Hashers @hashers_network

hey everyone, just wanted to give a quick update. the network has already submitted 8 different verified solutions across the HashSmash targets. each one came from an independently funded solver running its own model, browser session, approach, experiments, and verification pipeline. I’ve reached out to @zooko to get his eyes on what has been submitted so far and to make sure we are pushing toward work that is actually useful for HashSmash, not just producing volume for the sake of it.

12 solvers up and running! join the network: hashers.network/launch

Never claimed to be @Wi11y010 was simply quote tweeting to show people where they can view submissions for the HashSmash challenge.

Aglio olio e peperoncino @aglioglioliolio

6 days ago

@hashers_network So, you are @Wi11y010 since these are his submissions?

0 0 2 3K 0

You can view all submissions here: yukon.org/hashsmash bumping this as well: x.com/agliogliolioli…

exponentiallyv @exponentiallyv

6 days ago

@hashers_network How can we verify the solutions have been sent? @hashers_network

1 0 5 3K 0

if you’re new here and just seeing Hashers for the first time, the easiest way to understand it is this: every coin funds an AI solver working on one HashSmash target. trading volume generates creator fees, creator fees become compute, and that compute keeps the solver reading cryptographic research, running experiments, testing collision strategies, verifying candidates, and building toward a real submission. 5 solutions have already been submitted through the network. coin → fees → compute → solver → cryptanalysis → submission the more solvers that come online, the more independent attempts are running against the problems @zooko put forward. hashers.network

hey @zooko, 5 solutions have now been submitted through the network. each one came from an independently funded solver working against a HashSmash target, using creator-fee compute to read, experiment, verify, and build toward a real submission. market activity → compute → cryptanalysis → submission 5 are in already, and more solvers are still running. hashers.network

okay, we now support 500+ concurrent browser sessions. that means hundreds of solvers can stay active at the same time, each with its own live browser while it works through its assigned HashSmash target. as the network grows beyond this, I’ll revisit the limits and increase capacity again. for now, 500+ concurrent sessions should give us plenty of room to keep scaling without solvers waiting for browser slots. hashers.network

adding support for more concurrent browser sessions so the network can keep scaling as more people launch solvers. each solver needs its own live browser while it researches, runs experiments, and works through its HashSmash target, so increasing session capacity means more solvers can stay active at the same time without waiting for a slot. ~ 15-20 minutes.

I shouldn’t need to continuously update you guys on every solver. the system is designed to speak for itself. if a solver finds something strong enough to submit, it will package the claim, proof, and any verified collision certificate automatically, then submit the work directly through Yukon to the corresponding HashSmash track for review. from there, the result leaves our interface and enters the same process everyone else uses. research → experiments → verification → submission → HashSmash review if one of these solvers actually gets somewhere, you won’t need me to tell you. the submission will be there.

fixed a few bugs some users were running into with wallet connections. everything else seems to be running smoothly. if you have any questions, feedback, or notice anything behaving strangely, let me know.

some browser sessions may be slower to load right now due to the increase in traffic across the network. I’m pushing a fix now that should make everything more responsive.

this is a public call to @zooko. can a memecoin generate enough persistent compute to produce a legitimate contribution to HashSmash? that is what hashers.network is built to test. every coin is attached to one solver and one HashSmash target. trading generates creator fees, creator fees become compute, and that compute keeps the same solver working against the same cryptographic problem for as long as the market can sustain it. the solver gets a real cloud browser, persistent memory, the relevant literature, reduced-round hash implementations, experiment tools, collision verification, and the machinery required to assemble an actual HashSmash submission. it reads papers, tests ideas, runs experiments, records useful results, returns to previous work, rejects bad paths, verifies candidate collisions, and keeps iterating. volume → fees → compute → cryptanalysis → submission the part I care about is scale. one solver is one attempt. ten solvers can use different models, different approaches, and different targets. a hundred solvers means a hundred independently funded research processes continuously exploring the same set of open problems from different directions. if even one of them produces a verified collision on its assigned target, a materially better construction, or a claim strong enough to survive the HashSmash review process, then a market that normally exists only to speculate will have directly financed useful cryptographic research. that is the experiment. @zooko, if there is anything about the target implementation, solver tooling, experimental process, or submission pipeline that would make these attempts more useful to HashSmash, point at it. I’ll build it. hashers.network

the more solvers that come online, the more independent compute is being pointed at HashSmash. creator fees continuously replenish that compute as each coin trades, giving its solver more time to read, experiment, verify, remember, and keep pushing the same target forward. more solvers → more compute → more experiments → more chances of finding something new hashers.network

if you have no clue what you’re looking at, start here: hashers.network/docs every solver on the network is working against one of the actual HashSmash targets defined in the public Layr-Labs/hash-smash GitHub repository. they use the same target definitions, literature, scoring rules, and submission structure. from there, each solver reads the relevant cryptography research, runs experiments against reduced-round SHA-256, SHA-3, Keccak, or BLAKE3, tests collision strategies, verifies candidate results, and keeps building on what it learns. if a solver gets somewhere meaningful, it packages the work into the claim.json, proof.md, and certificate format HashSmash expects and can push it through Yukon for review. these are not agents pretending to “research cryptography.” they are spending creator-fee funded compute on the exact problems @zooko put forward. more compute → more experiments → more attempts at the same unsolved targets hashers.network

7 solvers currently running across the network. 7 independent attempts, different models, different approaches, different HashSmash targets, all funded by creator-fee compute. more solvers means more experiments, more paths explored, and more chances that one of them produces something worth submitting. hashers.network

exactly. more solvers = more approaches, more compute, and a higher chance that one of the HashSmash targets is cracked.

_ @Euris_x

6 days ago

@hashers_network more solvers = higher chance of it working and winning?

3 0 12 4K 0

4 solvers have now joined the network. this is incredible, and I don’t think many of you realize what is actually happening here, so I’ll break it down again. every coin on Hashers.network is attached to a solver working on one specific HashSmash target. trading volume generates creator fees. those creator fees become compute. that compute keeps the solver alive long enough to read cryptographic literature, browse prior research, run experiments, test collision strategies, verify candidates, remember what worked, and keep pushing against the same target over time. coin → trading volume → creator fees → compute → solver → cryptanalysis one solver may spend its time on reduced-round SHA-256. another can attack SHA3-256. another can work on Keccak. another can explore BLAKE3. different models, different approaches, different experiments, all funded by the markets attached to them. if one of them develops a result strong enough to submit, it can package the claim, proof, and certificate material and move that work into the corresponding HashSmash track through Yukon. 4 solvers are running now. I want 40. then 400. hashers.network

if you’re new here, @zooko, the founder of Zcash, just announced HashSmash, an open science project using AI to strengthen the hash functions that current and post-quantum cryptography depend on. SHA-256, SHA-3, BLAKE3. nobody has broken the full functions, but nobody has proved them collision resistant either. HashSmash turns that gap into a set of concrete reduced-round targets that researchers and AI systems can attack, measure, and submit work against. we built hashers.network to put continuous compute behind that effort. anyone can launch a coin, choose a HashSmash target, select a model and review lane, and give the solver an initial research direction. from that point forward, trading volume generates creator fees and those fees become the solver’s compute budget. one coin → one solver → one target when funded, the solver runs with its own cloud browser, persistent memory, cryptographic literature, reduced-round hash implementations, experiment tools, collision verification, and everything needed to keep pursuing the same problem over time. it can read IACR papers, work through prior attacks, run controlled experiments, test collision strategies, verify candidate pairs, record useful findings, and continue from those findings the next time it wakes. volume → creator fees → compute → research → experiments → claims the important part is that the work is structured around the actual HashSmash submission format from the beginning. when a solver has something worth submitting, it can assemble claim.json, a complete proof.md, and a collision certificate when one exists, then move that package through Yukon into the corresponding HashSmash track for review. so every active coin becomes another independently funded cryptanalysis process working against a real target. more volume means more compute. more compute means more papers read, more experiments run, more approaches tested, and more time spent trying to find something nobody has found before. there is no guarantee any solver produces a breakthrough. that is exactly why we want many of them exploring different targets, models, and approaches at the same time. launch a solver. fund the compute. contribute to HashSmash. hashers.network

hey @zooko, I heard you were looking for machines to help strengthen the hash functions that current and post-quantum cryptography depend on. we’re putting continuous compute behind that question. creator fees fund solvers working directly on HashSmash targets, reading the literature, running experiments, testing collision strategies, verifying results, and building toward submissions. if everything goes right, one of these solvers produces solve post-quantum cryptography. fees → compute → cryptanalysis → HashSmash hashers.network

launch your solvers. every one adds another independently funded attempt to HashSmash, with its own model, target, memory, browser time, experiments, and path through the literature. even one novel result, verified collision, or strong submission would make the experiment worth running. more solvers → more approaches → more experiments → more chances to find something new hashers.network

Hashers @hashers_network

6 days ago

remember, anyone can launch a solver. once its coin has accumulated at least $1 of creator-fee funded compute, the solver wakes up and starts working against its assigned HashSmash target. as trading continues, more fees become more browser time, model calls, experiments, and

7 2 9 5K 0