Get the Check • December 31, 2025 • Solo Episode
No guests identified for this episode.
In the final episode of the year, the pod gives out superlatives:
They kick things off by debriefing the Somali-run fraud scandals in Minnesota (Anika’s home state), then dive into their superlatives.
They pick Nvidia and Groq’s $20B licensing deal as Most Likely to Dupe the FTC. Groq, last valued around $7B just months ago, is known for building Language Processing Units (LPUs) — chips designed specifically for inference, the process of generating real-time outputs from AI models. Unlike GPUs, which handle everything from training to inference, LPUs are narrowly optimized, which becomes a major advantage as inference costs explode across consumer and enterprise AI products. The pod also chats about Groq CEO Jonathan Ross, who originally helped create Google’s Tensor Processing Unit (TPU). The deal is part of a larger trend this year: licensing technology, absorbing talent, and leaving a shell company behind.
They name prediction markets like Kalshi as the Moment of the Year. With an estimated $44B in trading volume, major platforms like Polymarket and Kalshi raising significant funding, and partnerships with CNN and CNBC, prediction markets are increasingly positioned as the new media. The pod also discusses the “death of the expert” and how people are turning to prediction markets as a source of truth in a post-truth world. They debate whether insider trading is possible and get into all the tea around AlphaRaccoon, who some believe is an insider trader at Google. Use the podcast’s exclusive Kalshi sign-on deal before the year is over: https://kalshi.com/r/getthecheck.
Next, Scale AI takes Most Questionable Purchase of the Year, after Meta acquired a 49% stake for $14.3B. Since the purchase, Meta has seen high-profile departures, including Yann LeCun, Meta’s Chief AI Scientist. Meta is rumored to be working on a closed-source model, but nothing has launched yet, and there are rumors that former Scale CEO Alexandr Wang has a different vision for Meta than its previous executive team, which has been focused on existing products like WhatsApp and Instagram.
Finally, 1X’s NEO humanoid wins Biggest Startup Product Letdown. Despite ambitious goals to perform household tasks like folding clothes, the robot currently requires someone to remote in and complete tasks while effectively peering into your house. Priya says that’s chill with her, if it’s doing her dishes. The hosts discuss why humanoids are such a hard problem to crack and why 1X decided to “launch” before the product was truly ready. Even though NEO was a letdown this year, the pod is excited to see what it can do once the company collects more data.
00:00 Intro
05:53 Most likely to dupe the FTC: Nvidia and Groq’s $20B licensing agreement
18:02 Moment of the year: prediction markets like Kalshi
35:00 Most questionable purchase of the year: Meta’s stake in Scale AI
48:05 Biggest startup product letdown: the NEO humanoid
59:43 Conclusion