June 2026, Adaption research grant, teaching at Mannheim, and three Arbiter investigations

This month, SimPPL is honored to be named among the inaugural research grantees by Adaption. The grant will support our ongoing efforts in building adaptive AI systems that scale reliable information access, directly powering and refining the technologies that drive Arbiter.

Dhara Mungra, our co-founder, wrapped up her first semester teaching AI Product Development at the University of Mannheim. Students in the course built public-interest AI prototypes focused on journalism, social media research, and online information ecosystems, demonstrating the potential of developer-led ecosystem health.

In our first Arbiter investigation, we analyzed 16,604 posts across X, YouTube, and Bluesky to trace the narratives surrounding Hantavirus outbreaks. Discourse was heavily driven by outbreak response and crisis management narratives, with the largest attention spikes tied to a cruise-ship outbreak and subsequent public health monitoring efforts.

What Arbiter Found: Public health responses to travel-related risks emerged as the highest-engagement narrative, generating 133.9M interactions across platforms. X generated the largest engagement spikes, driven by breaking developments around outbreaks, transmission concerns, and expanding geographic impact. YouTube emerged as the primary hub for crisis management discussions, emergency response coverage, and explanatory content. Public health guidance and prevention information attracted significantly less attention than outbreak and transmission narratives, with the fastest-growing spikes tied to the MV Hondius cruise-ship outbreak between May 3–7.

Our second investigation explored community reactions to Anthropic's reported $900B valuation. When AI becomes a market story, funding milestones, revenue projections, and valuation comparisons dominate engagement across platforms rather than product capabilities.

What Arbiter Found: Twitter/X generated the majority of engagement through finance and technology influencers, including accounts such as unusual_whales, APompliano, and PolymarketMoney. LinkedIn and Facebook hosted more investor-focused discussion, with investor/speculation framing appearing in 74% and 82% of sampled posts respectively, while YouTube contained the highest share of technical commentary. The most widely shared content focused on valuation milestones, revenue growth, and market projections rather than product capabilities. Among the 120 highest-engagement posts analyzed, 24 contained market-projection or valuation claims.

Lastly, we studied how immigration enforcement disputes in Baltimore turned into broader ideological narrative battles online. Across platforms, a small number of incidents became focal points for wider ideological arguments around public safety, policing, and political identity.

What Arbiter Found: Immigration enforcement dominated discussion volume, but conflict-oriented narratives such as police-community clashes and public-safety framing generated the highest engagement. Twitter/X drove the vast majority of viral activity, with more posts crossing both the 1k and 5k interaction thresholds than any other platform in the dataset. The largest engagement spikes followed reports of arrests and subsequent confrontations involving law enforcement, particularly around the May 4 incident that became the highest-engagement event in the study period. Corrective information and fact-checking played a limited role in the most-engaged discussions within the top 40 posts analyzed. The top 5 posts alone accounted for 69% of total interactions.