We're only year 3 of a decade (if not multi-decades) long transformation of work. 3 years ago we bet on building an horizontal platform for work with agents, a chance to invent a new operating system for companies, from scratch, with AI as a fundamental premise. Many people considered us crazy for going after that, praising verticalized AI products as the winning strategy. But here's the thing: the time horizon of tasks successfully handled by agents has been predictively increasing form minutes to hours and will in all likelihood reach the equivalent of days and weeks of human work equivalent in the coming quarters. This is were verticalized and/or single-player AI falls short. Single-player tools, one person, one agent, confined to your machine is the wrong architecture for what's coming. We're shifting from using AI to produce things, to managing fleets of agents that do the producing. 3 years ago I wrote[1]: "ChatGPT is the Pong of LLMs. [...] Imagine, one day we'll get the DOOM, Civ, Red Alert, and Counter Strike of LLMs. Let alone multiplayer modes." Weeks long tasks in companies are inherently collaborative and mechanically spanning multiple teams. The new bottleneck in harnessing agents within organizations is coordination: multiple humans and multiple agents need to work together, with shared context, shared tools, shared goals. Agents that can hand work off to other agents or surface decisions to the right person at the right time. Humans who can review, steer, and step in without losing the thread. Teams that can run parallel workstreams and actually stay aligned. This is Multiplayer AI, and that's what we've been building at Dust. Across Datadog, Clay, Persona, 1Password, Doctolib and 3,000 organizations globally, we've watched teams figure out what this looks like in practice. 300,000 agents deployed. 70% weekly active. 240% NRR. Today we're announcing a $40M Series B with Abstract, Sequoia, Snowflake, and Datadog to accelerate our vision. Designing the right interfaces for multiplayer AI is the next frontier. Join us to redefine work by defining multiplayer AI.

Most AI at work is still single-player. The gains stay trapped with each individual. Nothing compounds across the team. Today, we're announcing @DustHQ $40M Series B to scale multiplayer AI: humans and agents working in parallel, with shared context, shared tools, and shared goals. 🧵

What should designers produce to help teams decide faster? We've been running an experiment at Dust. Field report by our very own @duncid: dust.tt/blog/field-study-pro…

Engineering Night: London Edition 🇬🇧 Mondays just got better. We're bringing Engineering Night to the UK on Feb 23, and yes, it's on a Monday. 👀 We may be biased, but spending your Monday night with cool people, learning how to ship better agents, and eating pizza sounds like a pretty great way to start the week. Victoria Klein from Agentica and our very own Jules Belveze are sharing the real, technical details of shipping AI agents. Not what might work someday. What's working now. Then the important stuff: Q&A, pizza, drinks, talking shop with people who actually build this stuff. PSA: Our events fill up fast, so if you're interested, register now! 📍 London 📅 Monday, Feb 23 | 6pm GMT Link 👉 luma.com/w5keu427

🔧 Your agent fleet just got an upgrade. Today we’re launching Skills. Picture this: 1. Your sales ops specialist builds the perfect HubSpot workflow 2. Three agents use it 3. Twenty-seven agents don't even know it exists 4. Expertise doesn't spread, agents stay siloed Skills fixes this: - Package instructions, knowledge, and tools once - Deploy across every agent that needs it - Update once, improve everywhere Your sales workflow becomes a skill. Your data query patterns become a skill. Your compliance rules become a skill. One sales ops expert just became a force multiplier for 50 agents. Available now to workspace admins and builders 👷

At the end of last year, like many other teams, we felt a significant phase transition in models coding capabilities. Agents shifting from sycophancy vibes to constructive push back. Across the team, engineers progressively bumped their grid-size (number of independent checkouts one has on their machine to parallelise agent work), but we started experiencing friction due to pre-agent baked assumptions in our dev-setup (port conflicts preventing having multiple dev instances running on the same machine and slow warm-up times form fresh checkouts). It was clear to us that we had to prioritise engineering work in 2026 to upgrade our dev setup to the user of larger grid-sizes. @henryfontanier came back from holidays with a present for us and casually dropped on the team `dust-hive` a full repackaging of our dev-env for agentic coding with batteries included: github.com/dust-tt/dust/tree… Henry wrote a full blog post presenting the technical details behind dust-hive some thoughts on impac of agents on software engineering at large: dust.tt/blog/engineering-at-… Each dev env is a bit specific and it's likely that every team will have to build a bit of specific software to adapt their environment to large grid-size with proper containment and rapid branching. Henry's work can serve as a solid starting point to do so in days not weeks. The interesting questions to me are (most covered in the article): (i) can this be standardized? the answer seems to be yes but would require a very rich configuration layer that goes beyond a Dockerfile. (ii) What does it mean for software engineering? We strongly believe at Dust that it means more taste and more technical competency are needed from software engineers not less. Almost like an inverted Jevons paradox. The world we live in.

Build vs Run and zero-sumness: a simple framework to reason about how to scale teams post artificial intelligence abundance. Will AI annihilate all head-counts? I posit it's not quite the case especially for run functions that are inherently a zero-sum game in the market. dust.tt/blog/build-vs-run

Want a promotion? Get rid of the mundane part of your work. @gabhubert (Co-Founder of @DustHQ): "Automating tasks is not automating roles. It's actually the best way to get promoted." By using AI agents for the 20-minute painful tasks, you free up time for strategic thinking and creative input. De-inhibit users: Just start with the first draft, like you'd tell your mom what to do.

and "this is only the beginning" !