Rónán!

dev

\ projects:

[Rónán! @ronanyeah]

@suidevelopers Revive this https://t.co/ZDoCOmjDty

[Rónán! @ronanyeah]

@juzybits like

[Rónán! @ronanyeah]

@juzybits Now to explore the galaxy

[Rónán! @ronanyeah]

@_smbrian Advanced typescript isn't much easier than rust anyway

[Rónán! @ronanyeah]

@A_cRyPtIc_one @Silverhard_ @suilendprotocol @steammfi @0xrooter This is a useful tool but since it does batches of 500, the BRICK airdrop would have required 1000+ transactions.

Another you reason you may not want to airdrop directly like this is to require recipients to manually claim to prove wallet liveness or community engagement.

[Rónán! @ronanyeah]

@b1ackd0g Almost as hard to get as post.shit

[Rónán! @ronanyeah]

@0xrooter I think Sui is currently at a slight disadvantage for this kind of experimentation because creating multiple coins from a single contract is not possible

[Rónán! @ronanyeah]

@0xd34th @healthydege I'm an expert at wasting resources

[Rónán! @ronanyeah]

@kostascrypto Pushed

[Rónán! @ronanyeah]

Stateless Probabilistic Minesweeper

Done! @kostascrypto

The SAT solving is implemented with the `varisat` Rust crate, and is provided to the webapp as WASM.

Code and demo links 👇

[Kostas Kryptos @kostascrypto]

Stateless Probabilistic Minesweeper

A relatively easy random challenge (or not so easy, I 'm actually curious):

Prize: Live X podcast between @themoveguy (creator of Move book) and myself to the first person who will implement the following:

In this Minesweeper variant:

The board’s mines are not fixed upfront and hidden cells have no remembered state between moves.

- Every time you reveal a cell or query a hidden cell, its state (mine or safe) is sampled fresh, based only on the currently revealed visible state.

- The number shown on any revealed safe cell reflects the actual number of adjacent mines consistent with the visible board and the current random sampling.

- The algorithm must ensure every sampled board state satisfies all revealed clues, without contradiction.

- No persistent assumptions or stored states exist for hidden cells; all sampling and consistency checks happen dynamically on demand.

- On each user's move, sample (roll) the mine/safe state of the chosen cell and any relevant neighbors freshly, constrained only by the currently revealed board.

- Ensure the sampled assignment of mines and safes across the board is globally consistent with all visible clues (numbers on revealed cells).

- Maintain logical consistency without remembering any hidden cell’s prior assignments.

- Support configurable per-cell mine probabilities and their conditional neighbor dependencies.

- Automatically deduce forced mines or safe cells when only a single consistent choice exists.

- Implement flood fill to reveal safe neighbors of zero-valued cells.

- Guarantee the first click is always safe, so we can have a game.

- Provide interactive gameplay in a website.

- Explain the algorithm clearly and provide a Python or JS script.

Bonus
- Integrate efficient algorithms to prune impossible assignments early during sampling.
- Explore approaches like constraint satisfaction problem (CSP) solvers or SAT solvers adapted for stateless sampling.
- Provide a solver assistant suggesting moves with minimal risk based on fresh sampling.
- Implement as a Sui smart contract, using Sui's Native Randomness and No Nautilus, MPC or ZK (no cheats).
- For math nerds: if that game is not possible mathematically, provide a proof.

[Rónán! @ronanyeah]

@levelsio In your case, the HTML is the state (alongside any global+local JS assignments in use)

[Rónán! @ronanyeah]

@kostascrypto On it 🛠️

[Rónán! @ronanyeah]

DEMO: https://t.co/qTZTW28Qq0
CODE: https://t.co/FhLBfsIUSG

[Rónán! @ronanyeah]

@kostascrypto These are reasons it is technically possible, not reasons it is a good product idea

[Rónán! @ronanyeah]

@redacted_noah cputrampler

[Rónán! @ronanyeah]

I use arch

[Moon Dragon @frozenaesthetic]

Share a piece of lore about yourself

[Rónán! @ronanyeah]

In progress

[Kostas Kryptos @kostascrypto]

Stateless Probabilistic Minesweeper

A relatively easy random challenge (or not so easy, I 'm actually curious):

Prize: Live X podcast between @themoveguy (creator of Move book) and myself to the first person who will implement the following:

In this Minesweeper variant:

The board’s mines are not fixed upfront and hidden cells have no remembered state between moves.

- Every time you reveal a cell or query a hidden cell, its state (mine or safe) is sampled fresh, based only on the currently revealed visible state.

- The number shown on any revealed safe cell reflects the actual number of adjacent mines consistent with the visible board and the current random sampling.

- The algorithm must ensure every sampled board state satisfies all revealed clues, without contradiction.

- No persistent assumptions or stored states exist for hidden cells; all sampling and consistency checks happen dynamically on demand.

- On each user's move, sample (roll) the mine/safe state of the chosen cell and any relevant neighbors freshly, constrained only by the currently revealed board.

- Ensure the sampled assignment of mines and safes across the board is globally consistent with all visible clues (numbers on revealed cells).

- Maintain logical consistency without remembering any hidden cell’s prior assignments.

- Support configurable per-cell mine probabilities and their conditional neighbor dependencies.

- Automatically deduce forced mines or safe cells when only a single consistent choice exists.

- Implement flood fill to reveal safe neighbors of zero-valued cells.

- Guarantee the first click is always safe, so we can have a game.

- Provide interactive gameplay in a website.

- Explain the algorithm clearly and provide a Python or JS script.

Bonus
- Integrate efficient algorithms to prune impossible assignments early during sampling.
- Explore approaches like constraint satisfaction problem (CSP) solvers or SAT solvers adapted for stateless sampling.
- Provide a solver assistant suggesting moves with minimal risk based on fresh sampling.
- Implement as a Sui smart contract, using Sui's Native Randomness and No Nautilus, MPC or ZK (no cheats).
- For math nerds: if that game is not possible mathematically, provide a proof.

[Rónán! @ronanyeah]

@kostascrypto @D4yv33d @themoveguy @grok Next challenge hint

[Rónán! @ronanyeah]

Was trying to find this tool and cannot believe this worked

[Rónán! @ronanyeah]

@uberboyo F1 drivers also seem to peak <30