What's broken in the Technical Fellowship (and a small set of fixes)

Mark65 · May 4, 2026, 8:44am

I’ve been around Polkadot governance long enough to have read the manifesto a few times and to have followed Fellowship referenda on Subsquare for about a year. So this isn’t a drive-by post. It’s also not a hit piece on anyone. The Fellowship works. RFCs get ratified, runtimes ship, the dashboard exists, the monthly calls happen. As a coordination layer the body does the job.

But the manifesto as it is today (Draft 5) is silent on the three things that actually decide whether this body stays healthy:

how the Whitelist works
how the sub-treasury works
what happens when a Fellow’s day job at Parity overlaps with their Fellowship vote

And on top of that, only Masters can propose changes to the manifesto itself (§6.2). So the body cannot really fix itself unless its smallest group decides to.

A few things first. Nothing here is about individuals. It’s about the text and what the on-chain record looks like.

The problems, in plain language

The Whitelist is the most consequential thing the Fellowship does, and the manifesto doesn’t even mention it. Search Draft 5 for the word “whitelist”. It isn’t there. Yet the Fellowship is what fast-tracks runtime upgrades and OpenGov calls. There’s no checklist, no required written rationale, no list of named reviewers, no record of dissents. The recent runtime upgrades v1.6.1 and v1.6.2 (Polkadot #1645 and #1736) went out this way without a public review packet.

The sub-treasury isn’t in the manifesto either. After Polkadot #1722 was voted Nay, the Fellowship moved its operational funding into an internal RFP from a sub-treasury. Fine as a budget choice, but it removed the token-holder veto without replacing it with audited public accounts.

A lot of Fellows are paid by Parity and also paid a Fellowship salary. This is in the Q3 2025 Treasury Report: “a significant chunk of Fellowship members are Parity employees, who receive a generous Fellowship salary.” The manifesto has no rule against double payment for the same work, no register of interests, no recusal duty. The four “tenets” are nice words. They are not enforceable.

Voting weight grows fast with rank. 1, 3, 6, 10, 15, 21, 28, 36, 45. So three Masters can outweigh a couple of dozen juniors on any internal vote. Add the previous point and you get a small group making most decisions.

Promotion above Rank III gets fuzzy. “Discussion points”, no published rubric, no precedent register, no obligation to write down why you voted the way you did. This is fine until someone disputes a promotion and there’s no paper trail.

Only Masters can propose changes to the manifesto. §6.2. Members and Fellows have no constitutional voice. The people most attached to the status quo are the only ones allowed to change it. This is why nothing on this list has been fixed yet.

No one publishes turnout. No quorum, no per-referendum participation, no abstention rate, no roll-call. The events are on chain. Someone has to actually publish the numbers. Every internal referendum I sampled passed. That can be alignment. It can also be rubber-stamping. From outside there is no way to tell.

No legal wrapper. The manifesto says it directly: not a legal entity in any jurisdiction. In practice this means the Fellowship can’t sign a contract with a vendor, can’t hold a trademark, can’t sign a data agreement with a custodian, and (this matters) doesn’t protect active Fellows from being sued personally if something goes wrong. The Wyoming DUNA was created for exactly this. Swiss verein and Cayman foundation work too. Pick one.

No sanctions screening, no aggregate numbers on the roster. A body that pays salaries globally and doesn’t check OFAC / EU / UK lists is asking for trouble. Thirty-day fix.

No way for a Master to retire, and no clean way to remove one. No Emeritus status, no panel, no evidentiary standard, no appeal. The only formal route is a Polkadot OpenGov referendum, which is the nuclear option for what should be a routine HR procedure.

Patches happen ad-hoc. The September 2025 OpenDev call mentioned a “retention-vote loophole” patched only for the 32 founding members. That kind of one-off shows the rules don’t handle edge cases cleanly. Symptom, not cause.

What I’d actually do about it

A single annex package using the path the manifesto already provides (§6.2). Severable, so the Fellowship can take what it likes and reject what it doesn’t. Roughly:

A. Conflicts of interest. Public register, mandatory disclosure of employment by Parity / W3F / parachain teams, recusal when there’s a clear interest, no double pay for the same work.
B. Vote transparency. Publish roll-call, turnout, abstentions and a short written rationale within 30 days of each referendum. Quarterly report.
C. Voting cap. No single employer over one third of the rank-weighted vote. Add a one-Member-one-vote secondary count for ratifications.
D. Rubric and precedent register. Write down the criteria for each rank. Voters add a short rationale (500 chars). All of it goes into a precedent register.
E. Lower the amendment threshold. A Master plus a Fellow can co-propose. Or three Fellows on their own. End the Master monopoly.
F. Secretariat and Whitelist procedure. Five elected seats. Every whitelisted call gets a one-page checklist signed by at least two Architects, named reviewers, dissent log.
G. Treasury discipline. Quarterly audited accounts. Anything over 50,000 DOT goes to OpenGov.
H. Legal wrapper. Wyoming DUNA, or Swiss verein, or Cayman foundation. Thin, no governance power, just so we can sign things and protect Fellows from personal exposure.
I. Sanctions screening and basic roster numbers. Annual OFAC / EU / UK check. Anonymised aggregate stats on jurisdictions and employers.
J. Exit and discipline. Emeritus retirement status. Three-arbitrator panel for disputes. Real evidentiary standard. Appeal to OpenGov.

A few things I’m not saying

I’m not saying any individual Fellow has done anything wrong. The argument is structural.

I’m not saying Parity shouldn’t employ Fellows. The expertise is where it is. I am saying the same person shouldn’t vote on a matter their employer benefits from without recusing, and shouldn’t be paid twice for the same piece of work.

I’m not pushing the five-Fellowships split. It’s a bigger surgery than what we need.

I’m not trying to slow the body down. The whitelist checklist is one page. The rationale field is 500 characters. The rubric just writes down what already happens.

What I’m asking

Fellows of any rank: push back here on which annexes are wrong and why. Not in the abstract, point to the text. The package is severable, I’d rather adopt eight of these than zero.

Masters: would any of you co-sponsor at least the conflicts annex (A) and the legal wrapper annex (H)? Those are the two that matter most.

W3F: it would help if you said publicly where you stand on A and H. Your position will move adoption either way.

Happy to keep a HackMD with the full text and integrate redlines from this thread. If there’s appetite I’ll open a thread per annex so this doesn’t turn into one giant scroll.

Posted in good faith. Citations to manifesto sections, on-chain referenda and prior forum threads available on request.

shawntabrizi · May 4, 2026, 11:46pm

I think we are also missing dedicated support for mentorship and growth within the fellowship.

It should be a key function of those who are in a position to meaningfully contribute to Polkadot to in-fact help others push those contributions instead. However, the current design of the fellowship does not really allow members to rank up in primarily mentorship roles.

Regarding vote transparency, I really don’t know what you mean here. Any 10 second Claude session could create a script which gets you the details you are requesting.

jobyreuben · May 5, 2026, 4:24pm

This is a philosophical response, not a technical one. So readers kindly don’t expect the usual ritual of dressing weak foundations in complex terminology just to make them sound defensible.

I think this post gets to the real issue, and honestly its bigger than whitelist procedure or treasury structure. Polkadot has spent years trying to solve fundamentally human governance problems with increasingly elaborate game-theoretic machinery, as if every judgment call in life can be converted into incentive model, wrapped in elegant runtime conditions.

Because apparently trust, judgement and accountability were just inconvenient legacy features waiting to be patched out by another governance dashboard and runtime pallets.

That mindset is probably one of the deeper reasons why the Fellowship keeps running into the same legitimacy questions despite having some of the best engineering talent in the industry. Exceptional engineers, and best-infrastructure and somehow the social layer still feels like a comittee trying to invent human nature from first principles.

This is not really about blaming any one person, but the influence of certain design philosophies on the culture is impossible to ignore. Over time, a lot of developers and protocolists thinking around Polkadot has inherited the assumption that every problem should be solved through mechanism design, another incentive loop, another elegant economic mechanism that the very least should look smart on paper (far idealists I assume).

Personally, I think that is often the wrong direction.

Not every decision should be turned into a giant economic puzzle. Some things are subjective by nature : promotions, conflicts of interest, treasury judgement, responsibility, trust. These are not problems you solve by adding another weighted voting formula and pretending the dashboard has become morally enlightened.

A body that can fast-track runtime upgrades, influence OpenGov outcomes, manage internal treasury flows, and shape protocol direction cannot rely on unwritten or experimental norms where formal accountability should exist.

Most successful centralized companies with larger market cap and far more public trust than even a decentralized network, do not operate like Byzantine game-theory experiments for a reason. They delegate clearly, assign accountability and make people answer for outcomes. Because when things fail, there are actual people to blame and people expected to take the fall.

Very Futuristic concept, I know right?.

We often confuse decentralization with mechanization. They are not the same thing.

In fact, the more complex the mechanism becomes, the more it tends to favor the people who designed it or the people who have enough free time to study the rulebook like its the bar exam. Complexity gets sold as neutrality, but often it just hides where the real power sits.

Ever saw a Rube Goldberg’s Machine?

That is why things like conflict of interest rules, recusal duties, transparent whitelist reviews and clear upgrade paths matter more than another elegant formula.

Sometimes governance starts looking less like governance and more like a very expensive IQ test for token holders and supposedly ambassadors.

At some point, Ockham’s razor is the better principle: the simplest structure with clear accountability is usually stronger than a hyper-optimized system everyone spends their time trying to arbitrage.

Not every governance problem needs 4D chess.

Sometimes the answer is much simpler: pick responsible people, make them accountable, publish the record and stop pretending every human problem can be solved by turning it into a math competition (especially models that requires human-actors and their instinct based-decisions)

And honestly, as far as I can see and I think many people quietly feel the same that too much of the ecosystem energy has shifted toward figuring out how to extract from the treasury rather than how to create real long-term value and survival amidst the growing competition.

It often feels like the question is no longer “is this useful?” but “can this attract treasury’s funds?”

Somewhere between innovation and governance, we accidentaly built an Olympic sport for proposal writing.

That is really dishonourable. (Isn’t It?)

When a governance body becomes structurally closer to a funding pipeline than a stewardship institution, incentives change. People stop asking what strengthens the protocol and start asking what can be justified well enough to pass.

When incentives become too centered around treasury extraction, quality drops, accountability disappears and governance becomes less about stewardship and more about professional grant farming with better branding

That is also why, after all these years of Polkadot, Parity and W3F still remain the only real centers handling almost everything around the ecosystem, and they do great engineering, sure, but apparently “decentralization” means everyone else just watches professionally. There is barely even a second front, more like a legislative body without an opposition party (very unhealthy), unless we are counting wallets and DApps or their own partners as meaningful institutional decentralization.

And as the Fellowship continues, if it becomes evident that overlapping compensation for the same work is happening, then it proves the exact point: protocol gaming does not happen only in theory. It happens in plain sight, usually with excellent formatting and a very confident forum post.

When the same ecosystem is largely shaped by a small set of institutions, and the same circles are designing the mechanisms, approving the flows and benefiting from the outcomes, it becomes very hard to keep calling the system neutral with a straight face.

At that point it becomes fair to ask: why has the community been so willing to accept every new mechanism and governance design presented to it, instead of first questioning who benefits the most from the complexity itself?

Before asking whether a new mechanism works, maybe we should first ask the much simpler question:

who need it to exist in the first place?

And yes, I know I have not written “here are the fellowship and governance gaps” or “here are my own 10 specific fixes”, nor am I trying to say I know what is best and everyone should simply hear me out, nor am I trying to directly reply to OP’s findings point by point. Their post already does that well by identifying the hidden issues in plain sight.

What I’m pointing at is something more general, the design philosophy underneath most of these Fellowship issues, both the ones already visible and the ones that will likely emerge next. (I bet it soon will)

Because if the underlying mindset stays the same, the same problems will keep reappering under new names, new committees, and new voting formulas.

Specific fixes matter.

But philosophy determines what kind of system keeps getting rebuilt.

Time usually proves which one mattered more

So to OP @Mark65 thank you for saying it openly. A lot of people see these things quietly and hope someday someone will fix them, but very few are willing to put them into words publicly. That at least gives people like me a reason to speak our thoughts too, regardless of whether it will be seen, agreed with or acted upon.

Mark65 · May 6, 2026, 7:25am

On the philosophical reply just above: broadly agree, with one pushback.

Most of what I proposed is actually a move in the direction you describe, not away from it. A register of interests, audited quarterly accounts, named whitelist reviewers with named dissents, an arbitration panel with a real evidentiary standard, a Master Emeritus retirement status. These are boring, century-old institutional practices borrowed from how civil-society organisations have run themselves forever. None of them is a new economic puzzle. The voting cap (Annex C) is the only thing that adds a formula, and it’s there only because the existing formula is what gets abused. Fix one mechanism, retire most of the rest.

On the direction itself you’re right. If the body owns its accountability with simple rules and named people, the elaborate scaffolding around it tends to fall away. Less Rube Goldberg, more boring board minutes. That’s the bet.

On Parity and W3F, since they came up. The engineering is outstanding, that part isn’t in dispute. The uncomfortable bit is that W3F’s own December 2025 grants post describes Parity as “primarily entrusted with the delivery of the product-centric vision of Polkadot”. That sentence, written by W3F, is already a fairly precise description of where most of the weight currently sits. Either we accept that as the operating model and publish it as such, or we commit to a graduation roadmap with milestones and we report against it. Holding both positions at once, broad decentralization in the marketing and a two-org system in the runtime room, gets harder to defend with each cycle.

For what it’s worth, Annex A (conflicts) and Annex H (legal wrapper) are arguably in their interest too. Right now Parity and W3F absorb the reputational cost of every Fellowship structural issue, even when their actual work is excellent. A clean separation would let the engineering speak for itself, without being read through a “is this capture?” lens. That benefits them as much as it benefits the rest of the ecosystem.

Quickly to the earlier reply on mentorship and transparency, since we’re in the same thread. Mentorship is a real gap, worth its own annex K with a mentorship line in the rubric and mentee feedback at retention. On transparency, a script gets you the votes. It doesn’t get you the rationale or the recusals, because neither of those lives on chain. That’s the layer the annex would force the body to actually produce.

And on the closing thanks, no need. The post seems to land not because of how it’s written but because the gap between what the manifesto claims and what the on-chain record shows had become impossible to ignore for anyone reading both. The credit goes to whoever decides to act on it.

jobyreuben · May 7, 2026, 8:29am

I acknowledge the effort, but I fear the problem is deeper

I acknowledge your duty as a fellow contributor trying to make Polkadot better by proposing actual solutions. That deserves respect.

But I fear none of it gets implemented unless you are with Parity Technologies (which I genuinely hope you aren’t), because every proposal around conflict of interest, treasury allocation, power concentration, or who actually holds the steering wheel tends to die somewhere between:

“great point”

and

“we’ll take this into consideration”

That is exactly why your Annex E (Lower the amendment threshold) matters so much to “End the Masters Monopoly Game”.

If constitutional reform itself depends only on the smallest circle deciding it should happen, then reform becomes permission-based rather than governance.

The uncomfortable reality: competing with Parity is treated as a problem

I say that not theoretically, but because I’ve personally felt it and I’ve heard directly from W3F itself, the quiet but very real position of avoiding anything that might have “competing visions” with Parity, even when the contribution objectively improves Polkadot but happens to compete with Parity’s roadmap

And yes, I fully expect another Rube Goldberg governance innovation framework™ as v2.0 beautifully complicated, elegantly branded and somehow designed to preserve the inability of contributors to influence the roadmap of a decentralized network we keep proudly calling decentralized.

Because apparently decentralization means everyone can participate, as long as they don’t touch anything important.

That is also why even something like Annex C (Voting cap) becomes uncomfortable. If concentration of influence did not exists, nobody would even be discussing voting caps in the first place.

Your proposed fixes are great, until they threaten existing power

On your proposed fixes: accountability, transparency , conflict of interest controls. I honestly think that is the best-case scenario.

A public register of interests, recusals (of-course), no double compensation for the same work, that should not even be controversial. That is why Annex A (Conflicts of Interest) is probably the most important annex in the whole package

If that alone becomes politically difficult, it proves the structural issue more than any forum debate ever could.

The same applies to Annex B (Vote transparency) and Annex F (Secretariat and Whitelist procedure). Scripts can show us turnout. They cannot show us who recused, who dissented, who reviewed a white call or why a fast-tracked runtime upgrade was considered safe. That layer of accountability shall not live on-chain or anywhere to query it unless people are forced with serious-repercussions to produce it.

Until of course these fixes stop being in the interest of the organizations currently benefiting from opacity, authority and control.

Then suddenly everything becomes

complex
nuanced
requires further alignment

Conveniently.

My own W3F grant experience made this obvious

I was a grantee of Web3 Foundation and my latest grant proposal was not a small experiment or a random idea, it was a very-serious technical contribution, effectively a successor to one of Parity’s major pallet: pallet-staking

Result:

Silence
Rejection
No Reason
Out of Scope

And before someone says:

“maybe it just wasn’t strong enough”

No, and they usually provide proper reasoning even for much smaller proposals.

When I asked directly in Matrix, it became very clear what the real issue was:

Conflict of Interest

Their own people made it professionally obvious, they would not back something competing with Parity’s roadmap.

Read that again:

Parity’s roadmap. Not Polkadot’s roadmap.

Apparently if innovation originates outside the approved castle walls, it suddenly becomes:

“too complex”

Funny how “too complex” usually translates to:

“not invented here”

Because apparently innovation becomes suspicious the moment it doesn’t originate from the usual approved hands.

Funny how every “complex” problem somehow ends with

“maybe Parity should handle it”

So naturally their conclusion was:

“maybe ask the Fellowship”

And we already know how that loop ends to ask the same disguised landlord for permission to repaint your own house.

That is why Annex A (Conflict of Interest) matters far beyond salary disclosure. Conflict-of-interest rules are not about optics; they are about whether independent technical work can exist without first needing approval from the same people whose roadmap it challenges.

We have built it any way

No funding
No blessing
No ceremonial handshake

Eventually soon the tech will speak louder than this strange political theater around who exactly owns and steers Polkadot.

And honestly, we do not even want the funding anymore.

Because once you realize funding depends less on contribution and more on whether your existence inconveniences their roadmap, it stops looking like support and starts looking like permission.

Compensation with conditions is just governance cosplay

This is why Annex G (Treasury discipline) matters. If treasury authority moves inward while accountability moves outward, then governance becomes a funding gate, not stewardship. Quarterly audited visibility should increase as treasury control becomes more internal, not disappear.

The real problem is not Fellowship Procedure

My fear is for the worst-case scenario:

Fellowship gets a cosmetic fix
everyone claps
a few annexes get renamed
people post “great progress”
everyone shutdowns their Macbook
and calls it a day

… all while the fundamental issue remains untouched.

Because the real problem is not Fellowship procedure.

Its power residence.

Even something like Annex H (Legal wrapper) which is really required, can become dangerous if its only used to give cleaner legal structure to the same concentration of control. A legal wrapper should create accountability and institutional maturity, not just better paperwork for the same centralization.

W3F increasingly feels less neutral

Recent posts (After December 2025) and replies already show growing discomfort with both Web3 Foundation and Parity Technologies, especially since W3F moved from feeling like a neutral ecosystem enabler to what increasingly feels like

“a nonprofit customer support department for a for-profit roadmap”

People wanted to trust W3F because it once felt neutral:

pro-builders
pro-innovation
pro-ecosystem

Now neutrality seems to come with approval requirements.

Even the recent branding tells the story:

Instead of

“lets bring more parachains”

it shifted toward:

“more investor elegance”

“more polished decks”

More:

“look how decentralized we are”

… while the same two organizations still decide what actually matters and with W3F now jumping directly into governance proposals, it feels like your wing-man quietly left and started playing for the other team.

What keeps happening

Better projects get filtered out
Parity continues to define the roadmap
W3F continues funding projects already safe for that roadmap
Governance keeps feeling biased
The Fellowship remains structurally skewed

And we all keep pretending this is fine because the website gradients are getting better.

They made people believe excellence only comes from one place. so concentration of power became easier to excuse.

The moment people see strong engineering can exist outside those castle walls, the illusion will start to break.

That is also why your Annex F (Secretariat and Whitelist procedure) matters more than it first appears. Whitelist is where technical judgement becomes political power. If reviewers are unnamed and dissent is invisible, it stops being governance and starts looking like delegated legitimacy.

Before any annex or procedural fix, I think the real question is simpler

Is the Fellowship actually governing Polkadot’s technical direction?

Or is it mostly legitimizing decisions already shaped by Parity?

If Parity effectively defines the roadmap, then is the Fellowship an independent governance or just a clearer constitutional wrapper around existing power?

And if the goal is real decentralization:

where is the actual roadmap for reducing that dependency?

instead of endlessly formalizing it?

Because without answering that first, even the best fixes risk becoming better paperwork for the same concentration of control.

This is where I think your annex package could go even further, not just procedural correction, but an actual decentralization roadmap (Annex X, if we want symbol it that way towards centralization).

Because otherwise even strong Annex A through H risk becoming accountability theater around a structure whose center of gravity never moved.

Final point

I do not think conflict-of interest rules alone will solve much if the underlying incentive is still ownership and control

When parachains are treated as:

“Parity’s (Our) parachains”