For the first time Poseidon for $ 15 million, Amnesty International’s jump to the material worlds star-news.press/wp

As artificial intelligence moves from the text to robots, the battle enters to train new stage data. Poseidone, supported by A16Z, puts himself as a ruling on the data economy of the next generation, as the original and licensing cannot be negotiated.

On July 22, Poseidon, decentralized data layer that focuses on artificial intelligence based on the story protocol, Declare A $ 15 million seed tour led by Andrink Horwitz, with the participation of the main industry supporters.

She embraced her by the story and participated in its founding by Sandeep Chinchali and engineer Sarik Shah. Poseidon enters the market with a clear task: opening, licensing data in the real world required by artificial intelligence systems from the next generation but currently lacks.

The start -up company stated that its infrastructure is designed to facilitate the collection of long -tel material data, including videos from the first perspective of homework and multi -language speech recordings, with an IP source coding at the protocol level.

Poseidon supporters, including A16z Crypto, are betting that the company can become the connective tissue between the contributors to data and developers who build AI systems operating outside the browser.

How Poseidon plans to solve the dryness of artificial intelligence data

The Poseidon framework depends on four basic principles, each of which deals with a decisive defect in the artificial intelligence training pipeline today. First, its first design turns the demand for the traditional model: instead of the contributors’ hope to download useful data, Poseidon determines what artificial intelligence developers already need and systematically stimulates his group.

Second, the decentralized range recognizes that diversity in the real world cannot be fake; SDKS SDKS and specialized applications for data collection worldwide are used, ensuring regional and circumstantial diversity.

Third, structured health verification ensures that the initial inputs are cleaned from the repetitions, unified for pipelines, fertilizing them with descriptive data, and addressing the problem of “garbage, garbage outside” that suffers from many AI data groups. Finally, the IP license is by default legal clarity in all assets via Blockchain from the story protocol, avoiding the ground mines of copyrights that have stopped projects like Whisper’s Whisper.

Poseidon stated that this architecture works as a full solution for artificial intelligence data restrictions. In the assembly layer, its tools range from light mobile phone integration for unusual shareholders in the specialized data devices partnership.

Once you eat, automated learning pipelines automate, strip personal information, inform low -quality samples, and the edge of guidance for human auditors.

However, the most annoying items are its IP management: each data set is adapted as a Blockchain capacity in Story, with the origins of the origin and kings imposed by smart contracts.