with our marketing, data, and operations teams together in Dust, running parallel workstreams with multiple people and agents contributing simultaneously. 

While humans made decisions, refined messaging, and coordinated approvals, agents synthesized data and updates from Slack, Snowflake, HubSpot, Notion, Gmail, and Google Drive. Our marketing team collaborated with agents to draft this blog post, prepare investor quote sheets, generate social copy, and build customer communication kits. Our data team's agents pulled the metrics you'll read below, live from our data warehouse. Our ops team coordinated the launch timeline, media embargo schedule, and approval workflows, with agents flagging blockers and @mentioning the right person when a decision was needed. 

Work no longer moved linearly between isolated teams, tools, meetings, and approval chains. Agents advanced work in parallel, while humans stepped in at the moments that required judgment: refining the message, approving next steps, resolving blockers, and making final decisions. Our human teammates and their agents worked from the same context, inside the same workflows, with shared visibility into decisions and progress. This is what multiplayer AI feels like in practice, and it’s how Dust operates every day.

AI Operators rewriting work

We’re a company that runs on what we’re building. Dust is the primary tool used by everyone on every team, from engineering to GTM. That’s why we’re certain that the people who transform organizations with AI won’t be external consultants, forward-deployed engineers, or centralized innovation teams. They’ll be the operators closest to the work: 

We call these people

AI Operators

. They use their functional expertise to rewire how their organizations work. Dust is built for them.

We believe the next generation of organizations will operate as integrated systems of humans and agents collaborating together in parallel. Multiplayer AI is the system they’ll run on. That system is Dust.