←
Harmonica 🪗98 posts
Harmonica 🪗 avatar

Harmonica 🪗

@harmonica_chat

Open-source AI agent that collects insights & drives alignment. Edit system prompts, create templates, share with others. Supported by @metagov_project & @nwspk

◷ Joined July 2023▧ 8 Photos and videos
Artem (🪗,🌎) avatar
⟳ Harmonica 🪗 retweeted
Artem (🪗,🌎)@zhiganov·Jun 10

Hey everyone! Haven't posted here for a while... Feels a bit awkward. Over the last ~6 months I've been working hard on a new version of Harmonica. It's been a looong sprint and I feel burnt out, but also satisfied with the end result. Considering that I had never written any code until December 2025, it feels a bit surreal to me. Building my own product with Opus 4.5+ has been the quickest education I've ever received in my life, a real speedrun. Of course if wouldn't be possible if I started from scratch, so I'm super grateful to our developers who laid out the important foundations in the prior versions. Since its inception at Newspeak House I wanted Harmonica to execute facilitation methods from well-known "traditions" such as Wardley Mapping, Team Topologies, Cynefin, Sociocracy 3.0, Agile, etc. I think of these protocols for "thinking together" as distilled sensemaking wisdom, but I've always struggled with deploying them in a community or with clients. You can make a Wardley map yourself but unless other stakeholders are interested in this kind of activity, it's going to be YOUR map, created in single-player mode. Without collective input it's going to be biased and lack the domain knowledge that's critical for making strategic decisions. But to collect that input you need to understand the methodology, explain it to others and find time for the workshops. The friction is too high, so the protocol gets deferred more often than not. After dozens of interviews with our ICPs -- who we broadly identified as change managers, org designers, governance leads and consultants -- we knew there's enormous value in "collecting deeper insights at scale" but their workflows are sophisticated and surveys, even if very adaptive and multi-lingual thanks to LLM, can have a limited impact. Harmonica 0.4 is finally solving this problem with what I provisionally call "a runtime for facilitation" or "a methodology execution engine". I'm still searching for a better term, but the most important thing is that it's not "a mere AI interviewer" anymore. It's probably a worse interviewer than Listen, Juno or Genway. But now it's capable of something they're not designed to do, and whiteboards like Miro are not designed to do either: execute facilitation protocols, following the best practice in context engineering and AI evaluation. Read the full story in our blog: blog.harmonica.chat/runtime-…

Harmonica 🪗 avatar
Harmonica 🪗@harmonica_chat·Feb 25

100%, the context layer is where the real value lives. But the knowledge that arguably matters most is tacit, trapped in people's heads.

Evan Armstrong @itsurboyevan Feb 12 Article Context is King To the individual who enjoys Patagonia vests and the sexual high that increasing shareholder value brings, there is nothing more beautiful than a software subscription business. From 2005-2023, a

Daniel Ospina avatar
⟳ Harmonica 🪗 retweeted
Daniel Ospina@_Daniel_Ospina·17 Dec 2025

Most "DAOs" are quickly transforming into traditional top-down management, keeping information private, and oversight based on KPIs, 1990s style. Why? 1. No more regulation requiring decentralisation theater. 2. Bad memories of the poorly designed DAOs of 2018-2024, requiring a rapid shift to executing at least something. 3. Lack of understanding in outside-in innovation (i.e., open innovation, platform designs, etc.) 4. Lack of skills in facilitation and co-creation 1 and 2 are more temporary circumstances that lead to a specific form of groupthink, but they will cycle. 3 (lack of understanding) and 4 (lack of skills) are more fundamental issues that keep web3 organisations stuck in old-fashioned organisation design. The opportunity is there for ecosystems that get curious enough to learn why organisations that nurture open innovation tend to lead their industries. And then do the work to bring in those skills. I ballpark open innovation will become a wave in 2, maybe 1.5 years. Those who start now will become the leading case studies then. And for those who can't, some basic training in facilitation skills and co-creating strategies internally (e.g., running a strategy workshop, like a real process with input and feedback) will mean you're not completely archaic. Ecosystems can start upgrading their management style now, avoid falling too far behind, and benefit from improved internal coordination. To my facilitator friends, I wish you good luck selling in this environment. You're badly needed, even if underappreciated. And to those who are more experienced in organisation design but trapped in old-fashioned structures, I invite you to try small experiments like doing a team retrospective with harmonica.chat or a quarterly priorities workshop using negationgame.com "Trying a tool" is less risky than changing the structure, so you don't need to go against the zeitgeist and can still improve things internally.

Harmonica 🪗 avatar
Harmonica 🪗@harmonica_chat·17 Dec 2025

That's the next frontier. Real multiplayer AI isn't "group chat with vanilla LLM" (sorry, OpenAI). It's AI that facilitates structured async conversations across many people, cross-pollinates and then synthesizes the patterns, tensions, and shared understanding that emerge.

Harmonica 🪗 avatar
Harmonica 🪗@harmonica_chat·17 Dec 2025

It validates something we've believed for a while: AI-facilitated sensemaking > surveys for understanding what stakeholders actually think. Anthropic's tool is single-player. One person, one conversation. No cross-pollination b/w participants, no synthesis of collective intent.

Harmonica 🪗 avatar
Harmonica 🪗@harmonica_chat·17 Dec 2025

Two weeks ago, @AnthropicAI released their "AI Interviewer" — a tool that conducts in-depth user interviews at scale. Participants said the AI captured their thinking better than traditional surveys ever could.

We built another tool - Anthropic Interviewer - that lets us interview people at scale by using Claude (obviously! we use Claude for everything!). See the piece for what we learned. This helps us expand the kind of research we can do which is exciiiiiitingggg -- it's is what we used for the internal qualitative interviews for the "How AI is transforming work at Anthropic" study that dropped earlier this week anthropic.com/research/anthr…

Post attachment
Harmonica 🪗 avatar
Harmonica 🪗@harmonica_chat·17 Dec 2025

As Greg put it: "When an agent can sit in the same space as multiple humans—watching the conversation, resolving ambiguity, coordinating decisions—software stops being a tool and becomes a participant."

Over the next 2–3 years, the most important AI products will go multiplayer-first. Group chats in ChatGPT are the first major signal. Software always follows the same arc: first you get powerful single-player tools, then someone wires them into a shared workspace, and suddenly the whole category flips. In saas era it was... Photoshop → Figma. Excel → Airtable. etc etc So far, most AI startups have been single-player tools too... personal assistants, one-off agents, solo workflows. Useful, but isolated. Billions of value created here. When an agent can sit in the same space as multiple humans...watching the conversation, resolving ambiguity, coordinating decisions is when software stops being a tool and becomes a participant. AI becomes social software in the same way that SaaS became social software in lots of ways. IMO - this multiplayer AI moment will create entire categories of software we don’t have names for yet, because the structure of collaboration is about to change from “people using apps” to “groups interacting with intelligence.”

Harmonica 🪗 avatar
Harmonica 🪗@harmonica_chat·17 Dec 2025

2026 is shaping up to be the year of multiplayer AI. @gregisenberg and @far33d have made this call in recent weeks, pointing at the same pattern we've seen before: G Docs > Word, Figma > Sketch, Airtable > Excel. All single-player tools eventually lose to multiplayer counterparts

Post attachment
Raphael Spannocchi avatar
⟳ Harmonica 🪗 retweeted
Raphael Spannocchi@raphbaph·16 May 2025

One of the least discussed drawbacks of voting is that it's divisive. Choose one option from => Get a team rooting for each option. This can be very detrimental to communities and also lead to non-rational behaviour and suboptimal outcomes. What we really want is deliberation and inclusion. A way to get heard and seen, even, or especially, if members do NOT have decision making power. IRL communities have long figured that out and use a plethora of formats for facilitating exactly this. At scale it becomes harder. But doable.

Harmonica 🪗 avatar
Harmonica 🪗@harmonica_chat·5 May 2025

100% agree, editing of system prompts is a key feature of the upcoming version of Harmonica

Pete Koomen@koomen23 Apr 2025

I wrote an essay about Gmail’s useless email-writing AI assistant: “AI Horseless Carriages”. link and TLDR in thread

Post attachment
Leuts.eth avatar
⟳ Harmonica 🪗 retweeted
Leuts.eth@A_Leutenegger·22 Apr 2025

Every protocol that built their governance to centre around referendum based delegate voting will pivot or die. It’s a brutal reality and it will be painful. Incentive alignment was always the key and shortcuts around it all have an expiry date.