Turing Buys Failed-Startup Codebases for Training

AI firm Turing is buying codebases from failed startups, reportedly paying tens of thousands of dollars to repurpose them for model training. The acquired production-ready code is being cleaned, licensed or resold and used to improve coding and software-engineering capabilities for LLMs, creating a new intermediary data market. The practice raises legal and ethical concerns around IP, bankruptcy procedures, and data transparency.
Key Points
- 1Acquires production code: Turing purchases failed-startup codebases, reportedly for five-figure sums, to train models.
- 2Provides real-world production-ready engineering data that outperforms synthetic or incomplete open-source projects for training.
- 3Enables improved coding capabilities in LLMs but raises IP, bankruptcy, and transparency risks for practitioners.
Scoring Rationale
Practical industry trend offering actionable insights, but relies on company reports and lacks broad independent verification.
Sources
Public references used for this report.
Practice with real Retail & eCommerce data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Retail & eCommerce problems

