Block unveils a new voice model that runs locally and sounds eerily human

Block has introduced a new voice model that TechCrunch AI describes as sounding eerily human. According to the report, published October 24, 2026, the model can run locally on a MacBook M4 Max.

The report credits Block with the development and with the claim of local execution on consumer laptop hardware, specifically the M4 Max chip. The characterization of the speech output as “eerily human” comes from the publication’s own description, per the headline framing.

The supplied material does not elaborate further on the model’s technical specifications, its intended applications, or Block’s stated purpose for the release. The report centers on two points: the sound quality of the synthetic voice and its ability to operate without reliance on larger server infrastructure, running instead on a standard high-end laptop.

No additional detail from the source accompanies these two claims. The item is brief, framed as a notable development in voice synthesis technology, with TechCrunch AI as the sole cited outlet.

This bulletin reports only what the source states, that Block has produced a voice model, that it is described as sounding eerily human, and that it has been demonstrated running locally on a MacBook M4 Max.