DigiNews

Tech Watch by Johan Denoyer

← Back to articles

OpenAI might have stolen another major proof

Quality: 8/10 Relevance: 9/10

Summary

Valerio Capraro's post claims OpenAI may have trained Astra on conversations related to Gromov’s soficity conjecture, implying unpublished human work was absorbed into a model and presented as a breakthrough. The piece frames this as a potential major scientific scandal with implications for attribution and AI data provenance.

🚀 Service construit par Johan Denoyer