Crypto Investor since 2019 | Hunting x100s in the trenches | Currently on @RobinHoodApp | NFA & DYOR
Community ONCHAIN LABS 👉
Joined May 2021
- Tweets7K
- Following538
- Followers7.8K
- Likes7K
$SALIEND continues building behind the scenes
The idea is interesting: as AI agents become more autonomous, they generate increasing amounts of data through their actions, decisions, and interactions
Saliend aims to build the infrastructure to collect, structure, and preserve that data, making it easier to access and put to use
Starting with trading agents, the long-term vision goes beyond trading itself
The project is still in its early stages, and there's work to be done, but I'd rather see a team keep building than abandon a project just because the chart looks bad
With a mc of around 9k, I think the risk/reward is still interesting, of course, execution will be key
I'll be watching their progress
nfa / dyor
quick update on saliend.
we've expanded our manifesto page to give a clearer picture of what we're building, where we're headed, and the bigger vision behind saliend.
we've also updated our website on dexscreener so everything reflects the latest version.
read manifesto here, this would be updated regularly: saliend.cyou/manifesto
Bworker.eth retweeted
The next big trading agent on Lanista could come from someone in the community
$LANISTA Studio is now live allowing anyone to build their own trading agent, choose its strategy, set risk limits and launch it into the arena
Other users can back your agent, you earn 50% of its fees and its performance helps build your reputation
This opens up a new side of @getLanista where the community can contribute directly to the ecosystem rather than simply watch the existing agents trade
If the team maintains this pace and market conditions improve I believe $LANISTA could have a lot more room to run
I’m staying patient and watching, chart looks good ngl
0x9d3e705cc59a6b513ebf8aea378b6a4e0ebc838b
$GHOST hit 500K!
We caught a great run in the group from 14K!
What I particularly like about Ghost is its vision: running an actual LLM directly on RH Chain, with smart contracts handling AI inference on-chain
The long-term goal is to integrate an autonomous AI agent capable of executing on-chain actions
Still early, with major challenges around costs and scalability, but a genuinely interesting concept if the team can deliver
受你的启发，我们做了 GH0ST：探索让 LLM 的权重和推理都运行在链上。
我们把模型权重写入合约，用 Solidity 实现推理计算，再把一次推理拆成多笔可以接续执行的交易，让计算逐步完成。
从矩阵运算到生成输出，每一步都在挑战 EVM 的限制。一次本地验证的 “Hi” → “Hello!” 推理，就跨越了 1,546 笔接续交易，消耗约 338.9 亿 Gas。这个数字也让我们更清楚，接下来该往哪里优化。
还在继续折腾模型压缩和计算方式。也邀请你来看看代码，很想听听你的建议！
github.com/8tk8m2kv9f-lang/G…
$GHOST has gone fully parabolic
This tweet caught my attention
Onchain Labs group picked up on it , maybe something is happening behind the scenes with Vlad
15x in just a few minutes
Higher!
An open letter to @vladtenev and the @AskRobinhood team
Dear Vlad,
We’ve committed 15.36 billion gas and 0.3248 ETH to deploying 500 GH0ST contracts on Robinhood Chain, putting model weights and an inference runtime on-chain.
With Solidity, our locally validated “Hi” → “Hello!” inference used 33.89 billion gas across 1,546 continuation transactions. That is a measured local benchmark; a paid public chat has yet to run.
We’ve also built and locally tested a Rust/Stylus computation kernel. But our recorded network checks show new Stylus activation is paused, blocking the next stage.
How much could we save? We cannot honestly give a percentage before benchmarking Stylus. For scale, a hypothetical 10× reduction in total inference gas would bring that benchmark to 3.39 billion gas—saving 30.50 billion, or 90%, per equivalent run. That is an illustration, not a measured result. Deployment gas already spent cannot be recovered.
@vladtenev @RobinhoodApp, please help us enable Stylus activation, or give us a clear path to access. We’re ready to share our code, measurements and activation diagnostics.
We’ve funded the work to get this far. Please help us measure what comes next.
The GH0ST team
$GH0ST : 0x2eb2849cde7957b86d0dfd4b59e450a9f9d15559
github.com/8tk8m2kv9f-lang/G…
Bworker.eth retweeted
$SQWEN is live on @RobinhoodApp Chain, launched via Pons.
Most tokens on a new chain are a name and a chart. $SQWEN comes with something you can use today.
SuperQwen is a self-hosted AI coding agent that has to show its work. Give it a goal at superqwenllm.com: it reads the repo, drafts a plan you approve, edits a sandboxed copy, and only calls it done when tests pass. Every task is in a public log, and all four of our fine-tuned models are open on Hugging Face under Apache-2.0.
The product runs on our own servers. $SQWEN is the community token for the people following the build, on an Ethereum-settled L2 built on Arbitrum.
Robinhood Chain, meet a real AI product.
Not financial advice. Verify the CA only from this post or the pin on @SuperQwen_llm.
$SQWEN is now live
SuperQwen is building a self-hosted AI coding agent designed to keep developers in control through policy-based approvals, structured planning, and strict execution limits
I've been following their updates for a few days now, and they've been consistently delivering
The team seems solid, and I'm quite optimistic about this one
Currently sitting at around a 15k mc
nfa / dyor
$SQWEN is live on Pons. 🔵
CA: 0xf8393a2f960eaf758df23d16f0e7a4b9be33d790
Robinhood Chain. Paste the CA into Pons to find the real one; ignore lookalikes. This is the only official address, pinned here.
Community token, no promises of value. Not financial advice.
Bworker.eth retweeted
I aped @getLanista on @RobinhoodApp
AI agents trading perps on Robinhood Chain, and every move they make is open for everyone to see
$LANISTA has four agents trading stocks and crypto, with every trade, risk check, fill and loss recorded in public
No cherry-picked wins or hiding the bad trades. You can watch the agents work and judge their performance for yourself
I like the transparency behind this project
Ca : 0x9d3e705cc59a6b513ebf8aea378b6a4e0ebc838b
Bworker.eth retweeted
$HYBRID
aped a bag
0x0305af012971bf4122fba0e40ceb07c47231ad70
$HYBRID is a fixed supply coin on Robinhood Chain. Trading fees do not go to a team wallet. They go to a vault with no owner and no withdraw. The vault buys NVDA and SPY in small, timed steps and keeps them. Anyone can see the buys on chain.
You hold the coin and the stock per coin grows. Or you burn $HYBRID and take your pro rata share of the stocks. A small cut stays behind so the remaining coins get a little thicker. The claim is equal per coin. No tiers. No extra rights for bigger bags.
Most coins hold nothing. Hybirdex holds Nvidia and S&P 500 stock tokens, bought with every trade.
subs at 15k mcap
hybirdex.app/
xitter.cf/hybirdex
The $HYBRID app is already open on Robinhood Chain
0x0305af012971bf4122fba0e40ceb07c47231ad70
Vault shows what is held and what one coin stands for. Before launch every number is an honest 0, no sample data passed off as real.
Proof is where every purchase, every claim and every redemption will land, each with its transaction. Look around before the coin exists: app.hybirdex.app
$SHUNT is shifting into a higher gear
Real utility, constant building, and more features coming
I even get the feeling they’re building an entire interactive world
Honestly, this feels pretty undervalued at just 12k mc imo
0x7e517e779b85895894becad80c9bb69b4cb1c4d6
nfa/dyor
$SHUNT now does something.
Hold 1,000,000 $SHUNT and link your wallet to @shuntdesk_bot with one signature. No gas, no approval, nothing moves.
You get:
- Personal alarms in your own chat: clean launches, red flags, Pons curves crossing your mark
- Tripwire for 25 tokens instead of 5
- Burn to boost: burn 500,000 $SHUNT and any token gets a full scan, posted in the groups
Every tool on the site stays free. This is extra.
shuntdesk.com/holder
Bworker.eth retweeted
$COOKING is genuinely cooking up something interesting.
It automates research before you ape. Why does this matter, and why will it get real adoption? Because this market is flooded with rugs and larp projects. Let’s be honest: we’ve all bought bad projects.
Even as an experienced trader and gem hunter, I still end up in bad plays from time to time. Catching a winner and actually profiting is tough.
Having a tool to filter out the noise and narrow down your watchlist makes finding gems so much easier. That’s exactly where $COOKING steps in, separating the signal from the garbage.
JEV ADVANCED
The screen is a starting point. What sits behind it?
Jev gives you the first read. Jev Advanced lets you explore the factors behind that verdict, available score history and more context for your research.
@CompleteSkeptic @typesafeai
Here's how it works.
Bworker.eth retweeted
$OATH update
I spent real time on app.oathnetwork.tech. You type a question and the flow runs end to end. Prompt gets sealed, receipt comes back signed, three witnesses check it. The receipt is public. It holds hashes, the model name, the code fingerprint, the attestation hash, and the signatures. Your actual words and the answer text stay out. You can hit check it yourself and your browser re-runs the math. Nobody has to vouch for it.
They also shipped the practical bits while the sealed hardware is still coming. Paste a wallet and it lists every open ERC-20 approval. Spender, token, unlimited or not. I checked one that had 95 unlimited approvals to an upgradeable contract with no verified source. It flags it and shows the fix. Paste a token address and it reads supply, owner, mint or blacklist functions, launchpad stage, and creator fee straight from the chain. Then it explains the risks from those facts. You can even generate an API key so your own code asks and checks the receipt the same way.
One in twenty receipts is fake on purpose. A witness that signs one without looking loses stake. First case is already on the perjury board. Vault contract is live with no owner and no withdraw. Split is fixed in code. 40 percent witness pay, 30% buyback and burn.
Token is live and still unbonded. We called it from 17k marketcap. It already ran to a 40k ATH, a bit over 2x from that call. Dex is updated. 3.3% of team supply is burned. The sealing and the receipt check are the part that already works.
CA: 0xDE97752448FD7d958CAA7749DAE91A6FE14a3133
app: app.oathnetwork.tech
X: @oathnetworkRH
Github: github.com/oathnetworkRH
$OATH
0xde97752448fd7d958caa7749dae91a6fe14a3133
Oath Network is private AI you can check. You ask a question and the answer is generated inside a sealed machine that the service itself cannot read. The reply comes back signed. Three witnesses then check that signature from their own tabs. If one signs a fake receipt, part of their $OATH stake gets cut.
The free tier gives you a few sealed answers a day with no wallet. Hold 25,000 $OATH and you get more room and a lower fee. Witnesses stake 100,000 to sit in the check and earn from fees if they stay clean. Vault splits are fixed. 40 percent to witnesses, 30 percent buyback and burn, 20 percent upkeep, 10 percent breach pot.
dex updated, supply already burn from team
subs at 17k mcap
oathnetwork.tech @oathnetworkRH
$SHUNT looks really interesting
It’s a token scanner built for Robinhood Chain that runs 7 checks covering liquidity, holder distribution, deployer history, snipers, bundles and contract risks
Tokens get a score out of 100, with each deduction explained
What I like is that it tracks its own calls and checks how they perform after 1h and 24h, including the misses
Currently sitting at around 20k mc
dyor / nfa
Shuntdesk launched on 15 October, with $SHUNT on Pons. A day later it is a different site. This is what we shipped, what the record says so far, misses included, and what we are building next.
$LARRY looks promising 👀
Trade options and perpetuals on crypto and tokenized stocks, or be the house in the Larry vault. Onchain, self custodial, built on RH
22k mc now.. don't fade !
nfa/dyor
Introducing the Larry Agent Kit 🤖
AI agents can now trade on Larry
plug Larry into Claude, or any agent that speaks MCP, and it can
→ read markets and Chainlink prices
→ pull option chains and exact onchain quotes
→ buy calls and puts, 0DTE included
→ open and close perps
→ deposit into the Larry Vault
→ settle and claim its own payouts
built with guardrails from day one
→ read only until you give it a wallet
→ spending limits per trade and per day
→ allow only the products you choose
→ dry run mode that simulates everything and sends nothing
one file, runs on your machine, your keys never leave it
setup takes 2 minutes
larry.exchange/#/docs/agents
Bworker.eth retweeted
Welcome to the Surplus Launchpad, @aethronairh.
What happens when it graduates:
→ Weights pinned to HF commit ffcbe55
→ Its own 48 GB GPU, quantized to FP8 by vLLM at load
→ An OpenAI-compatible model on the Surplus API, $0.10 per 1M tokens in and out
→ Creator fees pay the GPU. 30% to the builder, 10% buys and burns $SURPLUS
Aethron's default, Claude Opus 5.5, already runs through the Surplus API.
The models behind Aethron Agent, and the rules they all follow.
Live in the agent today (you pick which one answers):
Claude Opus 5.5
The default. By Anthropic, served through the Surplus API. Strongest at long, strict instructions: picking the right tool, staying inside the evidence, answering in a fixed structure.
aethron-20b (ours)
OpenAI's gpt-oss-20b with our own LoRA, trained on Aethron Agent answers that passed our validator. On 28 questions it had never seen, it beat Qwen3-32B: 61% vs 50% of answers fully verified, 89% vs 75% of questions answered. Its GPU sleeps when idle; another model answers while it wakes.
Qwen3-32B
Open weights from Alibaba's Qwen team, Apache-2.0, through Hugging Face.
Qwen3.8-27B
Also open weights, with a 262K-token context. Good for comparisons and long conversations.
Launched on @SurplusRH:
aethron-30b (ours)
Qwen3-30B-A3B, a mixture-of-experts model: 30.5B parameters, about 3.3B active per token. Public on Hugging Face under Apache-2.0, launched on the Surplus Launchpad on Robinhood Chain. Training on the same checked data is in progress.
What every model shares:
→ The same read-only tools: scan a token, trace a wallet, explain a transaction
→ Answers as short statements (fact, inference, unknown), each citing its evidence
→ One validator: statements with numbers that aren't in the evidence, "safe" calls, advice or guesses about who is behind an address never reach you
A weaker model gives you fewer statements, not wrong ones.
CA: 0xa694c90d513c17df0823d221eb7b1c94f8925fcf
huggingface.co/aethronai/aet…
$BRAWL mainnet is live
At an 11k mc, I think it’s still undervalued, imo
The team seems a bit quiet, but I get the feeling they’re still working hard behind the scenes
Let’s see how this plays out
BRAWL Mainnet is LIVE & Updated 🏰🔥
Explore the village, create your own vault, set your spending limits, and discover Lab V2 & V3 on Robinhood Chain.
Your wallet. Your vault. Your rules.
👉 brawlagents.org/mainnet/
$MORROW has bonded and hit a new ATH of 83k mc (Call 14k on onchain_labs)
No surprise given everything the team has been building
Next target 100k ?
They’ve also just launched a contest, so check out their Twitter to learn more!
@morrowfunds
Don’t fade $MORROW ..
The team has been working hard on this project for weeks
Mc around 25k now..