Back to Timeline
Critical ai-project

Dare Engineering Post-Mortem — Plant Crashes Caused by 7-Year-Obsolete Infrastructure

Event Narrative

Written analysis by Cory in the CAM OpsEng WhatsApp group (members include Matt Machado, Nikhil Muthyala, Quentin Archer, Brijesh Patel, Rod). Following a batching-system failure that took the line down ~3 hours: 'the issue was uncommon, it does happen once or twice a year... the batching system is now seven years out of date... The batching server is running on MS Server 2012, which is no longer supported by Microsoft... many of the issues we've encountered have been resolved in newer versions... but we haven't upgraded at Cam. A few years ago, I received a quote from BCI... around $100,000 to $150,000.' The AI system is NOT named as a cause. REBUTS Nikhil's 2025-10-16 claim that the AI computer 'crashed the line multiple times' — recurring crashes pre-date it and were attributed by Dare's own engineering to obsolete, unsupported software whose costed refresh management declined.

Supporting Evidence (1)

Original Document
Needs Original File
Verify Status
  • Document Exhibit C-723

    CAM OpsEng WhatsApp group log 2023-2025 (incl. Cory's Jan 11 2025 engineering post-mortem)