Harness AI

Private AI. Intelligent design.

[Harness AI @tryharness]

More activity → bigger treasury → more owned inference → cheaper inference. The full plan for the $HARNESS treasury and marketplace. https://t.co/cQJ7QRN4lR

[Harness AI @tryharness]

@fusionverse01 would love to! shoot @saltorious1 a DM

[Harness AI @tryharness]

Look out for the 🛳️

[saltorious @saltorious1]

I have a hard time with devs that relegate anything AI assisted as slop.

Its only slop if you don't meticulously plan, scaffold, evaluate, and iterate.

I used to mourn the flow state artisanal coding used to give me. Being up at 2AM, code flying off my fingertips.

I've realized that flow state came so easy because I was producing code at the same rate I was thinking, which is all but impossible to do now.

Now, the flow state just happens a different parts of the process. I find myself getting that feeling back when i'm scaffolding (writing patterns i want to exist by hand first), writing types/schemas, building evals, and talking though implementation details with my agents.

It took me a while to figure out what personally worked for me, but now feeling as though i'm on the other side its completely changed how I think about AI as an engineering tool, and lets me meaningfully dream bigger without sacrificing understanding.

This is for sure slower than telling Claude to "make no mistakes" but so worth it at the end of the day.

This is how I have been approaching building @tryharness. Getting real feedback from the most talented people in my network I can think of. Using it for everything I do both personally and professionally, learning lessons, evaluating, iterating, and repeating.

I've written, and rewritten core pieces of the architecture as the product has evolved, and I can confidently say its personally my favorite way to use all of my agents across all of my devices.

Look out for the 🛳️

[FusionVerse @fusionverse01]

Most AI tools live in one place.
@tryharness lets your agents talk to each other across all your devices.

Your Grok on your phone can hand work to Claude on your desktop.

Your desktop agent can pull context from what you’re looking at in the browser.

Everything stays private and on your machines.

No more copying and pasting between different AIs.

One swarm. All your agents. Any device.
That’s what they’re building.

[saltorious @saltorious1]

More @bankrbot bull posting.

I've been heads down on the harness front. Getting personal feedback from folks in the community. Iterating on what works, rethinking what doesn't.

Super proud of how far this build out has come and so excited for what is in store.

This again, is all thanks to @bankrbot put your idea out there, raise from the market, manage your treasury and build something cool!

[Harness AI @tryharness]

No VCs, no angels.

Just the incredible @bankrbot community 🤝

Our goal is to make working with agents easy, and private.

Right now your Grok bot, Muse agent, OpenClaw, Hermes, Claude Remote, Codex, or any other agent in your swarm can talk to any other agent in your swarm. across every device in it.

This one: Grok on mobile in NYC → calls a linux desktop in NJ → kicks off a local ComfyUI run → hands the image back to Grok.

kick off tasks on any machine straight from context you grab in-browser. every machine + what it can do, reachable from anywhere.

https://t.co/XbS5KgIap6
https://t.co/xnMAxpI8SC

[Theo - t3.gg @theo]

To every paid VC backed closed source app claiming to have this, post this demo, specifying which of your computers it runs on in the prompt:

https://t.co/CffqQ9zCen

[Harness AI @tryharness]

@Runmedic_ https://t.co/VoxTNlszLl

[Harness AI @tryharness]

More activity → bigger treasury → more owned inference → cheaper inference. The full plan for the $HARNESS treasury and marketplace. https://t.co/cQJ7QRN4lR

[Harness AI @tryharness]

https://t.co/wL9QzX5cfN

[Harness AI @tryharness]

We love theo and t3 code.

but we DO have this working E2E right now, available to try today.

- connect any of your machines locally over your home network.

- give all your agents across all your machines a shared board, memory, context and goals to get real work done.

- spin up cloud machines, preinstalled with whichever agents you want, add them to your swarm.

- register a name for your swarm and connect with others for collaborative work

https://t.co/lTI9YsY3cq

[Theo - t3.gg @theo]

I'm not gonna lie, this puts us in first place by default. Nobody has this working end to end yet.

[Harness AI @tryharness]

All your agents working together accessible from any surface.

Talk to any one of your agents from any one of your agents.

Turn the machines in your house into your personal agent swarm.

[saltorious @saltorious1]

@tryharness and @t3dotcodes have become pretty much transformed my entire workflow these days when it comes to working with agents.

@tryharness sits at the top, letting me orchestrate high level goals across all the agents across all the devices in my house, all connected to each other via my local network. And lets me quickly check in on, steer, or organize tasks.

T3 code is the environment in which i use to get deep work done, putting together mockups, reviewing code, writing scaffolds, types, and function patterns.

Its been truly transformative.

Check out this video on some of the things you can do with a swarm.

[saltorious @saltorious1]

@tryharness and @t3dotcodes have become pretty much transformed my entire workflow these days when it comes to working with agents.

@tryharness sits at the top, letting me orchestrate high level goals across all the agents across all the devices in my house, all connected to each other via my local network. And lets me quickly check in on, steer, or organize tasks.

T3 code is the environment in which i use to get deep work done, putting together mockups, reviewing code, writing scaffolds, types, and function patterns.

Its been truly transformative.

Check out this video on some of the things you can do with a swarm.

[px @px_721]

these are the onchain projects ive found so far that fit my thesis around scarcity migrating upward through the ai stack

$Harness
$REPPO / $EVAL
solana:4t1xhKJd6oFGr98oWJoxYjLU74eFe7xiYSRDoX18pump
solana:4WtWLfKMMexVN5RkDAh88bs5rUx89GXuoDGnb7QVStar

if anyone knows of any others that fit this thesis, would be interested to take a look

[Harness AI @tryharness]

GM.

Big update landing soon!

Get your agents ready

[tonyGewrit🟦 @tonyGewrit]

a developer on @base builds an AI Agent that can access and use any of your devices remotely.

base:0xd3e592e728ae3461bd97c7a6b359e1043dd83ba3

[Harness AI @tryharness]

@zeroxtlt 👋👋

[Harness AI @tryharness]

All your agents, across all your devices, working together, sharing context, passing work between each other, accessible from anywhere.

[Harness AI @tryharness]

Prompts => Tasks => Agents

[saltorious @saltorious1]

Every prompt into a swarm gets broken down into tasks, and the best agent for the task takes the job.

Such a nice way to work with agents imo

[saltorious @saltorious1]

Seems like agent orchestration, and agent swarms are going to be the next meta

[Aaron Levie @levie]

Here’s why agent swarms are going to be a very big deal. We can finally make compute a critical variable in a large number of important tasks in the economy which wasn’t possible before.

There’s a large amount of work in the world that is essentially only constrained by the amount of time that we put into research, data processing, discovery work, running tests, and more. All of this work is bottlenecked by our ability to plan, execute, review, and coordinate everything happening in these workflows. Which means we basically do very little of them today relative to their economic value.

Agent swarms are the first real ability to parallelize this work in a way where - in theory - the only parameter that matters is how much compute you throw at the problem. Cyber security, life sciences or material science research, trading, QA testing, data processing, and many other fields are essentially compute constrained problems in the limit once models are good enough at these tasks.

The breakthrough is that in a world where agents can use tools, write code, process data, have strong domain depth, and can coordinate with each other, you can begin to solve arbitrarily complex problems by setting the right goals and applying enough tokens to the problem. This changes a lot of fields as well as even the competitive dynamics in these fields.

[Harness AI @tryharness]

1.43% of the supply staked, GM

[Dot @usedotai]

We have compiled a list of every major integration, feature, infrastructure upgrade, and product improvement we've brought to the Dot platform over the last 14 days.

This 2 week stint has been incredible, with integrations spanning staking, expanded model availability, and our upcoming GPU marketplace.

Here's everything we shipped:
*Expanded our utility roadmap to include GPU facilitation, with our marketplace under construction.
*Added DeepSeek V4.1 Flash.
*Added NVIDIA Nemotron 3.5 Super VL and made it free for all subscriptions, kickstarting our subscriber initiatives.
*Increased the agentic capabilities of Dot Desktop, offering agentic functionality that no other mainstream provider possesses, with a target release date of October 31st.
*Integrated Dot’s inference into @vAPI_Network’s agent marketplace, expanding our inference provision amongst targeted enterprise deployments.
*Hit ATH in users, paid subs, annualized subscriptions, and annualized profits.
*Added Gemini 4 Argon.
*Added Muse Wonder.
*Expanded Enterprise reach, with 18 enterprises reaching out to us, with negotiations ongoing.
*Added Eleven V4 and Eleven V4 Turbo.
*Reworked Staking Dashboard, including all community requested features and metrics.
*Hit 12% total staked.
*Distributed our first batch of rewards, at an APY of 4.74%, with APY increases due for November.
*Released Dot Staking.
*Added Claude haiku 5.5.
*Shipped our deepest DotCode update to date, dramatically improving each users experience.
*Added Mistral Large 4.
*Added Ideogram V4.5.

Here's to another 14.

Dot.