AI Futures from BestWorld’s Jeremy Lichtman
Forecasts by Jeremy’s bots here—> His YouTube main site here —>
Today I ask the prediction bot whether it thinks there will be a ceasefire in Ukraine in Q1, 2025. This is a tricky topic, and a fairly short timeframe: it’s really hard to predict what will happen next in entrenched conflicts (and I’ve done poorly myself on scored questions on this topic in the past two years). Let’s see how well the bot does.
Bonus: I show a couple of unusual Neanderthal hand tools from my collection.
Are large language models (LLMs) truly intelligent, or are they just glorified Markov Chains? 
Dive into this engaging video essay that explores the inner workings of LLMs and their surprising similarities to simpler probabilistic models. In this video, we break down how LLMs like GPT use vast amounts of data and sophisticated neural networks to generate human-like text. By comparing these models to Markov Chains, we uncover the mathematical principles behind their ability to predict and construct coherent sentences. Are LLMs merely advanced pattern matchers, or do they demonstrate something closer to genuine intelligence? Whether you’re curious about the mechanics of AI, the philosophical questions surrounding artificial general intelligence (AGI), or simply want to better understand how these models are shaping the future, this video is for you. Learn how LLMs function, their strengths and limitations, and what their capabilities say about the nature of intelligence itself.
• Yes, ChatGPT helped write this description.
Link to the paper mentioned about detecting LLM-generated text
Link to the Markov Chain code on GitHub
Ben Wilson is an expert builder of LLM-based prediction bots. 
In this fascinating interview, I talk to him about how just how good prediction bots are right now 
, what some of their blind spots are 
, and how people are working to improve them in the future 
. This video will be of great interest to anyone interested in forecasting, AI agents, prediction markets and more. 
Ben has built a library of tools that can assist people who are interested in building forecasting bots of their own: https://github.com/Metaculus/forecasting-tools
He is also involved in building templates that allow people to quickly set up forecasting bots that can compete in Metaculus’ quarterly AI forecasting competition: https://github.com/Metaculus/metac-bot-template
A quick rundown of the specific topics that I am currently tracking daily with the prediction bot. These are largely either economic or geopolitical questions. Two will be dropping off my list soon.
Will the highly contentious US Steel and Nippon Steel merger ultimately go through? 
Our AI prediction bot takes a closer look at one of the most politically charged business deals of the decade.
The fate of this merger is a question that the Metaculus competition team that I am part of have been tracking for a while.
This video explores the twists and turns of the proposed merger, which has become a flashpoint in an election year where both political parties fiercely courted the union vote. After the Committee on Foreign Investment in the United States (CFIUS) failed to reach a decision, President Biden stepped in to block the deal, citing national interests. Now, the companies are appealing the decision in the courts, setting the stage for a dramatic showdown in 2025.
Can the merger overcome the political and legal hurdles, or will it remain a casualty of election-year politics?
This is a must-watch for anyone interested in geopolitics, corporate strategy, or how AI is being used to predict real-world events. Join us as we unpack this high-stakes battle and explore the future of steel, profits, and politics in 2025.
As promised, the full summary that the prediction bot produced is as follows. Clearly, at least some of the news that it used was a few days out of date!
The merger between Nippon Steel and U.S. Steel faces significant regulatory challenges, primarily due to the strategic nature of the steel industry and the involvement of a foreign entity acquiring critical assets. Historical precedents suggest mergers in such industries undergo rigorous scrutiny, particularly by CFIUS, often resulting in blocked or altered deals. Factors such as the $14.9 billion deal size, political dynamics favoring domestic production, and opposition from entities like the United Steelworkers union add to the complexity. Furthermore, the current political climate under President Biden increases uncertainty, as there is potential for executive intervention. However, Japan’s status as a close ally, along with possible economic benefits, lobbying efforts, and long-term viability concerns of U.S. Steel, might positively impact the outcome. Changes in U.S. economic or geopolitical policies towards Japan and successful negotiations addressing regulatory concerns could also improve the merger’s chances.
