RIPEMD-160 collision truncated to 48 bits

The requested collision is in ../collision.json. Interpret each 0x string as hexadecimal bytes, not literal UTF-8 text.

Input	Full RIPEMD-160 digest
0x19bef10000000000	902e5fc86be47b1e2a17de2aab65d4e765623a05
0x1022730100000000	902e5fc86be4a99253faddc650510cb85e77233d

Both digests begin with 902e5fc86be4 (six bytes, or 48 bits). The inputs are distinct. This establishes the requested truncated collision for lambda = 24; the full digests differ.

Attributable computational evidence

../find_collision.py enumerated eight-byte little-endian integers from zero, using Python hashlib.new("ripemd160", data) and retaining the first six digest bytes. It found the pair after 24,322,577 evaluations, in approximately 35.77 seconds in this environment.

A separate invocation of Node.js v24.9.0 recomputed both full digests using crypto.createHash('ripemd160'), decoded the JSON's hex strings into bytes, checked distinctness, checked the exact four JSON keys and parameter values, and compared the first 12 hex characters. All checks passed. The runtime APIs are documented by Python hashlib and Node.js crypto; the numerical evidence above comes from the local executions, not these documentation pages.

Limits

These are local computational results. Python and Node.js can both use OpenSSL, so the second check is a separate execution and interface, not an independent cryptographic implementation. SIMD verification was not available or executed here; its result remains unobserved. No claim of a collision in the full 160-bit digest is made.