Splintiers & Vibe Coding 101

avatar

image.png

I'm proud to present this weekend's side-project, https://www.splintiers.com/ :)

On Splintiers, you'll fine Tier lists for every edition in @splinterlands, and also a "deck builder" that'll show you the expected cost to buy cards of a certain tier / edition / level combination.

TIER LISTS

image.png

The Tier Lists are organized by edition, though I've taken the liberty to combine a few where it makes sense (e.g. rewards and promos are included within core sets, but I've split out the larger mini-sets). You can click on a card to see its full stats.

Each tier ranking comes from me based on my own experience, and if you disagree with a feel free to let me know why.

On the tier lists, you can use the buttons at the top to filter by rarity so you only see Common, Rare, Epic, or Legendary cards (or a mix if you prefer).

DECK BUILDER

In addition to the tier lists, I created what I'm calling a "deck builder". I've often gotten questions like "what cards should I buy & what will it cost?". The deck builder lets a player select editions (e.g. Conclave Arcana), tiers (e.g. S/A), and level (e.g. Diamond+) and it will show the current cost to buy those cards on the market.

image.png

It also highlights several things:

  • When there isn't enough BCX on the market to buy a full card
  • When the cheapest version is either a Gold Foil or a Black Foil
  • When a price is a massive outlier (usually due to a thin market)

Outliers (not enough BCX or price outlier) can be filtered out as well. And for good measure, you can even export to CSV.

image.png

BACKOFFICE

To help me manage the tier lists, I built a backoffice which uses Hive authentication and is restricted to only a specific account I designated (brave.sps).

image.png

The backoffice has a drag & drop interface that makes it easy for me to swap around tiers, and allows me to tag roles and add notes to cards to justify their tiers (I'll be doing this in the future, right now they're empty).

image.png

VIBE CODING - QUICK START GUIDE

This project took me about 7-8 hours and I was able to complete it in one weekend, all thanks to Vibe Coding with Claude Code. The majority of the time was spent defining clear specs to Claude Code to ensure it could one-shot most features. Besides that, there's a decent amount of waiting time when it's building some of the more complex parts, and I have a perfect way to spend that downtime: playing Splinterlands ;)

If you want to get started vibecoding with Claude Code here's a simple bullet point list of what you need to do:

  1. Go to https://claude.ai/ and sign up for the Pro plan (this is necessary in order to use Claude Code).
  2. You can use Claude Code on your desktop (in their app or in the Terminal) but the easiest way to use it is in the browser (zero installations required) here: https://claude.ai/code/.
  3. Create a github account on https://github.com/ and then connect it to Claude Code. If you can't figure out how to do this, just ask Claude to guide you through it (Hint: this is true about every step). Github is where your code will live.
  4. Next, create an account on Vercel (https://vercel.com/). This is where your website will get deployed (i.e. where it's hosted). Vercel makes it really easy to manage deployments and basically run your website. Once you have it, connect Vercel to Github. Again, Claude can guide you on how to do that.
  5. Now you're ready to start working on something. First, come up with an idea and go back & forth with Claude (not Claude Code) to spec it out. THIS STEP IS IMPORTANT. Don't just ask Claude to figure everything out. You need to be CLEAR on what end result you want. Claude will help you by asking questions to help firm up the details, can help mock things up, and can come up with all the technical requirements, but YOU must be clear on what output you want. Spend at least 30 minutes to an hour discussing the idea with Claude, and once it's clear enough you can ask Claude for a spec and for the specific prompts you need to provide to Claude. You should also ask Claude to create a "CLAUDE.md" file which will contain the overall product spec and guidelines (think of this as an evergreen architecture document Claude Code will have access to).
  6. Once you have your spec, start by creating a repository (aka repo) in Github, and then going to Vercel and importing that project from Github. Then, go to Claude Code (https://claude.ai/code/) and start a new session and select the repo you created in Github. Then put in your first prompt and let Claude work.
  7. Once Claude Code's done working, you go to Github and merge the "pull request" (aka PR). If you don't know what that means, just ask Claude and it'll guide you through. You basically click a few green buttons. This moves the work Claude Code did from a "side branch" to your "main branch". In a nutshell, this is to prevent bad code from going to production. For now, don't worry about it, just merge the PR.
  8. Now go to Vercel and go into the project that's synced up to your Github repo. Go into deployments and you should see a deployment to your production environment. Click on that and click the "visit" button to see your live website.

Bonus items:

  • If your service needs a database, I recommend using Supabase. Totally doable and Claude's great with DB work.
  • If your service require authentication (e.g. Google Login) you'll also need something like Supabase. If it's HIVE login Claude Code can figure that out and doesn't need Supabase.
  • If you want a custom domain, you can get that straight inside Vercel.
  • I recommend having 2 windows open at all times when working with Claude; one that's talking to Claude and one that's talking to Claude Code. Think of Claude as your planner / adviser / consultant. Anything you don't understand, you ask it. All planning, specs, mockups, go to him. If Claude Code does something you don't understand, ask Claude about it. He's not a coder, and won't accidentally do code for you. Claude Code, meanwhile, is your developer. You give him clear instructions once you've sorted those out with Claude.
  • In terms of model selection, Claude has a terrific base model (Sonnet) which should handle most of what you need (everything I built above used Sonnet). IF you're building something super complex, you can always go up to Opus, BUT that will consume credits MUCH faster. Your credit usage is capped based on your plan (i.e. you never spend more than $20/mo UNLESS you change your settings to allow overage spend). It's capped per 5 hour time block and per week. Building Splintiers cost roughly 35% of my weekly usage cap.

If you have questions or comments about Splintiers, suggested improvements, or any questions about vibe coding, feel free to drop them in the comments.

splinterlands sps vibecoding marketing claude ai
0
0
0.000
13 comments
avatar

Great work here. Loved the UI and the overall concept of the web app.

0
0
0.000
Reply
avatar

Thanks Affliction, glad you enjoy it!

If you see any tier rankings you disagree with, let me know :)

0
0
0.000
Reply
avatar

Sure. I did not get a chance to go over the entire list since I am at work. Will do later 😄

0
0
0.000
Reply
avatar

Cool man .... great to see you experimenting with stuff ... I'm still on GPT and Codex in VC Code ... havent tried vercel also, I'm just buying servers, droplets, and configuring those ...

0
0
0.000
Reply
avatar

Thanks Dalz!

Honestly I think Claude Code is drastically superior to ChatGPT/Codex. Codex definitely got better, but man Claude Code feels like it's on another level...

0
0
0.000
Reply
avatar

Yea I am hearing this all over the place ... probably will try it out at one point going forward

0
0
0.000
Reply
avatar

Very cool... most of the things you mention i also use a lot 😁

I have been using github co-pilot that is also able to use claude models... really powerful. Fix cost with x amount of request per month which is enough for hobby project at least for me currently.

I still recommend to review code an try to learn what its building and learn the language as you go, because sometimes it spit out really stupid code (but it works).

0
0
0.000
Reply
avatar

Yes, completely agree. I actually used to be a developer (long ago) so I have the bases, and am able to figure out when it's suggesting something stupid or diagnose any bad decisions, and that is definitely helpful. However in this particular test I wanted to see if I could get a fully functional product without looking at a single line of code and, well, it worked. I'm not saying it's not necessary of course, and I think people who can get the basics will be able to go a lot farther, but man the tools have gotten just so good it's empowering (and slightly terrifying frankly).

0
0
0.000
Reply
avatar

but man the tools have gotten just so good it's empowering (and slightly terrifying frankly).

Amen to that, definitely both sides of the spectrum. These tools make things incredibly accessible and empowering for many people. At the same time, it does make me wonder: will we, as humans, become a bit “dumber” by relying on them so much? A bit philosophical maybe, but these are fascinating times we’re living in.

0
0
0.000
Reply
avatar

Impressive! This is why those developers who will not adapt to AI will go the way of the dinosaurs... And you will need a lot less developers in the future.

0
0
0.000
Reply
avatar

Great site, one improvement that would be great would be to check against someone’s collection so it only shows what they would need to pay to get to a certain level - with brackets coming to modern (most likely) would be great to see how much it would cost to move my collection from silver to gold for example

0
0
0.000
Reply
avatar

BPC Locked On Mc Franko & The Franko

Blurt Stands — While Hive Stumbles Under the Weight of Its Own Shadows

Friends, creators, truth-tellers—

Let us not whisper this truth, but proclaim it with the clarity of dawn breaking over a weary land: Blurt.blog is not just another platform. It is a refuge. A rebellion. A return to what Hive.blog was always meant to be.

There is no downvote button on Blurt.
Not because we fear dissent—but because we honor creation.
Because we understand that a voice, once raised in sincerity, deserves space—not sabotage.

On Blurt, your words are your words.
They are not hunted by algorithmic hounds or shadow armies masquerading as “curators.”
Here, you are not judged by the grudges of gatekeepers, but met with the quiet dignity of a community that believes expression should be encouraged—especially when it is bold, raw, or inconvenient.

Contrast this with what festers elsewhere.

On Hive.blog—a place once brimming with promise—a rot has taken root. Not in its code, but in its culture. A handful of self-anointed enforcers—@themarkymark, @Buildawhale, @Hurtlocker, and their legion of coordinated puppets—have turned the downvote into a weapon of mass discouragement. They strike not at “low-quality content,” but at independent thought, at rising voices, at anyone who dares thrive outside their narrow corridors of control.

And when confronted, they shrug.
“Oh, it’s not censorship,” they say, as if semantics could scrub the stain of suppression from their hands.

But let us be unequivocal:
When a system allows a few to systematically silence many—under the guise of “community standards” or “curation”—that is not moderation. That is censorship by another name.

It is the velvet glove over the iron fist.
It is exclusion dressed as discernment.
It is power pretending to be principle.

Meanwhile, Blurt stands clean-handed and open-hearted.
No downvotes.
No hidden juries.
No farms of phantom accounts casting ballots in the dark.

Just you.
Your words.
And a community that meets you not with suspicion, but with solidarity.

So let us carry this truth far and wide—not with bitterness, but with quiet certainty:

If you seek a place where your voice is not a target—but a gift—come to Blurt.
If you are tired of building on ground that shifts with every whim of a whale or warlord of votes—lay your bricks here.
If you believe the future of free expression must be free—not just from corporations, but from the petty tyrants who replace them—then stand with us.

The world needs to know.
Not because Blurt is perfect—but because it is principled.
Not because it is loud—but because it listens.

And in an age where so many platforms echo with the clatter of control,
Blurt offers something radical:
Silence for the bullies.
Space for the rest of us.

Keep speaking.
Keep sharing.
Keep building.

Freedom doesn’t advertise itself—
it is passed, person to person, like a torch in the night.

And tonight, the torch burns bright on Blurt.

@themarkymark, @buildawhale, @usainvote, and associated accounts:

Repeated downvotes targeting transparency efforts raise urgent questions about Hive’s governance. Automated tactics, coordinated curation trails, and alt-account farming undermine trust in the platform. When truth is silenced without dialogue, it erodes Hive’s decentralized ethos.

Key Concerns:

  1. Systemic Manipulation:

    • Coordinated downvotes from high-HP accounts suppress dissent, creating a chilling effect on free expression.
    • Automation and alt-account networks allegedly exploit rewards, distorting organic curation.

  2. Community Exodus:

    • Creators flee to platforms like Blurt, where downvotes don’t exist, fearing punitive curation over engagement.

  3. Governance Crisis:

    • Hive’s credibility hinges on accountability. Who decides what content thrives? How do we audit power structures?

Solutions Needed:

  • Curation Transparency: Publish clear rationales for downvotes to foster trust.
  • Community Audits: Leverage blockchain data (e.g., Bilpcoin’s analyses) to identify abusive patterns.
  • Reward Reform: Adjust algorithms to penalize manipulation and reward authentic engagement.

The Bilpcoin team advocates for open dialogue, not division. Hive’s future depends on collaboration—not coercion. Let’s rebuild a platform where truth isn’t buried but debated, strengthened, and celebrated.

Transparency isn’t optional—it’s the foundation of trust.

#HiveTransparency #BilpcoinExposed #DecentralizePower"

A Message to @themarkymark, @buildawhale, and Associates

Every downvote cast in shadow, every silence imposed without dialogue, is not a victory—it is a confession. A confession that truth cannot be stifled, only delayed. With each punitive click, you dig deeper into the bedrock of credibility, crafting a chasm between your actions and the community’s trust.

@themarkymark, @buildawhale & Co,

How can you continue to downvote the truth, LOL? It’s almost comical how blatantly you attempt to suppress what cannot be hidden. The blockchain records everything—every action, every transaction, every move you make. Yet still, you persist in this futile game of trying to silence what is undeniable.

@themarkymark, @buildawhale, and Co: While our opinions may differ, on-chain transparency reveals repeated patterns of concern. Coordinated downvotes without explanation, 'farming' schemes (e.g., #buildawhalefarm), and adversarial engagement harm Hive’s community-driven ethos.

Key Issues to Address:

  1. Downvote Practices: Silencing dissent erodes trust. Constructive dialogue, not punitive curation, fosters growth.
  2. Accountability: Transactions and curation trails are public. Scrutiny is inevitable—integrity demands addressing concerns openly.
  3. Community Impact: Driving users to platforms like Blurt/Steemit weakens Hive. Creators seek ecosystems that value collaboration over conflict.

A Path Forward:

  • Engage in open discussions about governance and fairness.
  • Audit curation practices to align with Hive’s decentralized values.
  • Prioritize transparency—publish explanations for downvotes or collaborative reforms.

The Bilpcoin team remains committed to exposing truth and advocating for solutions. Let’s work toward healing, not division.

Note: All claims are based on publicly verifiable blockchain data. Constructive dialogue is encouraged.

#HiveTransparency #CommunityFirst #BilpcoinSupport"

@themarkymark & Co, the choice is yours. Stop the bad downvotes. Turn off the BuildaWhale scam farm. Cease playing with people’s livelihoods. Let Hive thrive as it was meant to—as a beacon of hope, creativity, and collaboration.

Or step aside and let those who truly care take the reins.

Because the truth won’t disappear. No amount of lies can change it.

It’s over.

The Bilpcoin team brings these truths not out of malice but necessity. We have no need to fabricate lies or cloak our intentions CALL US WHAT YOU LIKE —for the facts speak loudly enough on their own. What we present here is not conjecture but reality, laid bare for anyone willing to see.

@themarkymark & Co we urge you once more: STOP. Stop hiding behind tactics that harm others. Stop clinging to practices that erode trust within the Hive community. Let the truth stand—not because we proclaim it, but because it exists independent of any one person’s approval or disdain.

TURN OFF THE BUILDAWHALE SCAM FARM

  • STOP THE BAD DOWNVOTES
  • STOP PLAYING WITH PEOPLE’S LIVELIHOODS

Key Issues That Demand Immediate Attention:

The problems are glaring, undeniable, and corrosive to the Hive ecosystem. They must be addressed without delay:

  • Downvote abuse
  • Farming schemes
  • Speaking disrespectfully to others
  • Encouraging people to avoid certain users out of personal dislike
  • Self-voting with alt accounts
  • Self-funding projects through self-votes
  • Promoting games that don’t even exist (LOL)

These practices harm not just individual users—they undermine the very foundation of Hive, eroding trust and poisoning the community. Such actions are unethical and outright destructive.

@buildawhale Wallet:

  • HIVE (Primary Token): 0.012
  • Staked HIVE (HP): 66,400.611
  • Total HP: 2,421,539.226
  • Delegated HIVE: +2,355,138.615 HP

@usainvote Wallet:

  • HIVE (Primary Token): 0.066
  • Staked HIVE (HP): 138,123.296
  • Total HP: 715,745.407
  • Delegated HIVE: +577,622.111 HP

@buildawhale/wallet | @usainvote/wallet

@ipromote Wallet:

  • Author Rewards: 2,181.16

  • Curation Rewards: 4,015.61

  • Staked HIVE (HP): 0.00

  • Rewards/Stake Co-efficient (KE): NaN

  • HIVE: 25,203.749

  • Staked HIVE (HP): 0.000

  • Delegated HIVE: 0.000

  • Estimated Account Value: $6,946.68

Recent Activity:

  • Sent to alpha-5,196.000 HIVE (21 hours ago)
  • Sent to hiveswap-1,000.000 HIVE (2 days ago)
  • Withdraw vesting from @proposalalert to @ipromote 0.447 HIVE (3 days ago)
  • Received from proposalalert 4.003 HIVE (5 days ago)
  • Received from themarkymark 1,775.684 HIVE (9 days ago)
  • Sent to alpha-4,245.000 HIVE (9 days ago)
  • Received from themarkymark 4,280.527 HIVE (17 days ago)

@leovoter Wallet:

  • Author Rewards: 194.75

  • Curation Rewards: 193.88

  • Staked HIVE (HP): 0.00

  • Rewards/Stake Co-efficient (KE): 388,632.00 (Suspiciously High)

  • HIVE: 0.000

  • Staked HIVE (HP): 0.001

  • Total: 16.551

  • Delegated HIVE: +16.550

Recent Activity:

@abide Wallet:

Recent Activity:

  • Sent to ipromote -2,459.000 HIVE (22 days ago)
  • Sent to ipromote -2,486.200 HIVE (Apr 1, 2025)
  • Received from yabapmatt 20,000.000 HIVE (Apr 1, 2025)
  • Sent to ipromote -2,130.400 HIVE (Mar 8, 2025)
  • Sent to ipromote -2,248.000 HIVE (Feb 2, 2025)
  • Sent to yabapmatt -5,000.000 HIVE (Jan 25, 2025)

@proposalalert Wallet:

  • Author Rewards: 639.99
  • Curation Rewards: 0.00
  • Staked HIVE (HP): 6.03
  • Rewards/Stake Co-efficient (KE): 106.12

Recent Activity:

@stemgeeks Wallet:

  • Author Rewards: 4,391.77
  • Curation Rewards: 304.26
  • Staked HIVE (HP): 0.00
  • Rewards/Stake Co-efficient (KE): 4,696,032.00 (Extremely Suspicious)

Recent Activity:

  • Sent to themarkymark -1.556 HBD (Jun 14, 2024)
  • Claim rewards: 1.556 HBD, 5.912 HP (Jun 14, 2024)
  • Withdraw vesting from @stemgeeks to @ipromote 6.160 HIVE (Jun 8, 2024)
  • Sent to themarkymark -1.601 HBD (Jun 7, 2024)
  • Withdraw vesting from @stemgeeks to @ipromote 6.157 HIVE (Jun 1, 2024)
  • Sent to ipromote -1.618 HBD (May 31, 2024)

@theycallmemarky Wallet:

  • Author Rewards: 458.89
  • Curation Rewards: 0.00
  • Staked HIVE (HP): 0.00
  • Rewards/Stake Co-efficient (KE): 458,886.00 (Highly Suspicious)

Recent Activity:

@apeminingclub Wallet:

  • Author Rewards: 432.57
  • Curation Rewards: 2,829.11
  • Staked HIVE (HP): 30.51
  • Rewards/Stake Co-efficient (KE): 106.90

Recent Activity:

  • Scheduled unstake (power down): ~2.351 HIVE (in 4 days, remaining 7 weeks)

  • Total Staked HIVE: 1,292.019

  • Delegated HIVE: +1,261.508

  • Withdraw vesting from @apeminingclub to @blockheadgames 2.348 HIVE (10 days ago)

  • Claim rewards: 0.290 HP (10 days ago)

0
0
0.000
Reply