ICP has a visibility problem

Developers Showcase
1

Here are my thoughts on ICP. When I talk about ICP, I talk about visibility.

Technology alone is not enough. A project can have brilliant technology and still lose attention, users, and capital if people don’t understand why it matters.

Look at projects with little real utility trading at 10x ICP’s market cap. Why?

Visibility. Narrative. Distribution.

They know how to reach people and turn attention into adoption. That doesn’t make their technology better. It means they understand the market.

Another project claims 40 million TPS. I can barely reproduce that performance on my local PC, yet they claim to run it on-chain.

The point isn’t whether every claim is true. It’s that bold, simple claims capture attention. ICP needs to communicate its strengths better.

DFINITY wants ICP to become infrastructure running in the background. I understand the vision.

But if ICP becomes invisible, the technology may succeed while the token and ecosystem struggle to attract attention.

Mainstream adoption and visibility must go together.

In 2023, a friend and I started building an application. I subscribed to an Azure database. The portal showed $14.

Ten days later, after creating and dropping the same tables, I received a $250 bill.

That sent me looking for alternatives.

I found ICP, and I was impressed. I built my first application and deployed it on-chain.

I asked my friend to check it out. He refused. Blockchain was full of scammers, he said, and I would lose my money.

I chose my path and kept building on ICP.

Then came UTOPIA, promoted as a way to make ICP known. But it burns no ICP.

Later came Caffeine AI. I hoped these products would help bring users into the ecosystem.

I kept waiting for the missing link between great technology and real ICP demand.

Then I found that Caffeine AI accepted only fiat.

Why not ICP first?

If an ICP-powered product makes it difficult to use ICP, how does that create demand for the token?

Great products should make using ICP natural, not optional or invisible.

OpenCloud was another disappointment for me. Card payment worked initially, but not for subscriptions.

I tried three cards and contacted support to ask if I could pay with ICP. No useful solution.

I eventually renewed with a card that accepts almost everything.

Next: MultiDex.

Instead of rewarding people according to their actual activity, the distribution felt like a tennis tournament: first, second, third—and everyone else gets nothing.

That is a poor way to reward the community helping build an ecosystem.

Giving 7,000 ICP to one winner who may already have sold—and could even be shorting ICP—is not the incentive design I’d choose.

Reward meaningful activity, usage, and long-term participation, not just a few winners.

Then came the idea of buying canister cycles with a card.

Convenient for some, sure. But why hide the ICP step completely?

Let users top up an ICP wallet, then decide whether to mint cycles, transfer ICP, or hold it.

Developers can already mint cycles from ICP.

A simple wallet top-up would preserve choice and create a direct connection between the product and the token.

Make onboarding easier without making ICP itself disappear.

Here’s my main point:

ICP doesn’t only have a technology problem. It has a visibility, distribution, and user-conversion problem.

If people don’t hear about ICP, don’t use ICP, and don’t need ICP, technology alone won’t create token demand.

My advice to DFINITY:

  1. Let users top up with ICP, not only cycles.

  2. Make ICP a first-class payment option in Caffeine AI, OpenCloud, MultiDex, and other ecosystem products.

  3. Launch and market MultiDex properly.

Reward MultiDex users based on meaningful activity, not a winner-takes-all model.

Give ordinary users a reason to participate, return, and contribute.

A broad, active community is more valuable than a small group receiving all the rewards.

Invest in marketing. Work with influencers. Tell clear stories.

Show real applications, real users, and real results.

Crypto is not just technology. Attention, community, and distribution matter.

Ignoring that reality doesn’t make ICP more serious. It makes ICP less visible.

ICP has powerful technology and real builders. But the market won’t reward potential forever.

DFINITY must make ICP visible, usable, and economically relevant—not just technically impressive.

Build the technology. Market the vision. Create real demand for ICP. The time to act is now.

Bonus 1 — Visibility is not hype

I’m not asking DFINITY to abandon technology for hype. I’m asking them to make the technology understandable, demonstrate its value, and get it in front of the people who need it.

Marketing isn’t the enemy of engineering. It’s how engineering reaches the world.

Bonus 2 — The missing connection

Every major ICP ecosystem product should answer three questions:

Who will use it?
Why will they choose it over alternatives?
How does its success benefit the ICP ecosystem and token?

If we cannot answer all three, we’re missing an important part of the strategy.

Bonus 3 — My challenge to DFINITY

Stop measuring success only by what ICP can do.

Show us how many people use it, how many developers stay, how many applications grow, and how much real demand the ecosystem creates.

Technology creates the opportunity. Adoption turns it into success.

2

Is this how 90% of newcomers to the forum introduce themselves (incl. me) :joy:

Bless you :fire::heart:

Have you seen the Jerry Banfield thread?

3

Give me where I am wrong, and I will remove the whole thing :slight_smile:

4

That’s quite a list, mate.

But it’s still not clear to us,

exactly what you want everybody to do.

We don’t think anybody said that.

If you want everybody to read

keep it short.

get to the point, mate.

If it’s too long nobody will read it.

5

The visibility just isn’t there.
@Dionysus needs to do better

6

That’s interesting.

don’t tell anybody, but we’re only half.

We really try our best, mate.

7

Not wrong, just a flag that gets waved and set on fire quite regularly. Water is wet, cheese goes well on crackers, ICP doesn’t get the attention it deserves, bears and Pope hats, you know what I mean

8

This is the real problem. DFINITY can only do so much, and what they have done is impressive. It’s the community that’s so far been letting the ecosystem down.

Most people stake yet pay no attention to active and diligent governance, most builders seem to focus on end products without taking time to build out missing pieces of infrastructure first (not all, but most), and many community members expect DFINITY to do everything so they grant them that power (forgetting that it defies the whole point of the network)

9

That’s certainly true. We’re all only human.

Are you suggesting that most people actually don’t want freedom, and even crave authority? Perhaps many often seek authority or blindly follow it, to avoid the responsibility of self mastery.

10

Is not this simpler? Is this not your natural state? It’s the unspoken truth of humanity, that you crave subjugation. The bright lure of freedom diminishes your life’s joy in a mad scramble for power, for identity. You were made to be ruled. In the end, you will always kneel

11

12

Out of curiosity, what are these missing pieces? Speed, connectivity, storage, security all seem fine

13
  • Event-driven architecture infrastructure (so that dapps no longer need to poll ledgers, which is slow and/or expensive, and/or requires a clunky UX)
    • I’ve spent the last few weeks working on this, and it seems crazy to me that we don’t already have this. I’ll launch a solution in a couple of weeks after a few ledger and SNS framework pull requests.
  • Trustless automated cycles topups
    • I’ve been working on this for several months and it’s almost mature enough to blackhole
  • Governance rewards for top followed neurons (to reword actual work rather than just finger bashing adopt/reject)
    • I’ve been working on bits and bobs towards this for several years now, but it will undoubtedly take several years more to fully address
  • Browser extensions that protect users from DNS repointing attacks
    • and/or putting extra effort into dapp design to limit susceptibility
  • Automated node swaps using an onchain optimiser to avoid so many unnecessary NNS proposals
    • worked on this for a bit and will return to finishing it at some point (it’s a lot of work)
  • key resharing that doesn’t require an entire subnet to be halted (taking everything on that subnet down for as long as it takes for someone to force through a subnet restart proposal - it shouldn’t be possible to force such rapid adoptions)

I could keep going…

14

Instead large sums have been raised over the years for people to develop the most ridiculous things, because everyone gets sold on an end product idea.

15

Is this an accurate picture of what you referred to (AI interpretation)

It aims to solve a real pain point (dapps repeatedly polling ledgers to detect changes are slow, cost cycles, and create laggy UX), but current ICP reality is still a mix of “timers + polling” and emerging event standards.

Why polling is the problem

On ICP, if a dapp wants to know when a user receives tokens or a ledger balance changes, the simplest approach today is for the canister to set a timer and repeatedly call get_transactions or query logs. This consumes cycles on every poll, and if the polling interval is short (e.g., a few seconds), costs accumulate quickly; if the interval is long, the user sees stale data and the UX suffers . A developer forum discussion explicitly notes: “I wouldn’t want to run the indexer on-chain until there is a good mechanism to avoid polling” .

What “event-driven infrastructure” should mean

Ideally, ledgers/index canisters would actively push events to subscribed dapps, rather than dapps asking “has anything changed” every few seconds. This requires standardized subscription, broadcasting, and event formats—similar to Webhooks or SSE.

Currently, ICP has an ICRC-72 standard in alpha exploring this direction. A community-developed event system uses the ICRC-72 standard: canisters can register as subscribers, broadcasters push events to subscribers, and subscribers can return cycles to offset the broadcaster’s costs . This is a meaningful exploration toward “event-driven,” but it is not yet a native, stable feature provided by the ledger.

Practical implications for interpretation

· Direction is correct, but don’t treat it as an out-of-the-box capability: If you are evaluating whether to build on ICP based on this, you need to know that today you may still need to write timer polling yourself, or rely on third-party index services.
· Cost model is key: The core value of event-driven architecture is eliminating invalid polling. On ICP, because canisters pay for their own computation (reverse gas), the cost of polling is borne by the dapp developer; event push can theoretically reduce costs and improve response speed simultaneously .
· Watch ICRC-72 and community event systems: If this infrastructure matures, dapps will no longer need to maintain their own polling logic, and UX will approach Web2 real-time experiences. But currently, this is more of an infrastructure direction that is “being built,” rather than a reality that “has been solved.”

In short: the problem is real, the direction is clear, but it’s still in the infrastructure-building phase—evaluating ICP dapps should not assume event push already exists.

16

Indeed.

I’d push back on this. Ledgers and index canisters should keep their single responsibility. A publisher canister is needed inbetween the ledger and index. The index canister itself becomes a subscriber to the publisher and then only uses polling as a fallback.

I have a pull request that’s currently being reviewed by DFINITY for the minor change needed to ICRC index canisters.

The ICRC label makes this sound more official than it should do. I don’t think it stands up to critical review. What’s being referred to is this → https://forum.dfinity.org/t/icaibus-internet-computer-ai-bus-pubsub-made-easy-on-the-ic/40606/34

17

As a paying customer, I agree with most of the points. Lots of stuff just do not work. And can not get real support. Support case does not get closed and have no SLA. Feel like throwing money into the vacuum. And we are glad to pay DFINITY support fee (reasonable amount). If things are really working (hard token like Yubikey, OpenCloud etc), why not have enterprise customers (paying customers)? dada.clouds@gmail.com

18