Fatima

[Fatima @FatimaSaud95]

𝗔 𝗹𝗼𝘁 𝗼𝗳 𝗽𝗿𝗼𝘁𝗼𝗰𝗼𝗹𝘀 𝗰𝗮𝗻 𝘁𝗲𝗹𝗹 𝘆𝗼𝘂 𝘄𝗵𝗮𝘁 𝗵𝗮𝗽𝗽𝗲𝗻𝗲𝗱

The harder problem is deciding and executing what should happen next

A condition is observed

Now something still needs to fetch the required information run the computation coordinate the workflow and return the result where it can be acted on

That gap is where @fermah_xyz comes in

Fermah is building Protocol Agency Infrastructure: a programmable execution layer that lets protocols turn verifiable conditions into complete workflows

At the center is Kernel Fermah's Protocol Agency Engine

Kernel can observe a condition execute the required workflow in a sandboxed environment and return an attested cryptographically verified result onchain without requiring a human to coordinate every step

So the idea goes beyond simple automation

It's about giving protocols the ability to move from:

From this happened to now execute what comes next

With Kernel underneath workflows can be composed across different protocols and applications from ZK proof generation through Froben to automatically resolved prediction markets through Flashcast Social

That's the part I find interesting:

Fermah is building the execution layer that can turn a verifiable condition into an entire sequence of actions

@7wealthh

[Fatima @FatimaSaud95]

𝗝𝗼𝗶𝗻𝗲𝗱 𝗙𝗲𝗿𝗺𝗮𝗵’𝘀 𝗳𝗶𝗿𝘀𝘁 𝗼𝗳𝗳𝗶𝗰𝗶𝗮𝗹 𝗔𝗠𝗔 𝗹𝗮𝘀𝘁 𝗻𝗶𝗴𝗵𝘁

It was genuinely good to hear from the team directly and get a better understanding of what they’re building

What stood out to me was how openly they addressed the community’s questions the answers gave me more context around the product the vision and the thinking behind what @fermah_xyz is building

That made the AMA feel more useful than just another community session I left with a much clearer idea of where Fermah is heading

And honestly it made me even more curious about what they’re building next

@vanishree_rao @7wealthh

[Fatima @FatimaSaud95]

@zeeharison Exactly if validators don’t agree speed doesn’t matter much

[Fatima @FatimaSaud95]

𝗪𝗵𝗮𝘁 𝗶𝗳 𝗭𝗞 𝗽𝗿𝗼𝗼𝗳 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻 𝗯𝗲𝗰𝗮𝗺𝗲 𝘀𝗼 𝗶𝗻𝘃𝗶𝘀𝗶𝗯𝗹𝗲 𝘁𝗵𝗮𝘁 𝗯𝘂𝗶𝗹𝗱𝗲𝗿𝘀 𝗯𝗮𝗿𝗲𝗹𝘆 𝗵𝗮𝗱 𝘁𝗼 𝘁𝗵𝗶𝗻𝗸 𝗮𝗯𝗼𝘂𝘁 𝗶𝘁?

That’s the direction @fermah_xyz is taking with Froben

Generating ZK proofs can be resource intensive

There’s compute to coordinate
Proving infrastructure to manage
And workloads that need to be routed to the right resources

Fermah turns that into a shared infrastructure layer

Froben is its Universal Proof Market connecting protocols that need proofs with the compute required to generate them

But the part I find most interesting is what happens behind the request

Kernel orchestrates the workflow:

match the prover → generate the proof → verify it → settle the result

No separate system for every step
No need to build the entire proving workflow from scratch

And this is already running on mainnet

Froben supports any proof system any chain and any VM with Kernel handling the proof generation workflow underneath

That changes the role of proving infrastructure

The complexity doesn’t disappear

It moves underneath the application where it can be handled as infrastructure instead of becoming another problem for every builder

And that’s what makes @fermah_xyz interesting to me

Because when infrastructure works well enough people stop thinking about the infrastructure

They just build

And that’s exactly the kind of shift ZK needs

@7wealthh

[Fatima @FatimaSaud95]

𝗧𝗵𝗲 𝗵𝗮𝗿𝗱𝗲𝘀𝘁 𝗽𝗮𝗿𝘁 𝗼𝗳 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗽𝗵𝘆𝘀𝗶𝗰𝗮𝗹 𝗔𝗜 𝗶𝘀𝗻’𝘁 𝗷𝘂𝘀𝘁 𝘁𝗵𝗲 𝗿𝗼𝗯𝗼𝘁

It’s also about the data that teaches it how to operate in the real world

@PrismaXai describes its approach through three connected pieces:

➜ Robots
➜ Data
➜ Intelligence

The connection starts with human interaction

Through teleoperation and human in the loop systems people can operate robots while real world interactions are collected as training data

And the data is more than just video

PrismaX focuses on robotics foundation model data that can pair video with action information while also capturing robot state such as joint positions and other sensor data

Why does that matter?

Because physical AI models need to learn from how actions relate to what is happening in the physical environment

That makes data quality a core part of the problem

By operating the full loop themselves PrismaX says they can see what separates data that helps models learn from data that quietly degrades their performance those lessons then shape how robots are operated how tasks are structured and how the next round of data is collected

That creates a real feedback loop:

Human interaction
→ Real world robotic data
→ Training and evaluation
→ Lessons from model performance
→ Better data collection
→ The next cycle

PrismaX is building around this cycle by connecting teleoperation robotics data human work and intelligence within physical AI systems

The interesting part is that the challenge isn’t simply making robots smarter

It’s building the systems that allow robots people and data to work together so each deployment can contribute to the next stage of learning

That is the role PrismaX is building as the service layer for physical AI

[Fatima @FatimaSaud95]

@kai_dranzerr @RialoHQ The supermodularity angle is what makes this really interesting

[Fatima @FatimaSaud95]

𝗪𝗵𝗮𝘁 𝗺𝗮𝗸𝗲𝘀 𝗭𝗞 𝗽𝗿𝗼𝗼𝗳𝘀 𝘀𝗼 𝗽𝗼𝘄𝗲𝗿𝗳𝘂𝗹?

You can prove that something is true without revealing the information behind it

Imagine you need to prove that you know a secret

You can prove that you know it without revealing the secret itself

That’s the basic idea behind zero knowledge proofs

But in real applications the proof still has to be generated from an actual computation and that can require significant compute as proving workloads grow

This is where Fermah comes in

Fermah Froben is a Universal Proof Market for ZK proof generation

A protocol can request a proof, while Kernel orchestrates the workflow around getting that proof generated and verified

The process can look like this:

➜ Request a proof
➜ Match the prover
➜ Generate the proof
➜ Verify the result

So ZK proofs provide a way to verify computations without revealing all the underlying information

Froben focuses on the infrastructure needed to coordinate the work behind generating those proofs

That becomes an important part of making proof generation practical as ZK workloads continue to grow

@7wealthh

[Fatima @FatimaSaud95]

𝗪𝗵𝗮𝘁 𝗵𝗮𝗽𝗽𝗲𝗻𝘀 𝗮𝗳𝘁𝗲𝗿 𝗮 𝗽𝗿𝗼𝗼𝗳 𝗿𝗲𝗾𝘂𝗲𝘀𝘁 𝗶𝘀 𝗺𝗮𝗱𝗲?

In a distributed proving network sending a proof request is only one part of the problem

The bigger challenge is coordinating that request with the compute needed to generate the proof

That’s where @fermah_xyz Universal Proof Market comes in

It connects two sides:

→ developers and protocols that need proof generation
→ compute suppliers providing hardware mostly GPUs

Fermah’s orchestration layer then matches proof requests to machines

This creates a clear separation between requesting a proof and finding the infrastructure needed to generate it

And that coordination matters

A proving network isn’t simply about having more GPUs available

It needs a way to connect proof demand with available compute and make proof generation faster more reliable and cost efficient

That’s the role Fermah is building around its Proof Market

Instead of every protocol having to manage proving infrastructure on its own the network connects proof requests with compute suppliers through an orchestration layer

The interesting part to me isn’t just the proving hardware

It’s the coordination layer that connects the request to the compute behind it

That’s what makes distributed proof generation practical at scale

@7wealthh

[Fatima @FatimaSaud95]

@melophile_279 @miketwinks @widyakrnwn The yield model stands out

[Fatima @FatimaSaud95]

@VoltLedger Sometimes staying out is the better trade

[Fatima @FatimaSaud95]

@VoltLedger A good reminder to focus on your own journey

[Fatima @FatimaSaud95]

@VoltLedger FOMO can turn a good setup into a bad trade

[Fatima @FatimaSaud95]

@VoltLedger @BloFin_Official The puzzle angle makes this one different

[Fatima @FatimaSaud95]

@VoltLedger Position size matters a lot in trading

[Fatima @FatimaSaud95]

@VoltLedger @binance The TradFi angle is interesting

[Fatima @FatimaSaud95]

@melophile_279 @stargirl_hills The art and learning combo works really well here

[Fatima @FatimaSaud95]

𝗧𝗵𝗲 𝗵𝗮𝗿𝗱 𝗽𝗮𝗿𝘁 𝗼𝗳 𝗭𝗞 𝗽𝗿𝗼𝗼𝗳𝘀 𝗺𝗶𝗴𝗵𝘁 𝗻𝗼𝘁 𝗯𝗲 𝘁𝗵𝗲 𝗽𝗿𝗼𝗼𝗳 𝗶𝘁𝘀𝗲𝗹𝗳

It’s everything that has to happen around it

A proof request comes in

The right prover needs to be matched the proof needs to be generated verified and ultimately settled

So what looks like a single computation is actually a sequence of coordinated steps

That’s where @fermah_xyz gets interesting

Fermah’s Froben is its Universal Proof Market built for ZK proof generation it supports any proof system any chain and any VM while Kernel orchestrates the full proof generation workflow

The important part is that Kernel isn’t just designed around ZK proving

Kernel is Fermah’s Protocol Agency Engine

Its broader idea is simple but powerful: when a verifiable condition is true a protocol should be able to do something with that information

Instead of stopping at:

this condition has been verified

the workflow can continue:

condition → action → verification → onchain result

Kernel can take that verified condition trigger the required workflow execute conditional logic in a trustless environment and return a cryptographically verified result onchain

That changes the role of infrastructure

It’s no longer only about providing the computation a protocol needs

It’s also about coordinating what happens before during and after that computation

Froben shows this idea through ZK proving

A proof request enters the workflow the proving process is coordinated the result is verified and the completed result can move back into the onchain environment

But the underlying architecture is broader than proof generation

Fermah is building toward protocols that can act not simply react

A smart contract can already respond when something happens onchain

The more interesting question is:

what happens when a protocol can observe a verifiable condition execute the required workflow verify the result and continue the sequence automatically?

That’s the part of Fermah I find worth watching

ZK proving is one application

The bigger idea is turning verified information into coordinated action

And that could become an important piece of how more autonomous onchain workflows are built

@7wealthh

[Fatima @FatimaSaud95]

𝗜 𝗹𝗼𝗼𝗸𝗲𝗱 𝗶𝗻𝘁𝗼 𝗮 𝗽𝗮𝗿𝘁 𝗼𝗳 𝗯𝗹𝗼𝗰𝗸𝗰𝗵𝗮𝗶𝗻 𝘀𝗰𝗮𝗹𝗶𝗻𝗴 𝘁𝗵𝗮𝘁 𝗱𝗼𝗲𝘀𝗻’𝘁 𝗴𝗲𝘁 𝗲𝗻𝗼𝘂𝗴𝗵 𝗮𝘁𝘁𝗲𝗻𝘁𝗶𝗼𝗻: 𝗵𝗼𝘄 𝗱𝗮𝘁𝗮 𝗺𝗼𝘃𝗲𝘀

A blockchain can have fast execution but that still depends on data reaching the nodes that need it

Traditional gossip can involve redundant transmission using more bandwidth and slowing propagation under load

This is where @get_optimum caught my attention

Instead of changing how a blockchain reaches consensus Optimum focuses on the propagation layer

Its mump2p protocol uses Random Linear Network Coding (RLNC) to encode blockchain data and send encoded packets through the network

The interesting part is that RLNC isn’t simply about making the network faster

It changes how the data is transmitted

Optimum reports 6–20x faster block and blob delivery and around 90–95% lower bandwidth usage compared with Gossipsub

And mump2p is designed to integrate with existing node setups through an API without requiring changes to blockchain consensus

That makes the problem more interesting to me:

scaling isn’t only about processing more data

It’s also about making sure the data can move through the network efficiently in the first place

[Fatima @FatimaSaud95]

𝗭𝗞 𝗽𝗿𝗼𝗼𝗳 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻 𝗶𝘀𝗻’𝘁 𝗷𝘂𝘀𝘁 𝗮𝗯𝗼𝘂𝘁 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗻𝗴 𝗮 𝗽𝗿𝗼𝗼𝗳

There’s a whole workflow that needs to happen around it

A verifiable condition is observed

The required workflow needs to be executed the computation needs to run, and the result needs to be delivered and settled onchain

That’s where @fermah_xyz gets interesting

Fermah is building Protocol Agency Infrastructure with Kernel at the center as its Protocol Agency Engine

Kernel can turn a verifiable condition into an executable workflow run conditional logic in trustless containers, and deliver cryptographically verified results back onchain without asking a human to coordinate each step

For ZK this powers Froben Fermah’s Universal Proof Market

Froben supports any proof system any chain and any VM while Kernel orchestrates the full proof generation workflow

What I find interesting is that Fermah isn’t only focused on the computation itself

It’s building infrastructure that makes workflows executable composable and autonomous

That’s the shift: from a protocol reacting to a condition to a protocol being able to act on it

@7wealthh

[Fatima @FatimaSaud95]

𝗪𝗲𝗯𝟯 𝘀𝗰𝗮𝗹𝗶𝗻𝗴 𝗶𝘀𝗻’𝘁 𝗷𝘂𝘀𝘁 𝗮𝗯𝗼𝘂𝘁 𝗵𝗼𝘄 𝗳𝗮𝘀𝘁 𝗮 𝗯𝗹𝗼𝗰𝗸𝗰𝗵𝗮𝗶𝗻 𝗰𝗮𝗻 𝗽𝗿𝗼𝗰𝗲𝘀𝘀 𝗱𝗮𝘁𝗮

There’s another part that matters just as much:

how efficiently can that data move across the network?

Blocks transactions and data blobs need to propagate between nodes

Optimum focuses on making this data movement more efficient by reducing redundant transmission and improving bandwidth efficiency

That’s where @get_optimum focuses its infrastructure

Its mump2p protocol uses Random Linear Network Coding (RLNC) to encode blockchain data into combinations that can be forwarded and reconstructed even when some pieces are lost along the way

This approach is designed to reduce redundant traffic while improving latency and bandwidth efficiency

What I find interesting is that mump2p integrates with existing node setups through an API without requiring changes to blockchain consensus

Optimum is focusing on a part of blockchain infrastructure that is easy to overlook:

not just how data is processed but how efficiently it gets there