Do We Need DAOs? A Critique
If DAOs are the future, why do they keep failing? Decentralized Autonomous Organizations (DAOs) sound revolutionary. But when examined closely, many of the problems they claim to solve are either already solvable with simpler tools, or are made worse by the DAO structure itself.
That said, I am not arguing for the preservation of the status quo either.
We do need something better than what we have today. But before reaching for new names, new tokens, and new governance experiments, we need to clearly understand the problems we are trying to fix.
The next evolution will not come from new tech alone. It will come from thoughtful design by those who are building within a decentralized mindset.
Why? Because decentralization is not merely a stack of protocols. It is a design philosophy rooted in three goals:
Give people real participation. Protect against unchecked power. Make systems that outlive their founders.
Those are the tools we should be building the future with. This article examines the arguments made by @VitalikButerin in his post below, and reframe it on what problems DAOs are actually trying to solve, where they fail, where they succeed, and how we might move forward without copying the wheel and calling it a wing.
We need more DAOs - but different and better DAOs. The original drive to build Ethereum was heavily inspired by decentralized autonomous organizations: systems of code and rules that lived on decentralized networks that could manage resources and direct activity, more
So... Do We Actually Need DAOs?
The answer is complicated. Because DAOs have both failed spectacularly in many ways, and yet, the problems they aim to solve aren’t going away.
Here’s a full breakdown of my arguments against DAOs, rebuttals to my points, and where we together might go from here.
1. “We need DAOs to create better oracles.”
Vitalik claims that we need better oracles (do we?!) and that current oracles aren’t secure or decentralized enough. DAOs could help curate or validate better ones.
But Oracles are ultimately authority systems. Humans will act the same as they always have regardless of the technology mediating them.
Token voting does not solve manipulation. Large token holders can still distort outcomes, especially when the data is subjective. If humans curate the oracle, it is no longer autonomous or neutral. It is simply a committee with a new name.
Replacing flawed traditional systems with structurally similar DAO systems does not meaningfully change outcomes. It often slows improvement by adding governance overhead to what is primarily a technical problem.
Solution? Pure DAO voting is not a good oracle. Pure centralization is also brittle.
The most resilient designs combine automation, cryptography, limited human escalation, and transparent oversight.
Governance should support the system, not run it.
2. “We need DAOs for onchain dispute resolution.”
There are enough failed DAOs out there that anyone can see that these are bad for decision making.
Token voting leads to low participation, whale dominance, and shallow engagement.
Most people do not want to vote every week.
Over time, governance becomes symbolic rather than substantive.
Justice requires expertise, privacy, and neutrality. Crowds rarely provide these qualities.
The size of the crowd does not determine how right their collective decision is. If everyone was jumping off a cliff, would you jump too? Human psychology would say most people would without question.
Traditional organizations make fast decisions, but often behind closed doors, with limited visibility and little recourse for participants.
So, the failure here is not decentralization itself. It is immature governance design.
Better models include delegation, reputation based influence, scoped voting, and layered decision rights.
We do not need constant voting. We need selective participation with clear accountability.
This can exist with or without a DAO.
We have to remember, DAOs weren't optimized for speed.
Community-based arbitration is already biased. Public votes in emotionally charged disputes can become popularity contests.
Token-based voting doesn’t work well in subjective cases. It incentivizes manipulation (e.g., vote brigading, vote buying).
You could build reputation-based juries, specialized arbitration courts, or legal-integrated contracts without invoking DAO governance models.
Bottom line: “More votes” doesn’t mean better judgment. Justice requires expertise, privacy, and neutrality, not crowd dynamics like DAOs have now.
2. AI + DAO = Decision Overload?
AI can solve decision fatigue! But at what cost?
When AI votes for users, agency is lost.
Governance becomes abstract and opaque. Using AI simply replaces one form of centralized power with another.
AI as a voting assistant opens huge doors to manipulation: Who trains the AI? Who audits its “values”? What if it gets socially engineered?
If an AI votes for you, you’re not governing, you’re outsourcing agency to code written by others with do not have your best interests at heart.
Most people don’t want to be involved in governance, even passively. That’s not fatigue. That’s disinterest or misaligned incentives.
The more abstract we make decision-making, the easier it becomes to hide bad actors, misaligned goals, or poor outcomes behind the AI veil.
Bottom line: Decision fatigue is real, but AI does not have to replace agency.
Instead of using AI to vote, use it to reduces cognitive load without removing the human choice.
Local user controlled models combined with privacy tools can allow assistance without central control.
Humans are lazy and most will give up their power to something that will just take care of things, but I will always advocate that AI should support judgment, not substitute it.
So in sum, the problem is not AI. The problem is relying on AI to avoid designing better human systems.
AI can enhance tooling but cannot and should not replace governance or informed human participation.
Otherwise, why even pretend it’s a DAO?
3. “We need DAOs to maintain lists.”
Original Point: DAOs can help manage lists of safe apps, token contracts, verified projects, etc.
I think list curation should be handled by human audits, protocols, and transparent documentation. In practice, a small group often controls visibility and access anyway.
At scale, lists become power structures.
Who controls the list controls legitimacy and access.
As systems grow, neutrality and continuity matter more than speed. In some cases, shared governance provides legitimacy when no single entity should own the registry. Most of this work is better done by protocols, indexes, or community tools, not by bloated voting systems.
“List curation” sounds decentralized until a handful of whales determine which projects get visibility or funding.
Security lists and allowlists should prioritize audits, evidence, and verifiable records, not token-based groupthink.
Bottom line: Curation is a task that calls for experts, transparency, and standards, not for generalist voters.
Using DAOs to manage lists is unnecessary.
Blockchain can help solve this, but the way DAOs do this? Do nothing but add friction without improving quality. Maybe in the future when they have evolved.
Lists become power structures at scale.
Who controls the list controls access, legitimacy, and visibility. As systems grow, neutrality and continuity matter more than speed.
DAOs can provide shared stewardship when no single entity should own the list.
For small systems, simplicity beats governance.
For large shared infrastructure, stewardship matters. But DAO based list management needs far more thoughtful design than what exists today.
4. “We need DAOs to get projects off the ground quickly.”
Original Point: DAOs let project start quickly because it gets around all the red tape of forming a legal entity. Early stage projects need leadership, clarity, and speed.
But as mentioned above, DAOs are not built for speed, and they stall during formation because momentum slows, founders leave, and results in zombie governance with no execution.
Most DAOs spend weeks to agreeing on things.
Fast-moving teams are already self-organize with clear leadership.
DAOs often bring legal gray areas that can block future progress (e.g., securities issues, lack of accountability, tax problems.)
From what I have seen, DAOs add more red tape than they remove. You don’t need a DAO to crowdfund and collaborate.
So my recommendation? Design for handoff. Succession can be supported through legal entities, foundations, councils, or onchain tools. Progressive decentralization can help systems survive leadership transitions.
Governance should emerge when it adds value, not before.
I mentioned before that it feels like DAOs are just trying to reinvent the wheel.
Early stage projects do not need DAOs for at this time, they adds cost and complexity without proportional benefit. Most cases can be handled with simpler tools.
BUT, I will acknowledge that trust does not scale easily, and most people have lost trust in current systems that we use today.
When participants are global, pseudonymous, and not legally bound, traditional tools break down. In these environments, onchain coordination can become useful.
In sum: 1. Use the simplest tool that fits the trust environment. 2. DAOs come with their own problems and issues. 3. They are situational tools to help larger problems.
DAOs are not replacements. They are situational tools for specific coordination problems.
5. “The communication layer matters.”
Original Point: DAOs need better forums and platforms. Token votes alone aren't enough.
Most DAOs fail at communication, but this isn’t a DAO innovation. It’s a UX/community design challenge.
There are already better tools (e.g. Discourse, Reddit, GitHub discussions) that support productive asynchronous input. They aren't perfect, but they are where people are. So improving what those can do, rather than replacing with something completely new should be a priority.
Crypto has mostly ignored this layer in favor of token mechanics.
Bottom line: Better communication should be a priority for all online communities. Culture is infrastructure.
So without strong communication systems, no governance model works.
6. “Convex vs Concave Problems” Smart framing, wrong assumption
Original Point: DAOs should be used based on the type of problem. Some problems need compromise, some need decisive action.
Great in theory. In practice? Most real-world decisions are mixed. you need nuance, timing, and flexibility.
DAOs often lock decisions into rigid processes with tokenized weight. There’s little space for intuition, pivoting, or contextual judgment.
You can still design collaborative systems without creating a DAO.
Just build around the problem, not the acronym.
We Need DAOs, But Not as an Ideology
We do not need DAOs because the technology is elegant.
We need better coordination because the problems are real.
-Trust is low -Power concentrates quickly -Founders leave etc etc.
DAOs are one response, not The Response.
The mistake we have been making is that we are treating them as an end state rather than a design space.
The best path forward is hybrid.
Take from DAOs: Transparency Shared ownership Composability Permissionless participation
Combine with traditional strengths: Clear leadership Accountability Human judgment Legal clarity
We should build systems focused on solving real problems, not proving philosophical points.
If we stop worshipping structure and start designing for reality, we do not need to keep trying...
We simply build better.