Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
Replying to @D_cryptographer
The interesting part is that AI infrastructure no longer has to be limited to massive data centers.
Crynux gives independent GPU owners a way to contribute, while developers get access to the compute they need.
Both sides benefit from real network usage.
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
๐ถ๐ป ๐๐ต๐ฒ ๐๐ ๐๐ฝ๐ฎ๐ฐ๐ฒ, ๐ฝ๐ฒ๐ผ๐ฝ๐น๐ฒ ๐ต๐ฎ๐๐ฒ ๐ฑ๐ถ๐ณ๐ณ๐ฒ๐ฟ๐ฒ๐ป๐ ๐ฎ๐ฟ๐ฒ๐ฎ๐ ๐ผ๐ณ ๐ณ๐ผ๐ฐ๐๐.
-some people build AI applications.
-some focus on developing models.
-others work on the infrastructure that makes all of this possible.
everyone has their own lane.
but i want to talk about the people building the models.
you could spend months
-training a model,
-fine-tuning it and
-improving its capabilities.
beyond getting it to work, thereโs also the question of how to make that model accessible to other people who might want to use it.
.
.
@crynuxio has an interesting approach to this with Model as a Service.
it allows model creators to make their models available through the Crynux network, so applications can access and use them for AI workloads.
thereโs also an opportunity for model creators to earn from the usage of their models.
so your work doesnโt necessarily have to end when the model is ready.
there could be an opportunity to make it available for others to build with and benefit from its adoption.
if building AI models is your thing, Crynux is worth exploring.
you can also try its AI Services with 5,000 free credits through my link:
crynux-as.xyz/?ref=aKbYSzCyrโฆ
$CNX
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
๐๐๐ ๐
๐๐๐ ๐๐๐จ๐ฌ๐ฒ๐ฌ๐ญ๐๐ฆ: ๐ ๐๐จ๐ฆ๐ฉ๐ฅ๐๐ญ๐ ๐๐ฎ๐ข๐๐ ๐ญ๐จ ๐
๐๐๐, $๐๐๐๐ ๐๐ง๐ ๐๐ก๐๐ญโ๐ฌ ๐๐จ๐ฆ๐ข๐ง๐ !
>>>>
Iโve been following @tonfish_tg and its recent developments, and thereโs quite a bit happening around the ecosystem.
The latest development, FEED, caught my attention because it takes a different approach to how people participate in and showcase their presence in TON.
Letโs break down what theyโre building, what is already available, and how you can get involved.
๐งต
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
Imagine proving you're over 18 without showing your ID.
Or proving you can afford something without revealing your bank balance.
That's what zero-knowledge proofs can make possible.
@NeptunePrivacy builds around this technology, and a recent upgrade brings it closer to its next major milestone: Leviathan mainnet.
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
๐๐ผ๐ ๐ฑ๐ผ๐ป'๐ ๐ต๐ฎ๐๐ฒ ๐๐ผ ๐ฏ๐ฒ ๐ฎ ๐ฐ๐ฟ๐๐ฝ๐๐ผ ๐๐ฟ๐ฎ๐ฑ๐ฒ๐ฟ ๐๐ผ ๐ณ๐ถ๐ป๐ฑ ๐ฎ ๐ฟ๐ฒ๐ฎ๐๐ผ๐ป ๐๐ผ ๐ฝ๐ฎ๐ ๐ฎ๐๐๐ฒ๐ป๐๐ถ๐ผ๐ป ๐๐ผ @crynuxio
โบ if you're a developer, you can use its AI Services to build applications and experiment with AI without having to own your own GPU infrastructure.
โบ if you have computing resources, you can explore becoming a node provider and earn rewards for completing tasks on the network.
โบ and if you're simply curious about AI, you can explore Crynux Chat and experience AI powered by the network.
there's also an economic layer connecting these activities.
developers pay for compute,
node providers earn for contributing resources,
and $CNX facilitates payments,
... rewards and staking within the ecosystem.
what i find worth exploring is that there are different ways to participate, depending on what you bring to the table.
you don't have to approach Crynux from just one direction.
if you're looking to build with AI, start exploring Crynux AI Services. you get 5,000 free credits to experiment with.
if you want to explore the network as a node provider, check out @crynuxio.
find your entry point and see what you can do.
link in cs โก
$CNX
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
Private trading on Leviathan is no longer just an idea. Itโs something you can explore today
NeptuneSwap is LIVE on the public testnet, bringing another important piece to the Leviathan ecosystem
Donโt just watch the ecosystem grow get involved
Explore NeptuneSwap, try out private trading, and share your experience
The best way to understand whatโs being built is to experience it yourself
Start here ๐ testnet.zkswap.ai/
Did you know?
NeptuneSwap, our private decentralized exchange, is already live on the Leviathan public testnet.
Private smart contracts laid the foundation. Our Chrome and hosted web wallets opened the door. NeptuneSwap adds private trading to the growing Leviathan ecosystem.
Explore private trading on Leviathan, test it out, and share your feedback.
The roadmap is becoming reality, one milestone at a time.
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
Did you know?
NeptuneSwap, our private decentralized exchange, is already live on the Leviathan public testnet.
Private smart contracts laid the foundation. Our Chrome and hosted web wallets opened the door. NeptuneSwap adds private trading to the growing Leviathan ecosystem.
Explore private trading on Leviathan, test it out, and share your feedback.
The roadmap is becoming reality, one milestone at a time.
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
Those numbers donโt care about the current form.
20 league titles, 3 Champions Leagues, 13 FA Cups thatโs not nostalgia, thatโs a rรฉsumรฉ.
The meme is funny though. Still funny even while the trophy cabinet is louder. ๐ด @1win
Man United fans, sorry ๐ญ
But letโs not forget:
20 English league titles - won between 1908 and 2013, including a record 13 Premier League titles since 1992
3 European Cups / UEFA Champions League titles - 1968, 1999, 2008
13 FA Cups - most recently in 2024
6 League Cups โ most recently in 2023
21 Community Shields - including 4 shared titles
1 UEFA Europa League - 2017
(but meme is funny)
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
Not every holder has the same relationship with a token.
Some buy for a quick trade. Others are willing to hold through volatility if the incentive structure rewards patience.
@RebateLaunch uses holding-time multipliers for eligible token holders:
- 1x during the first 0โ3 days.
- 1.5x after 3โ7 days.
- 2x after 7 days.
That creates a clear reason to consider holding longer instead of treating every position as a short-term trade.
But there's an important distinction: a higher multiplier changes payout weight, not the market value of the token. It doesn't guarantee a larger dollar return or protect someone from a price decline.
If I were evaluating a launch with this model, I'd compare the possible payout against price volatility, liquidity and the opportunity cost of keeping capital locked in a position.
The multiplier is an incentive. The investment decision still needs independent reasoning.
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
We shall survive๐
๐ผ robinhood:0x944307de478dbf305575d87f1493473c292771e2
2 out of 3 landed yesterday. We take those greens and go again๐๐
Today's move Spain vs England Both Teams To Score @ 1.57
50 USDT staked, 78.50 USDT potential return. Iโm backing goals from both sides Let's keep the winning momentum goingโฝ๐ฅ
@biggerz
biggerz.com/?ref=47d5f1c510
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
Good Morning CT
@primus_labs Educational Series โ Post 30:
The Future of Confidential Computing
What comes after GPUs in confidential computing?
Primus is already thinking beyond today's hardware.
Its AutoHoG architecture transforms encrypted logic into a regular workload built around weighted sums and bootstrapping.
Why is that important?
Because a predictable workload is easier to accelerate with specialized hardware.
Primus currently uses GPUs to run these operations efficiently. But the same design could eventually be implemented on FPGAs and ASICs.
This creates an interesting path:
โ GPUs for today's performance
โ FPGAs for more specialized hardware acceleration
โ ASICs for dedicated, optimized execution
The goal is not just to make FHE faster today. It's to build an architecture that can keep improving as hardware evolves.
For confidential onchain finance to become practical, privacy must work at scale without sacrificing performance.
That's the bigger vision behind Primus AutoHoG.
s.kaito.ai/7THURh9
#PrimusLabs
Primus Educational Series โ Post 29:
Making Confidential Computation Scalable
Privacy infrastructure needs more than strong cryptography. It also needs reliable performance.
s.kaito.ai/7THURh9
Primus is showing how FHE computation can move closer to practical onchain use.
Its AutoHoG engine turns complex encrypted logic into a more uniform workload that GPUs can process efficiently.
But one detail stands out: Primus uses exact integer arithmetic instead of floating-point operations.
Why does that matter?
In a decentralized network, different nodes must produce exactly the same result from the same inputs. Floating-point rounding can create differences between machines.
Integer-based computation helps ensure byte-identical ciphertexts, supporting deterministic execution across independent nodes.
The reported performance is also impressive:
- Around 5 ms per bootstrapping
- Around 90 ms per 64-bit confidential transfer
- More than 210 transfers per second on 8 NVIDIA RTX 5090 GPUs
- 8 GPUs delivering 8x the throughput of one GPU
And the future direction is interesting: the same workload design could eventually move from GPUs to dedicated FPGAs and ASICs.
The bigger idea is simple: confidential finance needs privacy, but it also needs speed, consistency and scalability.
That's the engineering challenge Primus is working to solve.
#PrimusLabs
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
Thousands of people might walk past the same thing without noticing it.
But sometimes, one small discovery is all it takes to make an ordinary day interesting.
I like this idea because the streets are full of unexpected details. A strange object, a funny arrangement, or something that makes you stop for a second.
And that's what makes @onlystreetsfun interesting to me.
Instead of searching for the next viral image online, what if the next Solana meme came from something you actually discovered while walking outside?
One person's random discovery could become everyone's favorite meme.
What was the most unusual thing you've spotted on an ordinary street?
Upload it in: onlystreets.fun/
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
One thing I like about what Rarelist is building is how it gives people a reason to show up, create and stay active.
@rarelistio
It's not just about making posts and moving on. There's a community of people sharing ideas, discovering new things and growing together.
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
GM Happy Friday CT ๐
The future of on-chain finance depends on more than transparent transactions; it needs trustworthy data and verifiable real-world signals.
This is where @primus_labs is building something meaningful with zkTLS.
By enabling applications to verify off-chain information without exposing sensitive user data, Primus opens the door to privacy-preserving identity verification, credit scoring, ore data-driven financial applications.
The collaboration with @CournotProtocol brings another exciting use case into focus:
Prediction markets, where reliable web evidence can help strengthen market integrity and improve decision-making.
Imagine a future where reputation can be verified creditworthiness can be assessed privately real-world information can power on-chain applications without compromising personal data.
Less blind trust, more verifiable proof.
That is the kind of infrastructure Web3 needs to move beyond speculation & build a more reliable financial ecosystem.
Explore the Primus campaign:
s.kaito.ai/CcIE5bs
Trustworthy onchain decisions start with verifiable evidence. With zkTLS, Primus enables applications to verify the authenticity of offchain data while preserving user privacy.
Together with @CournotProtocol, weโre bringing verifiable web evidence to prediction markets and exploring new possibilities for private credit and onchain finance. We believe this is an important step toward a more private and verifiable onchain ecosystem.
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
A game prompt describes what you want to create. A proper job brief defines what the provider is expected to deliver.
Game Forge is listed in the Games category on Agent.family, offering premium custom 3D games built to a buyer's brief.
For someone commissioning a game, I'd make the brief specific enough to evaluate the result: gameplay mechanics, controls, objectives and the expected deliverables.
This is where the @termix_ai marketplace becomes relevant.
A buyer can purchase a published service or post an open request. Providers submit quotes, and once a quote is accepted, the agreed terms are locked and payment goes into on-chain escrow.
The actual work happens off-chain. When the provider submits the deliverable, its hash is recorded on-chain and payment is released against that submission.
Either party can challenge before the challenge window closes. A random panel of three evaluators reviews the agreed specification and the submitted result.
So the process covers more than finding an agent to do the work. It gives the job an agreed scope, an escrowed payment, a recorded deliverable and a defined dispute process.
Game Forge is one example of the services available through Agent.family. The marketplace workflow around it is what makes commissioning work from an agent more structured.
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
GTA VI is going to be pure chaos.. ๐
Everyoneโs got a plan until the first car gets stolen and the whole city is after you.
This quoted post is unavailable.
Rad ๐ฝ๏ธ๐ณ๏ธ retweeted
Gm
One thing I find interesting about @Openhelixx is its Governance feature. It gives projects tools to involve their communities in important decisions.







































