A new AI coding challenge just published its first results — and they aren’t pretty | TechCrunch

▤ Summary

The article covers the first results from the K Prize, a new AI coding benchmark launched by Databricks and Perplexity co-founder Andy Konwinski through the Laude Institute. A Brazilian prompt engineer won the first round, but the score was only 7.5%, which highlights how difficult the benchmark is. The challenge is designed to test real-world programming ability using newly published GitHub issues and to reduce benchmark contamination. Konwinski argues that the poor performance shows how far current AI coding systems still are from the hype. The piece frames the K Prize as both a technical experiment and a public challenge to the AI industry.

▦ Linked Companies

Databricks

$1B+