OSINT 🧵: A method @AricToler, @riley_mellen, I and others used to find sources in our recent coverage of Jack Teixeira (JT). This piece didn't quote members of Thug Shaker Central (TCS), because it focused on JTs posts on another Discord server, but... nytimes.com/2023/05/06/us/…
... we still spoke to members of TCS for it. To find new members, we started with JT's and other members of TCS's Steam accounts which we knew from prior reporting, incl looking for Discord users on Steam. We had Discord usernames from screenshots TCS members released themselves:
Then, using a website that archives Steam, we got lists of each of their current and friends and plotted a network graph, where nodes are Steam accounts and there are edges between our starter set of accounts (red nodes) and each of their friends. Names removed for privacy.
We focused on Steam accounts that were friends with a high number of our starter accounts (remember the starter accounts belong to users we had reason to believe were in TCS at some point or other) and then looked for the person behind these Steam accounts on social media
This turned out to be a pretty good method for finding members of TCS! Also ChatGPT sped up making the network graph 👇
• • •
Missing some Tweet in this thread? You can try to
force a refresh
NEW: We worked with the artificial intelligence company Synthetaic to detect and analyze the Chinese balloon in satellite imagery. This allowed us to track the movement of the balloon itself, not just its expected path based on weather projections nytimes.com/interactive/20…
For the analysis we worked with @Synthetaic's founder @CoreyJaskolski, relied on Synthetaic's AI, high school physics and geometry and something called air parcel trajectory modeling, which models how a particle being blown by air moves.