📈 🚀 🌎 ✨
@PlanePowers is hiring globally with 31 open roles across 6 countries: US (14), India (8), Canada (3), Germany (2), Italy (2), UK (2).
We are 100~ people today and profitable. Capital efficiency deep in the DNA by our superstar founders @vamsi_kurama and @viharkurama.
The reason we’re growing the team this fast: we’ve >10X’d ARR two years running since our first revenue year in 2025, and we expect that to continue.
We just crossed 4,000+ paying business customers, including many dozens of enterprises running Plane in production with hundreds to tens of thousands of users org-wide.
Significant penetration across aerospace, defense, government, finance, manufacturing, engineering, and retail.
Come build the future of AI and agents working together : https://t.co/Ukj7Eb7A9g
LFM2.5-2.6B is a compact reasoning model with just 2.6B parameters: small enough to run on a phone and specifically trained for agentic workflows (planning, calling tools, and tackling multi-step tasks.)
Its post-training included agentic reinforcement learning inside real agent harnesses such as OpenClaw and Hermes Agent.
If you're already using OpenRouter, trying it is as simple as changing the model slug: liquid/lfm-2.5-2.6b:free
Give it a try: https://t.co/fGndbYNuMw
Massive refactor underway at @planepowers.
A brand-new design system is coming to both Community and Enterprise Editions. Around 7,750 legacy imports still need updating, and we’re aiming to finish the migration by the end of September.
We’ll open-source the design system too.
People keep saying SaaS is dead and UI is dead.
AI has accelerated almost every kind of execution. Code, research, writing and QA all move faster.
The same company still needs multiple apps and agents to answer a basic planning question: “What is blocked, and why?”
The context and coordination around the work have not kept pace.
Work is not optional, so every company ends up with a work graph. Large enterprises often build internal tools to keep that graph under their control, while mid-market companies often sit somewhere between a core stack and “let teams pick whatever.”
Tool choice can work when the underlying context stays connected. Without that, projects, calendars, wikis and support end up in separate systems, with identity, permissions, licenses and audit trails managed separately in each one.
When Elon bought Twitter and started talking about turning X into an everything app, I didn’t get it. A social network becoming the home for messaging, payments and identity felt like a stretch.
I’m starting to understand the ambition: the feed, DMs, payments and identity are meant to share one home.
Work software went the other way.
Subscription sprawl is easy to see. Missing context is harder to measure, and it costs more.
Humans made this fragmentation look integrated. We kept five tabs open and carried the joins in our heads.
An agent needs an authoritative place to read the work, change it and leave a trail; it can query five tools, but if those tools disagree about state, ownership or permissions, it inherits the disagreement.
It’s really quite mind blowing how @ramin_m_h has gone from academic to superstar startup founder CEO with 10s of millions of users, multiple $100m production enterprise customers… in 3 years flat. @liquidai.
Such an honor to know him as a friend, too. The journey is the reward.
with todays AI vintage of startups i see a few VCs calling for “50x a year growth or i dont invest” and that attitude will exclude them from fantastic companies
the “saasocalypse” is a myth and we will continue to see huge compounders in the future
dont compare yourself to the seven AI companies that went from 0 to $1B in revenue
all it takes is time
Proud to see my co-founder, phd co-advisor and postdoc advisor Daniela Rus, on the TIME's 100 most influential people in AI.
Daniela is one of the defining figures of modern robotics and AI. it is a humbling experience to share a 9 years professional journey with her as a mentor, advisor, and business partner.
As she describes it herself about her work @MIT_CSAIL and @liquidai, it is "all about the science & engineering of intelligence, & especially embodied intelligence” ✨
https://t.co/sdoeOtigNc
@planepowers agents are bot users.
Each has its own credentials and access only to the projects you choose. Mentions, assignments, comments and activity are recorded alongside everyone else’s.
They run from the work graph, not a chat sidebar. When a work item is created, a state changes, someone [@]mentions an agent or a schedule fires, a run starts.
Missing required field? Blocked transition? The run waits. It never guesses or invents an assignee.
Ask never loads write tools. Build plans the actions and waits for confirmation. A write-scoped agent executes within its granted scope.
Coming very soon!
I’m proud to share that WordPress has signed the Open Weights and American AI Leadership letter.
More people, not less, should shape the future of AI. Open Weights gives more builders the freedom to participate, continuing what WordPress stands for.
https://t.co/hgHIWAWYXI
So proud to see our @LiquidAI co-founder Daniela Rus, head of @MIT_CSAIL recognized today as top 100 people in AI by @TIME. 👏🏼 🥰
https://t.co/zQqEhevJpv
Turn product discussions, meeting notes, and rough ideas into actionable work. Give Plane AI a simple prompt, and get assigned, due-dated work items ready to pick up.
How smart is Apple’s built-in model vs the open LLMs on the same phone? Nobody had measured it, so I built a leaderboard.
Answer: 4th of 10. Three open models score higher; the clear leader is 1.2B, a 1.7 GB download. Apple’s is built in and free.
https://t.co/BqYxmvR1oR
Plane AI connects to the tools your team already uses through MCP and native integrations, bringing conversations, code, designs, and other context into your workflow.