Just a random guy figuring out Web3.
Joined June 2024
- Tweets167K
- Following879
- Followers2.9K
- Likes236K
Pinned Tweet
A dataset is worthless until someone outside the company actually uses it.
@axisrobotics now has this proof.
As Booster Robotics' first sim-data partner, Axis rebuilt Booster's actual workspace as a digital twin. Distributed contributors collected 42,000+ simulation episodes on it.
With just 30 demonstrations using real robots, the model showed 87.5% success versus 37.5% for an out-of-the box π0.5. It achieved full π0.5 performance using half the real-world demonstrations required.
This is the actual value proposition backed up with real numbers, not something that is just claimed.
Beyond Booster Robotics, the partnering companies include: Embodiment companies like Feagine Robotics. Modeling companies like Manycore Tech and Dexmal. Industrial Automation companies like Lotus Cars and Geely Auto. Last but not least SomaStacks.
Axis also provides 2 on-chain robotics networks directly, BitRobot on Solana and OpenRoboto on Bittensor, therefore even other crypto-native robotics projects are using Axis’s pipeline to get the data.
This is a wide variety covering: Hardware companies, Modeling companies, Industrial Companies, Competing on-chain Networks and and all of them are using the same underlining data engine.
One case study with hard data is better than 10 partnership announcements without case studies. Axis has both.
Most NFT projects consider their mint as a means to generate revenue. 0xCyberThrone views their mints as opportunities to make consequential decisions.
@0xCyberThrone Genesis is a free mint, no fee, no upfront cost. That wasn't the easy default either, they spent months actually rethinking how a mint should work before landing there. The reasoning is straightforward once you hear it, optimizing for short-term mint revenue trades against the thing that actually matters long term, getting the collection into as many hands as possible and building a community that sticks around instead of flipping out within a week.
The mint itself runs in two phases, guaranteed access for people who registered early, then first come first served for everyone else. Simple structure, no hidden tiers.
There's also a mechanic in the whitepaper worth knowing about going forward. Every new use case they add comes with a new token airdropped directly to Genesis holders, each one unlocking a new room in their digital world and a new revenue stream for whoever holds onto it. That's a growth model built to keep expanding after mint, not a one-time utility list that goes stale.
gThrone
Keeping an eye on @ComputersRh today, good momentum building.
Over 23,000 people have already applied for Early Access, strong numbers for something still this early. Collab approvals are rolling out in batches, so if you're waiting, that's normal, not a sign something's wrong.
Early Access applications are closing soon though, worth applying now if you haven't yet.
New gameplay footage is coming too, possibly with a few mechanics revealed early. Keeping this one on the radar.
Before late apply fast 👀
Gm gm 🌻
Testing out some new PCs for @ComputersRh today.
Quick recap:
• 23,000+ people have now applied for Early Access. ✅
• Collab approvals are rolling out batch by batch, so please be patient with your applications.
• Early Access applications will be closing soon.
• We’ll also be dropping a new gameplay video and maybe revealing a few mechanics along the way. 👀
Note: Any numbers shown in the screenshots are from our testing environment and do not represent final in-game values.
Back to testing. 🖥️
Mr_Random retweeted
A dataset is worthless until someone outside the company actually uses it.
@axisrobotics now has this proof.
As Booster Robotics' first sim-data partner, Axis rebuilt Booster's actual workspace as a digital twin. Distributed contributors collected 42,000+ simulation episodes on it.
With just 30 demonstrations using real robots, the model showed 87.5% success versus 37.5% for an out-of-the box π0.5. It achieved full π0.5 performance using half the real-world demonstrations required.
This is the actual value proposition backed up with real numbers, not something that is just claimed.
Beyond Booster Robotics, the partnering companies include: Embodiment companies like Feagine Robotics. Modeling companies like Manycore Tech and Dexmal. Industrial Automation companies like Lotus Cars and Geely Auto. Last but not least SomaStacks.
Axis also provides 2 on-chain robotics networks directly, BitRobot on Solana and OpenRoboto on Bittensor, therefore even other crypto-native robotics projects are using Axis’s pipeline to get the data.
This is a wide variety covering: Hardware companies, Modeling companies, Industrial Companies, Competing on-chain Networks and and all of them are using the same underlining data engine.
One case study with hard data is better than 10 partnership announcements without case studies. Axis has both.
Mr_Random retweeted
Finally!!🥳
I got the "No Days Off" badge in @axisrobotics hub.
Here's my current progress :-
- Total 1550 trajcetories completed.
- Total 2238 Points.
- 10 Badges Collection Unlocked Yet.
- Current Average Score: 79.5
7 Days Rank:- 566
All time rank:- 843
📌 My goal is to still acheive "Seek the Unseen" badge.
Mr_Random retweeted
The future of NFTs goes beyond digital collectibles.
@0xCyberThrone is exploring how digital identity, creative ownership and technology can come together in one ecosystem.
With its robot collection and expanding vision across art, music, finance and digital content CyberThrone is building a universe where ownership can become the starting point for something bigger.
The real value lies in what the community can create around it.
Beyond the art.
Beyond the hype. Building a digital universe with purpose.
Web3 is more than just NFTs, tokens, and trading.
It’s about the people who support creators help communities grow and keep building even when the hype fades.
People like Tma420 remind us how important genuine support is in this space.
That’s why recognizing real contributors matters.
The Honorary CyberThrone NFT is a small way to appreciate someone making a lasting impact on the Web3 community.
@0xCyberThrone is building a vision that goes beyond digital collectibles.
By connecting digital identity creative ownership, art, music, and technology, CyberThrone is exploring new ways to give NFTs a bigger role in Web3.
But a strong ecosystem is not built on technology alone.
It also depends on the people who support creators help others grow and contribute to the community.
Recognizing contributors like Tma420 through an Honorary CyberThrone NFT reflects the importance of genuine support.
Great communities are built by people not just projects.
🤝 Paid partnership
The response to the @ptsdshow Non-FUDdable Ticket WL raffle has been massive.
16,000 submissions, 500 guaranteed WL spots and now an extra batch of FCFS spots for the community.
More chances for everyone to get involved.
The most interesting projects are the ones that turn real experiences into meaningful stories.
That’s what makes ptsdshow stand out to me.
By combining storytelling, AI characters and crypto culture, it creates an experience that goes beyond the usual Web3 narrative.
Looking forward to seeing how the ecosystem evolves.
ptsdshow is building a unique space where storytelling, AI characters and crypto culture come together.
The Non FUDdable Ticket WL raffle has already attracted 16,000 submissions with 500 guaranteed WL spots and additional FCFS opportunities for the community.
Beyond the numbers what stands out is the project's approach to turning real experiences into engaging stories.
A creative direction worth keeping an eye on as the ecosystem continues to grow.
Mr_Random retweeted
Imagine uploading a detailed spatial representation of the inside of an industrial facility
Suddenly the question:
“Where is this data stored?”
does not sound like boring compliance anymore
Spatial information can expose far more than location
- Layouts
- Entrances
- Equipment placement
- Operational areas
- Infrastructure
- Movement patterns
That makes sovereignty particularly important for Physical AI datasets
@vangrid_io's enterprise documentation currently describes regional data routing across US, European and APAC regions, with cross-region access requiring explicit account configuration
I think that design choice reveals where the enterprise product is trying to go
A global contributor network and sovereign enterprise data are not mutually exclusive
The collection layer can be geographically distributed while storage and access remain bounded
That matters for companies operating across jurisdictions
It matters even more for governments, infrastructure operators and organisations working with sensitive facilities
Physical AI will eventually create enormous datasets about real environments
Treating all of that information like ordinary cloud media would be careless
Companies will want to know not only who can access an observation, but which region it can leave
The internet made data globally movable by default
The next generation of physical-data infrastructure may need the opposite capability:
global collection combined with deliberate boundaries
Sometimes the most important feature of a data network is being able to prove where the data is not going
Just 14 people make up the core team behind a network that’s already verified over 2.4 million captures on-chain.
Robert Brighton, the co-founder & CTO of @vangrid_io, confirmed the team is made up of only 14 people. He stated hiring will be done for data and engineering operations in the next quarters. 14 people are a small number for a large production network, especially given the high levels of automation in the industry. Vangrid doesn’t capture the data, the crowd does. They just provide the framework for verification and payment. Vangrid let a distributed network of people capture the data, and leverages a crowd-sourced model.
The company itself doesn’t go out and collect the data. It provides a verification framework and payment and bounty mechanics to encourage a distributed crowd to collect data. In a traditional mapping company, if a community finds more data, it usually results in the company acquiring/buying more vehicles. For Vangrid, it means 14 people will lead to millions of data points captured.
The actual leverage lies in the crowd.
Mr_Random retweeted
. @ComputersRh giving 333 GTD WL for Creator.
So it's a good opportunity to booked your WL.
Creators can submit one post every 24 hours, so the people consistently putting out content actually have a way to keep participating.
Total supply 3,333 in Robinhood
1,889 Collabs
1,111 players
3,33 Creator's
And Creator NFTs aren’t just a badge either. They’re supposed to give a significant boost to in-game token earnings.
They are already partner with many big name .
So if you are apply for Computers Rh WL don't forget to submit post .
Early is early.
. @ComputersRh giving 333 GTD WL for Creator.
So it's a good opportunity to booked your WL.
Creators can submit one post every 24 hours, so the people consistently putting out content actually have a way to keep participating.
Total supply 3,333 in Robinhood
1,889 Collabs
1,111 players
3,33 Creator's
And Creator NFTs aren’t just a badge either. They’re supposed to give a significant boost to in-game token earnings.
They are already partner with many big name .
So if you are apply for Computers Rh WL don't forget to submit post .
Early is early.
In August, they got about 100,000 captures. On September 7, they got 502,752 captures. Now, they're at over 2.4 million. It's not growing linearly. It's growing exponentially.
@vangrid_io got a funding announcement, and about 2 months later, had a 24x increase in captures verified. Enterprise bounty completions were at 415 back in early September. That's worth keeping an eye on, considering how much the total capture volume has increased.
It's also worth mentioning this all happened without the token. There were no incentives. This was all organic growth, and the captures were paying out directly in USDC, resulting in the network exploding to the millions from under 100,000 captures.
A growth like this, before a token, is usually a stronger signal than a curve built on incentive farming.
Keep in mind that the capture volume was this low a month ago.
Check out the enterprise completions and see where it stands compared to a month ago.
@NucleusCodes
According to Axis estimates, Physical AI is likely to require about 100 million hours of training data to achieve general-purpose capability. Currently, their dataset totals 49,548 hours. This means they are approximately 0.05% to a general-purpose capability.
Compare that to how the current leaders in this space operate. Scale AI runs its Physical AI Data Engine out of one dedicated lab in San Francisco, over 100,000 production hours completed using paid contractors in a controlled facility. That's a real, serious operation, but it's still one location, one company deciding what gets collected and how.
@axisrobotics is betting the only way to close a gap this large is to stop trying to do it from one building entirely. 222k+ contributors worldwide, generating data in parallel instead of sequentially through a single lab's throughput limit. Centralized collection scales linearly with headcount and budget. A global open network scales differently, every new contributor adds capacity immediately, not after another hiring cycle.
100 million hours is a genuinely massive number to sit with. Nobody gets there through one lab alone. Worth understanding the actual size of the problem before judging how far along any single approach has come.
