REP v3, October is done.
15 Guardians earned rewards this month, 2,991 SOMI in total, sent today. Top 3: @ShinyViq, @Bhupixhot, @vasyl_khlibko. Congrats, and thank you to everyone who showed up for Somnia this month.
A special thank you to @Plzgivemehopium, @vothaithinh102, @0xtranh, @ArunikaLiem, @MogiwaraD, @0x_Empire and @djaikoku. You posted like yourselves: your own trades, your own takes, your own memes. That is exactly what the rebalance was built to reward.
October was the first month under it. Human posting patterns got a boost, bot-like pace and AI-written replies got a reduction, and the board shows the difference. The same rules carry into November.
Everything that changed is on the rules page: mammothtm.github.io/somnia-r…
Thank you all for building this with me.
How do you feel about it? Tell me in the replies. 👇
REP v3 update.
You may have noticed I’ve been quiet. I’ve been collecting data on how everyone posts, and I ran a full analysis.
What I found: some accounts post at a pace no person writes at, hundreds of replies a day and dozens in a single hour, much of it AI-written. Others post less, but it’s clearly them: their own trades, their own takes, their own memes.
So instead of removing anyone, I’m rebalancing REP, starting with October:
- Human posting patterns get a boost.
- Bot-like pace and AI-written replies get a reduction.
- Posts showing your own use (your PnL, your rank, your quests) and your own memes now score fairly instead of being marked as thin.
The leaderboard looks very different now. That’s the point: REP is recognition for real contribution, not volume.
Also: quote @0xPaulThomas’s Rekt Replay post with your own Rekt Replay, and that quote post gets a 1.4x REP boost. One per Guardian, until Wednesday 28 Oct.
Thank you to everyone who keeps showing up as themselves 🫡🫡
REP v3, September is done.
23 Guardians earned rewards this month, 2,986 SOMI in total, sent today. Top 3: @thomasthomnb, @bhupixhot, @jundohud. Congrats, and thank you to everyone who showed up for Somnia this month.
A special thank you to @censarboyzzz, @vasyl_khlibko, @ShinyViq, @eleremen7, @0x_Empire and @Trong_Hatachi for the original content this month. Your own takes, your own research, real criticism where it was due. That is exactly what REP is for.
A reminder of what REP is: recognition for contributing. It is not farm to earn, and posting volume for points is not something we will encourage.
Replies matter, and most of this month's points came from them. But a hundred quick replies in a day is not a hundred contributions. So from 1 October, only your best 20 replies each day count toward your score. Reply as much as you like; the extra ones just won't add points. Original posts are where the effort should go.
And thank you all for building this with me. I know it was frustrating at times, with fixes and adjustments as we went, but it's slowly coming together.
How do you feel about it? Tell me in the replies. 👇
REP v3 update. The new scoring is live since 26 September.
Thanks to everyone who replied to the preview. Most of you were for it, so I switched it on and scored all of September again under it.
Reminder of how it works: you're scored on what you add on top of what Somnia already said. Your own use, your own numbers, a real opinion. Rewording an announcement scores low.
Since every score is lower now, the minimum to qualify is 8 REP instead of 10.
The board already shows the new numbers. September closes on the 30th, and rewards go out after that.
How it works: mammothtm.github.io/somnia-r…
Before and after for the top 25: mammothtm.github.io/somnia-r…
How is it feeling so far? Tell me in the replies.
REP v3 update. A lot of you said the board was filling up with AI-written posts. I agree.
I scored every September post a second time under a new scoring. It reads what Somnia's official accounts posted before your post, and scores you on what you add on top.
Rewording an announcement scores low, however polished. Your own use, your own numbers, a claim you stand behind, a specific criticism: that scores well.
Before you look: almost everyone's number goes down, because the new scoring is stricter for everybody. That is not you losing anything. The pool is shared, so what counts is your place on the board, not the size of your number. Plenty of people move up.
It is not AI detection. Nobody is accused of anything. It stops giving credit for the part a tool can do for you.
Nothing is live yet. Here is how the top 25 would move, with the reason for each: mammothtm.github.io/somnia-r…
Read it, then tell me in the replies: how do you feel about this change? I am reading everything before I decide.
REP v3 update. A lot of you said the board was filling up with AI-written posts. I agree.
I scored every September post a second time under a new scoring. It reads what Somnia's official accounts posted before your post, and scores you on what you add on top.
Rewording an announcement scores low, however polished. Your own use, your own numbers, a claim you stand behind, a specific criticism: that scores well.
Before you look: almost everyone's number goes down, because the new scoring is stricter for everybody. That is not you losing anything. The pool is shared, so what counts is your place on the board, not the size of your number. Plenty of people move up.
It is not AI detection. Nobody is accused of anything. It stops giving credit for the part a tool can do for you.
Nothing is live yet. Here is how the top 25 would move, with the reason for each: mammothtm.github.io/somnia-r…
Read it, then tell me in the replies: how do you feel about this change? I am reading everything before I decide.
Rescoring the whole month with a new scoring system. People are complaining about AI slop, and as I said earlier this week, I have been looking into it.
Patience is key, but I also know people will be angry. Community members who are happy with the change are welcome to answer those who are not. I won't play referee here.
What did I do here:
I looked at why generated posts were scoring so well. The answer was the scoring itself. It rewarded clean writing, good structure, and showing that you had read about a feature. An AI does all three for free. Hand it a Somnia announcement and it writes a polished, confident explainer in seconds, and the old scoring called that "substantive" and paid it top marks.
The new scoring gives the judge what Somnia itself has posted recently, and then scores your post on what it adds beyond that.
Rewording an announcement, however polished, however long, with however nice an infographic, scores low under it. It names the official post you restated.
What scores well: something you actually did or found out. A trade you made, your own screenshot, your own numbers, a bug you hit, a prediction you are staking, a criticism with specifics. If the only thing you had to do to write the post was read the announcement, it is a restatement.
This is not AI detection. I am not guessing who used a tool, and I use one myself to tidy my posts. It simply stops paying for the part a tool can do for you.
I am running it in parallel with the current scoring, over every September post, so I can review it first. The board you see today has not changed. Once I have compared the two, I will decide and tell you.
Rescoring the whole month with a new scoring system. People are complaining about AI slop, and as I said earlier this week, I have been looking into it.
Patience is key, but I also know people will be angry. Community members who are happy with the change are welcome to answer those who are not. I won't play referee here.
What did I do here:
I looked at why generated posts were scoring so well. The answer was the scoring itself. It rewarded clean writing, good structure, and showing that you had read about a feature. An AI does all three for free. Hand it a Somnia announcement and it writes a polished, confident explainer in seconds, and the old scoring called that "substantive" and paid it top marks.
The new scoring gives the judge what Somnia itself has posted recently, and then scores your post on what it adds beyond that.
Rewording an announcement, however polished, however long, with however nice an infographic, scores low under it. It names the official post you restated.
What scores well: something you actually did or found out. A trade you made, your own screenshot, your own numbers, a bug you hit, a prediction you are staking, a criticism with specifics. If the only thing you had to do to write the post was read the announcement, it is a restatement.
This is not AI detection. I am not guessing who used a tool, and I use one myself to tidy my posts. It simply stops paying for the part a tool can do for you.
I am running it in parallel with the current scoring, over every September post, so I can review it first. The board you see today has not changed. Once I have compared the two, I will decide and tell you.
gm.
A declaration on REP v3.
First: this is my own product, built for the community, and I run it. The rules and the outcomes are mine to decide. There is no court here and nobody has a right to a payout. If you get caught in a loop and I fix the loop, be annoyed for a minute, then move on. I am not going to argue every rule with every person. Call it a dictatorship if you like. It is.
Second, why v3 exists at all. I did not build another farm-to-earn platform, and this is not your farming ground. It is a contribution platform. People post about Somnia because they want it to win, and I wanted a way to see that and give something back for it. Something symbolic, from the community's side, not a salary. The pool is 3,000 SOMI a month. If the program settles, the gaming stops, and what comes out the other end is real contributors, I will think about growing it. That is the order. Contribution first, budget second, never the other way around.
If that direction does not match your expectations, I am genuinely sorry 🙏 Feel free not to participate. No hard feelings either way.
⮑ Now, back to the weekend.
I went through the data and the feedback. Two of my rules were wrong. Half of all replies were Guardians replying to each other, counting as Somnia content only because they inherited the tags from the post above, and each one paid twice. A few accounts were posting 70+ generated images a day, one a minute.
Fixed, for everyone, applied to the whole of September:
- A reply counts only if you reference Somnia yourself.
- Guardians no longer raise each other's engagement score.
- A post counts only if it is 15 minutes after your previous one. A burst counts as one.
The rules page is updated. The board is reordered. Nobody was singled out.
Thank you to everyone who sent feedback. It worked. Keep it coming.
REP v2 was manual, a bit clunky, and people disliked it to some extent. Fair enough.
Well, today it ends. I built REP v3, and it comes directly from X.
Thank you to @ShinyViq for giving me the skeleton and a bit of the meat, and to @ErnstKummer for helping me on all things LLM and scoring.
So post about Somnia, the first-party products, and the official accounts. It gets picked up, scored, and on the platform the same day. Nothing to submit.
A model reads the post and rates the thought in it, that's 85% of your score. Engagement is the other 15%, counted as your rank rather than a raw number, so one viral post doesn't bend the whole board.
It never sees your name, your followers or your likes. Just the words. A 50k account and a 200 account writing the same thing score the same. That's the part I cared most about getting right.
I'll be distributing 3,000 SOMI a month, shared across everyone who qualifies. Your score starts fresh on the 1st and builds as you post. September is the first month that pays.
It is right here: guardian.somnia.builders
Full rules: mammothtm.github.io/somnia-r…
Disclaimer: This is a community-built tool created by an independent member of the Somnia community. It is not affiliated with, endorsed by, sponsored by, or officially associated with Somnia or its affiliated entities. All outputs are generated independently and are provided "as is" without warranties.
I live on both sides of this. I FUD sometimes, I write constructive criticism too. I'm a crypto degen, so I know exactly where the line sits.
Constructive criticism comes from someone who wants the thing to work:
→ assumes good faith
→ specific about the actual problem
→ leaves room for a response
→ built on trust, invested in a better outcome
Even "positive frustration," the annoyed-but-still-rooting-for-you tone, sits here. It stings, but it's a gift. It tells you what to fix, and it comes from someone who still cares.
FUD is different in intent. It's not there to make things better, it's there for a reaction:
→ personal jabs instead of the issue
→ one-liners built for screenshots
→ trolling for its own sake
→ no actual ask, no outcome it's chasing
Criticism wants a response. FUD wants a rise.
The hard part: when someone loses all hope in a project, their emotions bottom out and criticism curdles into FUD. It stops being fair. And it's hard to bring optimism back to someone who's never optimistic, especially when 90% of this market is dumps and rugs. I get why people get there.
I'm saying this because I've banned a few people I respect. I'd love for them to be here, it's the tone that needs adjusting, not their presence. I'm open to reversing a ban if I see the person has hope again, or wants to be around for where this is going.
And respect to the ones who go quiet when things aren't going their way, more than to the ones who stick around only to talk the project down and prove their take was right. I don't want pep-talk merchants or engagement farmers either. I want good feedback.
A few do it really well. Nikolaj came to mind first, but hats off to plenty of others, I always listen to those people. And then there are the ones on a blade's edge, slipping between feedback and FUD, not sure which they want to give. I get that too. It's a fine line, and most of us wobble on it.
In the end it's all subjective. Outsiders often don't have the internal info to judge fairly, and you can't pass a rightful judgment without the full picture. So I try to hold my own judgments loosely, knowing I might be missing something. I'd ask for the same in return.
A house divided against itself cannot stand. So criticize, push, hold people accountable, all of it. Just don't tear the house down from the inside while you're still living in it.
Also, on Discord I have a ban button. On X, I don't. I think if @nikitabier gave me that option, @elonmusk would come ban me in real life. 😅
1
5
296
People think community work is easy. It isn't. You absorb a hundred of these every time something goes wrong, and it wears on you. The mental toll is real, and 'just ignore it' isn't the answer people think it is.
This is exactly why some people stay away from socials and public chats, keep their distance, or never get active in the first place. Because dealing with this is genuinely bs, and not everyone wants to sign up for it.
Got this in my DMs today after banning someone for verbal abuse in the community.
Racial slurs, death threats against my family, the works. Sharing it with everything identifying removed, not to shame one person, but to be clear about a line: this is not how human beings speak to each other, and anyone who behaves this way has no place in our community.
You can disagree with the team, criticize us, be frustrated, all of that is welcome. This isn't criticism. It's abuse, and it gets you banned. Every time.
Got this in my DMs today after banning someone for verbal abuse in the community.
Racial slurs, death threats against my family, the works. Sharing it with everything identifying removed, not to shame one person, but to be clear about a line: this is not how human beings speak to each other, and anyone who behaves this way has no place in our community.
You can disagree with the team, criticize us, be frustrated, all of that is welcome. This isn't criticism. It's abuse, and it gets you banned. Every time.
Dear {S}omnia community,
This message is not about Season 5, and I will not discuss it here. Feedback has been shared internally, and the team is reviewing the situation.
What I want to speak about today is how we communicate with each other.
Frustration is understandable. Criticism is allowed. Disappointment is allowed. Nobody is asking people to stay silent.
But there is a difference between giving feedback and letting anger, trolling, sarcasm, insults, or vulgarity take over.
In moments like this, I also have to reflect on myself. As Community Lead, I need to keep improving how we protect healthy community discussions, surface useful feedback, and separate criticism from behavior that does not help the conversation.
That is something I will keep working on.
But I also want the community to reflect.
Some people have been patient, calm, and diplomatic. Some stayed silent instead of adding to the noise. I respect that. Hats off to you.
Others chose mockery, insults, anger, or emotionally charged reactions. I understand emotions are high, but that does not make every reaction useful or right.
I want to share a few verses I personally reflect on in moments like this:
◇ Let every person be quick to hear, slow to speak, slow to anger.
◇ Death and life are in the power of the tongue.
◇ A soft answer turns away wrath, but a harsh word stirs up anger.
◇ Let no corrupting talk come out of your mouths, but only such as is good for building up.
◇ A fool gives full vent to his spirit, but a wise man quietly holds it back.
◇ What comes out of the mouth proceeds from the heart.
And finally:
You can be right and still lose people because of how you speak. The goal is not to win the reaction. The goal is to guide the outcome.
So my ask is simple.
‣ Say what you think.
‣ Give feedback.
‣ Disagree when needed.
But do it with control, respect, and clarity.
When emotions override your mind, your message usually becomes weaker, not stronger. A harsh message can bury a valid point. A calm message can actually move things forward.
This applies everywhere. Whether you are speaking to a teammate, friend, family member, stranger, or someone online, the way you speak matters. People do not only receive your words. They receive your tone, your intent, and your discipline.
When anger and vulgarity take over, the conversation changes.
a) Some people step back because they do not want to feed the noise.
b) Some stay calm and try to understand.
c) Some escalate and make the situation worse.
d) Some damage their own point by the way they speak.
I am saying this as a person who has interacted with many people, made mistakes, learned from them, and still fails sometimes. None of us are perfect. But we can still try to be better.
We are still growing as a community. We are still maturing. There will be difficult phases, mistakes, disagreements, and frustration. What matters is whether we let those moments divide us or help us grow stronger.
Thank you to everyone who is still here, still giving feedback, and still trying to push things in a better direction.
GG WP. Have a good weekend. See you next week.
At this stage, Thanos should log off, and Gandalf should take the staff.
Meet @fingersnapped, Community Lead at @Somnia_Network
From IT and security to onchain research and community, Alex built his path hands-on, starting in IT at 17 before moving deeper into crypto through Bitcoin, Ethereum mining, and running validators. Over time, he combined technical knowledge with product insight and community experience to operate fully onchain.
At Somnia, he leads the community team, ensuring operations run smoothly across moderation, support, and rewards. He also builds internal tools and workflows, working closely with ecosystem and product teams to improve efficiency and shape the community experience.
Outside of work, coffee keeps him going, and his faith keeps him grounded.
9
24
1,098
Don't buy keys from OpenSea. Thank you guys. BRB.
Due to the market explotation concerns raised, we are going to pause the program while we do a proper review on how to move forward from here.
We strongly advise people to stop buying from the secondary market until an official plan is shared.
We will communicate the next steps once that review is complete.