hemis07 retweeted
The first encrypted financial infrastructure connecting NFT-owned AI Agents with DeFi and @Polymarket markets.
20,000 NFTs. 20,000 AI Agents. Powered by $MOTES.
→ Applications open: meetmotes.com/apply-form
hemis07 retweeted
Winner announcement in 24 hours! We'll randomly draw 100 eligible winners with help from @grok. Make sure you follow us, like & repost to be in the running. Good luck! 🍀
OSF explains how grid bots buy dips and capture volatility
“It’s essentially as prices go lower, it’s adding more risk, adding more risk, adding more risk.”
“Then the hope is that from here we bounce up, bounce up and then it starts to sell out of the risk.”
“So even though I’m not like putting on massive risk, this bot has organically and naturally bought the dip.”
TL;DR for the CN Community 🦭
CRC-20 已经有代币,但还没有 NFT 市场。Seals 是第一个。
· 1,500 枚|69 种特征|8 枚 1/1
· 市场手续费:60% 持有者|30% 回购|10% 协议
· 路线:等候名单 → 铸造 → 持仓页 → 交易市场 → 分配与回购
加入浪潮 👉 seals.garden
实验性协议,请使用新钱包。
hemis07 retweeted
I no longer worry about trading an unverified token or giving approval to contracts.
You know why?
Because I found a meta dex aggregator that has got me covered.
Yes you heard that right
That meta dex aggregator is OMNISWAP that has an inbuilt trading wallet for each user using the Dex for trading.
This is made possible by the platform engine called the Batcher Engine.
Also transactions are also very fast and comes with no extra cost to the platform users
Trades from Robinhood Chain to base chain delivers within seconds
🫡 to @namidefi team for this great innovation.
wagNami
hemis07 retweeted
Trying to keep a distribution tool trained on billions of views under wraps before launch? Wild strategy.
It's been brought to our attention that multiple individuals have found Clipur.ai
It is our unreleased product that allows anyone to launch viral distribution
Trained on 500+ successful campaigns accumulating in billions of views and tens of millions of dollars in business value
We were not planning on releasing or announcing this yet, but we needed to shed clarity on some nefarious accusations that have surfaced recently
hemis07 retweeted
Axis Robotics Epoch 4 Points have been distributed
If you participated in the @axisrobotics hub tasks for Epoch 4, your points should have arrived in your portfolio
I got 547 points for my Epoch 4 tasks
This is the least I have had so far since epoch 1 🤦🏻♀️
I really hope to do better for this new epoch
The new epoch is currently live for everyone to participate
If you are new, join here:
s.kaito.ai/U7aW2rx
Also, @axisrobotics announced a community sale of 1% of the total $AXIS supply on @echodotxyz on @base chain at $100M starting FDV
The pre-registration is currently open, and the commitment starts by 21st of September, 2026
The sale will last for 7 days, and the settlement begins on the 30th of September
Don't miss out on both the Axis Robotics airdrop and community sale frens
NFA!
Axis Robotics Recent Partnerships and Their Contribution to the Growth and Adoption of Robotics 🧵
I have been looking more closely at the partnerships @axisrobotics has been making recently
and honestly, I think there is a bigger picture here that is easy to miss by most people:
These partnerships are literally connecting @axisrobotics to different parts of the Physical AI stack
Now, let us look into some of these partnerships and how they contribute to the growth and adoption of Robotics⤵️
hemis07 retweeted
While everyone seems focused on NFTs on Robinhood Chain and other ecosystems right now, I recently came across a project that genuinely caught my attention for a different reason. The project is @namidefi, and after spending time going through its documentation and testing its live product, I think it is one of those opportunities that many may be overlooking at the moment.
What first drew me in was OmniSwap, which is already live and available for use. At first glance, you might think it is just another aggregator, but it goes a step further. Instead of making users jump from one aggregator to another to compare rates, routes, fees, and execution options, OmniSwap does the heavy lifting itself. It compares aggregators, bridges, and solvers to find the most efficient route for a transaction, helping users get the best possible value without having to manually search for it. In simple terms, we can say OmniSwap aggregate the aggregators.
What makes this even more interesting is how it handles transactions. Whether you want to swap assets on the same chain, bridge assets across chains, top up gas, or perform multiple token actions, everything can happen within a single transaction flow. This is made possible by Nami's proprietary Batcher Engine, which bundles multiple calls together and delivers a much smoother experience for the user.
As I continued reading through the documentation, I found another aspect of the project that made me pay more closer attention. Nami plans to launch the ogNAMI token, and according to the model, users who stake ogNAMI will be entitled to a share of the value generated within the ecosystem. Specifically, 10% of swap fees are intended to be distributed to ogNAMI stakers.
When I thought about what that could mean, it immediately stood out to me that most users interact with protocols every day without ever participating in the value created by those platforms. @namidefi is taking a different approach by creating a structure where users who stake ogNAMI can potentially share in the growth of the ecosystem itself. That was one of the biggest reasons I became interested in learning more.
Let's talk about the model that enticed me, the token supply caught my attention. The total supply of ogNAMI is set at just 250,000 tokens. Out of that, 50,000 tokens are reserved for liquidity provisioning, leaving only 200,000 available for public participation.
The team didn’t stopped there either. The protocol is also designed in such a way that makes the OgNami token deflationary by setting aside a portion of the protocol fees to be used to buy back and burn ogNAMI tokens. From my perspective, that shows an effort to create long-term alignment between protocol activity and token holders.
After spending time with the documentation, I decided to test the product myself rather than rely solely on what I had read. The experience I had testing out the OmniSwap only strengthened my interest. It makes everything feel connected and straightforward, which is exactly what users need in an increasingly fragmented space.
What ultimately stands out to me is that this is not a project built around hype alone. There is already a live product solving a real problem, a clear utility behind the ecosystem, a revenue sharing model for stakers, a limited token supply, and a mechanism designed to reduce supply over time. These are the reasons I have been paying close attention and positioning myself for the upcoming ogNAMI sale.
If you are the type that like discovering innovative projects before they become widely discussed and appreciate products that focus on improving the user experience, @namidefi is definitely worth taking a closer look at. For now, I will keep exploring OmniSwap and follow the development of the ecosystem closely so as to position myself for what's to come from Nami team.
Follow @namidefi, turn on notification to stay updated on the latest updates and how the project is progressing.
GG guys. 🚀
hemis07 retweeted
N3on pulled up to Target dressed like Mario, turning a simple shopping trip into a full-on Jack.com side quest 😂 @N3onreload
hemis07 retweeted
Axis Robotics Recent Partnerships and Their Contribution to the Growth and Adoption of Robotics 🧵
I have been looking more closely at the partnerships @axisrobotics has been making recently
and honestly, I think there is a bigger picture here that is easy to miss by most people:
These partnerships are literally connecting @axisrobotics to different parts of the Physical AI stack
Now, let us look into some of these partnerships and how they contribute to the growth and adoption of Robotics⤵️
Had the opportunity to host the Post-Training Lesson on the Axis server today.
Big thanks to everyone who joined and contributed to the discussion, and a special shoutout to @iamlogtun for trusting me with the session.
Really enjoyed sharing what I know and learning from everyone who showed up.
We go again at the next event. 🤖
🚨Post-Training Onboarding Session Ongoing
If you're struggling with tasks or finding it difficult to find your way around the @axisrobotics hub, this is for you.
@0xhu7i is currently taking an onboarding session that will equip you with the knowledge to contribute data trajectories toward robotics intel on the Axis hub.
If you miss tonight, don't worry.
We host this session weekly (Tuesday & Thursday) and will be excited to see you there.
Join us NOW!
Link below ⬇️
hemis07 retweeted
Check Your Activity Score on Axis Robotics Hub
yesterday, I stumbled upon an interesting tool built by @Crypto_Pranjal used for checking your activity score on the @axisrobotics hub
so I decided to check mine and guess what ?
I have 91/100 score with the axis hub 👏
you can also check yours here:
axis-robotics-checker.vercel…
copy your axis dashboard wallet address and paste in the search box to check
quote this tweet with your score, let’s see how active you’ve been on the Axis hub mate
P.S: the stats showing on the website might not be as accurate as the one on the dashboard as it takes time to update
Kudos to @Crypto_Pranjal for building this useful tool and gAxis to all my axis frens
@iamlogtun @plpiaoliang
Why I Think Axis Robotics Could Become the ChatGPT of Robotics 🧵
one thing I really like about @axisrobotics is that it isn't just another robotics project building robots behind closed doors
@axisrobotics is building a hub where users can actually contribute data to the development of better robots
think about what ChatGPT did for AI
it gave millions of people a simple way to interact with AI, while all those interactions helped push the artificial intelligence technology forward
Axis Robotics is taking a similar approach, but for robotics
instead of just using a robot, YOU can help TRAIN it!
through the Axis hub, contributors can interact with robots, complete different tasks and record trajectories
these trajectories become valuable training data that can be used to improve robotic models
every useful trajectory adds another piece to the dataset used in building better robots, just like every prompt adds more knowledge to the AI
and the more high-quality data we can generate, the better robotics models can become
this is basically creating a feedback loop:
more contributors → more data → better training → smarter robots → more useful robots
this is why I see @axisrobotics as more than just a robotics platform
it is a data contribution hub for the next generation of intelligent robots
ChatGPT showed us what happens when you put powerful AI technology in the hands of millions of people
@axisrobotics is exploring what happens when you do something similar with robotics
and honestly, we are still very early
the future of robotics might not only be built inside research labs
it could be trained by people like YOU and I!
gAxis frens.
@iamlogtun @plpiaoliang
hemis07 retweeted
What's going on with @TrustWallet
The wouldn't allow me send funds out
Below is the error message
My network is fine
App up to date
Where then is the technical error coming from.
Is the wallet safe at this point?
190k+ users on the Hub and I’m somehow #2,332 points wise.
Honestly, this feels like Hyperliquid points before they got converted
If you’re wondering why your tasks keep showing FAILED, here’s a quick guide to what might be happening and what you can do:
Why tasks fail:
Backend verification failure — the trajectory didn’t pass the system-side checks. This is the most common one.
Bugged task — something in the task itself prevents it from being completed or validated properly.
Checker failure — the validation checker errors out. Less common, and if other trajectories are signable, simply retake the task.
How to increase your chances of SIGNABLE:
1. Disable checkpoints → Temporarily turn them off in Hub settings. Fewer interruptions = a smoother task flow.
2. Don’t rush → Tasks completed in under ~4 seconds have a higher chance of being flagged as FAILED or scoring lower. Take your time.
3. Go manual → Complete tasks manually. Avoid automated/scripted patterns and make sure your actions are deliberate.
4. Aim for 6–15 seconds → This seems to be a good window for balancing speed, accuracy and consistency.
5. Read & execute carefully → Follow the instructions exactly. Small mismatches can cause a trajectory to fail validation.
Basically: Read → Execute → Don’t rush → Submit → SIGNABLE.
Hopefully this saves someone from rage-quitting the Hub
gAxis
🤝 Paid partnership
hemis07 retweeted
Seeing two frontier models build the exact same email flow side by side is wild.
Fable definitely takes this round.
hemis07 retweeted
Goat Street NFT GTD/WL form is now live.
48 hrs for applications.
Supply: 6666
Chain: @RobinhoodApp
MP: Free Mint
Form:
docs.google.com/forms/d/e/1F…
Axis just crossed 123,000+ contributors.
at some point you'd expect a project to hit a ceiling. "we have enough people now, let's focus on quality over quantity." that thinking would be wrong here, and it comes down to how distributed data collection actually works.
most robotics AI is still trained on data from one lab. one room, one camera setup, one small team repeating tasks over and over. then the robot meets a real, messy, unpredictable environment and falls apart, because it was never shown enough of the world to generalize.
axis flips that. instead of a handful of engineers generating data in a controlled space, a global network of contributors captures different environments. different homes, different lighting, different objects, different ways people actually move through the world. that's the shift from lab to global.
here's the part that's easy to miss: this isn't just "more data." it's a network effect, and network effects compound instead of adding up.
more contributors join, so more diverse data flows in. more diverse data means the model generalizes instead of overfitting to one lab's conditions. a model that generalizes becomes genuinely useful. a useful, valuable project pulls in even more contributors. each loop makes the next one stronger. that's the actual mechanism behind "the bigger the community gets, the stronger the data engine becomes." it's not a growth slogan, it's cause and effect.
which is why 123k isn't a ceiling. the world doesn't run out of environments. a robot needs to function in every kitchen, every warehouse, every street, every messy human space that exists, that's not thousands of scenarios, it's effectively infinite. coverage is never actually complete. a robot trained mostly on data from a few regions fails in ways nobody predicted the moment it meets a new culture, a new layout, a new way people handle objects. more contributors means more of those blind spots get filled before they become real world failures.
tasks keep multiplying too. robotics isn't one use case, it's folding laundry, sorting packages, navigating stairs, handling fragile items, working alongside humans. each task category needs its own depth of examples, and more contributors means more task diversity covered at once instead of one at a time.
data also goes stale. this isn't a dataset you build once and reuse forever, environments and objects and human behavior keep shifting. a static pool of contributors eventually produces static, aging data. a growing pool keeps the pipeline fresh.
and redundancy here isn't waste, it's verification. more people submitting on similar tasks is how bad data and inconsistent labeling get caught. quality at scale needs volume underneath it.
this is why the comparison to normal hiring breaks down. in a typical team, more people past a certain point creates redundancy and inefficiency. in a network effect, more people creates more value for everyone already in it. wikipedia didn't get worse at 10 million editors, it got more reliable.
so 123k isn't a milestone to plateau at, it's proof the model works. the honest answer to "should there be a cap" isn't "no, growth is good for its own sake." it's "no, because the problem axis is solving, getting AI to actually understand the physical world, doesn't have an edge. it just has more territory the bigger the community gets."
AI learned language by digesting a giant pile of text that already existed online. A robot has to learn something nobody wrote down anywhere: how to physically pick up a mug, sort objects, stack boxes, without breaking anything. There's no "internet of robot movements" sitting around for a model to learn from, someone has to go generate it on purpose, in huge volume, from scratch.
And quality matters way more here than with a chatbot. If AI gives you a dumb answer, worst case you're annoyed. If a robot arm trains on sloppy, mislabeled movement data, it can crush something or damage real equipment on a line.
So every company building what the industry calls "physical AI" (AI that controls a body, not just a chat window) is stuck between two walls: not enough real-world movement data exists, and what does exist is often messy or doesn't transfer between different robot hardware.
Axis brands itself as a physical AI "data engine" built to fix three things:
-not enough training data exists for robots
-the data that does exist doesn't generalize well
-data collected on one robot often doesn't transfer to another
Here's the flow, start to finish:
Global contributors (100,000+ people worldwide)
|
Raw movement data (Teleop,sim & mobile capture)
|
Cleaning and labeling pipeline (Filters and standardizes data)
|
Enterprise customers (Buy dataset to train their models)
Walking through it: at the bottom of the funnel are ordinary people, not robotics engineers.
More than 100,000 contributors around the world are pitching in, together producing roughly 1,200 hours of simulated robot footage plus about 20,000 hours of real-world footage every month operating within a large-scale "human-in-the-loop" global data engine.
They generate that footage two main ways: piloting robots in real time from an ordinary web browser tab, no extra hardware or local computer setup required, or, for real-world capture, using a phone app that tracks hand movement and translates ordinary human motion into robot motion.
That raw footage isn't useful on its own, though it runs through Axis's own pipeline first. There's a system that builds tasks out of plain instructions, matching them against a library of over 2,000 3D objects and covering more than 90 task categories, things like picking up items, pouring, sorting, and stacking. Afterward, the footage gets filtered to strip out corrupted or failed clips, then smoothed out so it's actually clean enough to train on.
That cleanup step is the "high quality" part of the topic, it's the difference between data that teaches a robot something useful and data that just adds noise.
Why enterprise customers pay for this instead of building it themselves
Standing up their own robot-data operation is slow, expensive, and a completely different skill set than building robots or training models.
So instead, Axis packages this up commercially and sells ready-made data bundles to robotics hardware manufacturers, physical AI model companies, and industrial automation players, they order the exact training data they need instead of building the pipeline in-house.
This isn't theoretical: real companies are already buying in, including Booster Robotics, Geely Automobile, and Manycore Tech. And there's a concrete result to point to, feeding its dataset into a robotics model called π0.5 reportedly lifted its success rate by 4.9 percentage points on a benchmark called LIBERO-Plus.
So the whole pitch in one line: robots can't get smarter without real-world movement data, that data is genuinely hard and expensive to make well, and Axis exists so robotics companies can buy the hard part instead of building it themselves.
































