Last week, Tally announced it is winding down operations after six years of powering governance for over 500 DAOs. The coverage has been predictable:

"DAOs are dead."
"Governance tooling has no business model."
"Decentralization was just regulatory theatre."

We think this diagnosis is wrong. Not because the symptoms aren't real, but because it confuses the failure of a specific product category with the failure of coordination itself.

Tally built a version of what governance tooling was understood to be: voting portals, delegation dashboards, proposal workflows. It served many users, but they couldn't build a sustainable VC-backed business. That fact deserves examination, not a eulogy for the entire field.

The first generation of governance tooling did not play out as expected. But the deeper issue is that even where DAOs did thrive, the experience of participating in governance remained fundamentally broken. Voter turnout stayed low. Proposal quality was uneven. Treasury decisions were made by tiny fractions of tokenholders.

Voting portals can shine a spotlight on governance, but we need real innovation if the ecosystem is going to evolve.

At Lighthouse Labs, we've spent the past three years working at the intersection of mechanism design, identity infrastructure, and governance research. Not because governance tooling seemed like a good market opportunity, but because we believe the coordination layer for internet-native organisations hasn't been invented yet.

That conviction has produced work across several areas that, taken together, represent a fundamentally different approach to the problem.

If you can't describe what an organisation is, you can't govern it.

The Node Metadata Standard, currently in draft status as an ENSIP, introduces a composable schema for attaching structured metadata to ENS names. Think of it as a way for any entity represented by an ENS name or subname to also describe its role, capabilities, and relationships in a machine-readable format that lives on-chain.

This is more than just profiles and avatars, it's the precondition for governance that understands who is participating, in what capacity, and with what context. Without it, every governance interface is flying blind.

The foundations of this new standard are already out there and being discussed, especially within the context of debates about the future of agent identity architecture on ENS. It's our most recent contribution to the ecosystem, and it's the foundation everything else builds on.

We proposed one of the first major DAO retrospective’s together with James (🔥_🔥). What has since been produced by the Metagov team creates an empirical baseline for what in the past has been mostly anecdotal.  

Concurrent to this work, we ran AI-assisted grant evaluation experiments. Using language models to systematically assess SPP2 applications against consistent criteria, we explored the idea of automated or augmented screening to reduce operational burdens in decentralized organizations. The results weren't a replacement for human judgment, but they revealed how much can be done with the data we already have if you just run it through the right tools and processes.

We also implemented and improved upon the well-renowned Talk to the City method and ran it in parallel with the Retro to compare and contrast findings.

This research matters because it demonstrates something most governance teams skip: empirical accountability. If you're going to build tools for collective decision-making, you should first understand how collective decisions actually get made and where they break down.

Token-based, closed-list voting has been the centerpiece of collective decision-making for too long. Real-world experience so far has shown that forcing users into a binary “yes/no” choice results in difficult decisions and inflexible initiatives.

Using similar economic systems to allow voters to instead express a preference (and preference intensity) before moving to the final vote can expose true community sentiment and result in better-planned proposals going up for a vote.

This is what we provide with the Signals Protocol, which allows voters to temporarily lock up capital in support of their preferences. The result is better use of the “wisdom of the crowd” while the economic game theory of anonymous blockchains still works to enforce good behavior.

We're clear-eyed about the market. Tally's closure will accelerate consolidation in governance tooling, and many teams are already moving to fill the gap. This is natural and healthy. The ecosystem needs multiple independent frontends and diverse approaches.

But we'd gently observe that most of the emerging responses to Tally's shutdown are variations on the same theme: better voting interfaces, more security monitoring, slicker dashboards. Although these are necessary, they are not sufficient. We need deeper structural reforms in our behaviors and an openness to experimentation if we truly believe that internet native organizations  have a viable future.

Lighthouse Labs builds coordination infrastructure for decentralised systems. Our contributions to Ethereum include Harbor Protocol, Signals Protocol, EGPL, and the ENS Metadata Standard. We are active contributors to ENS DAO governance and the broader Ethereum ecosystem.

Last week, Tally announced it is winding down operations after six years of powering governance for over 500 DAOs. The coverage has been predictable:

"DAOs are dead."
"Governance tooling has no business model."
"Decentralization was just regulatory theatre."

We think this diagnosis is wrong. Not because the symptoms aren't real, but because it confuses the failure of a specific product category with the failure of coordination itself.

Tally built a version of what governance tooling was understood to be: voting portals, delegation dashboards, proposal workflows. It served many users, but they couldn't build a sustainable VC-backed business. That fact deserves examination, not a eulogy for the entire field.

The first generation of governance tooling did not play out as expected. But the deeper issue is that even where DAOs did thrive, the experience of participating in governance remained fundamentally broken. Voter turnout stayed low. Proposal quality was uneven. Treasury decisions were made by tiny fractions of tokenholders.

Voting portals can shine a spotlight on governance, but we need real innovation if the ecosystem is going to evolve.

At Lighthouse Labs, we've spent the past three years working at the intersection of mechanism design, identity infrastructure, and governance research. Not because governance tooling seemed like a good market opportunity, but because we believe the coordination layer for internet-native organisations hasn't been invented yet.

That conviction has produced work across several areas that, taken together, represent a fundamentally different approach to the problem.

If you can't describe what an organisation is, you can't govern it.

The Node Metadata Standard, currently in draft status as an ENSIP, introduces a composable schema for attaching structured metadata to ENS names. Think of it as a way for any entity represented by an ENS name or subname to also describe its role, capabilities, and relationships in a machine-readable format that lives on-chain.

This is more than just profiles and avatars, it's the precondition for governance that understands who is participating, in what capacity, and with what context. Without it, every governance interface is flying blind.

The foundations of this new standard are already out there and being discussed, especially within the context of debates about the future of agent identity architecture on ENS. It's our most recent contribution to the ecosystem, and it's the foundation everything else builds on.

We proposed one of the first major DAO retrospective’s together with James (🔥_🔥). What has since been produced by the Metagov team creates an empirical baseline for what in the past has been mostly anecdotal.  

Concurrent to this work, we ran AI-assisted grant evaluation experiments. Using language models to systematically assess SPP2 applications against consistent criteria, we explored the idea of automated or augmented screening to reduce operational burdens in decentralized organizations. The results weren't a replacement for human judgment, but they revealed how much can be done with the data we already have if you just run it through the right tools and processes.

We also implemented and improved upon the well-renowned Talk to the City method and ran it in parallel with the Retro to compare and contrast findings.

This research matters because it demonstrates something most governance teams skip: empirical accountability. If you're going to build tools for collective decision-making, you should first understand how collective decisions actually get made and where they break down.

Token-based, closed-list voting has been the centerpiece of collective decision-making for too long. Real-world experience so far has shown that forcing users into a binary “yes/no” choice results in difficult decisions and inflexible initiatives.

Using similar economic systems to allow voters to instead express a preference (and preference intensity) before moving to the final vote can expose true community sentiment and result in better-planned proposals going up for a vote.

This is what we provide with the Signals Protocol, which allows voters to temporarily lock up capital in support of their preferences. The result is better use of the “wisdom of the crowd” while the economic game theory of anonymous blockchains still works to enforce good behavior.

We're clear-eyed about the market. Tally's closure will accelerate consolidation in governance tooling, and many teams are already moving to fill the gap. This is natural and healthy. The ecosystem needs multiple independent frontends and diverse approaches.

But we'd gently observe that most of the emerging responses to Tally's shutdown are variations on the same theme: better voting interfaces, more security monitoring, slicker dashboards. Although these are necessary, they are not sufficient. We need deeper structural reforms in our behaviors and an openness to experimentation if we truly believe that internet native organizations  have a viable future.

Lighthouse Labs builds coordination infrastructure for decentralised systems. Our contributions to Ethereum include Harbor Protocol, Signals Protocol, EGPL, and the ENS Metadata Standard. We are active contributors to ENS DAO governance and the broader Ethereum ecosystem.