What I Read This Week…

President Trump signs executive order to overhaul federal procurement rules, OpenAI releases o3 reasoning model, and a brief history of federal funding for American universities

Watch All-In Episode 224

Read our Short Dive: The Trump Administration’s Fiscal Strategy

Read our Interview with Jason about the Evolution of Media

Caught My Eye…

• President Trump's recent executive order "Restoring Common Sense to Federal Procurement" launches the largest overhaul of federal procurement rules in over forty years. The order gives officials 180 days to simplify the 2,000-page Federal Acquisition Regulation that dictates how the government buys products and services. This reform is part of Trump's push to cut government red tape, which started with his January executive order "Unleashing Prosperity Through Deregulation." Under the new plan, officials will preserve only the rules that are required by law or truly necessary for making government procurement work more effectively. The order includes a sunset feature where non-required rules automatically expire after four years unless someone actively decides to keep them. Companies that sell to the government can expect fewer hoops to jump through, especially when selling regular commercial products. The changes should make it easier for new businesses to compete for government contracts by reducing the amount of paperwork and requirements. This procurement reform follows other recent Trump actions affecting government contractors, like ending diversity requirements, canceling the contractor minimum wage increase, and removing climate-related rules – all aimed at making federal procurement more efficient.

• OpenAI released their o3 reasoning model, showing state-of-the-art performance in math and coding and native integration with tools. OpenAI enabled the model to autonomously decide when and how to use all of ChatGPT's tools during its reasoning process, allowing o3 to seamlessly perform actions like searching the web, running Python code, reading files, examining images, and generating images within one session. Despite improved reasoning, o3 was observed to hallucinate about 33% of the time, double that of o1 (16%). Some users also observed that o3 occasionally over-uses tools or gets distracted by the tools, performing redundant web searches or calculations even for straightforward questions. With o3's ability to autonomously use tools, we're seeing AI models transition from assistants to agents that can decide to use basic tools to solve problems. o3's launch received more measured media coverage compared to earlier model unveilings like GPT-4, as people have become desensitized to new AI model announcements.

• Federal funding for American universities has increased over the last 75 years, driven by the twin pursuits of excellence and technological primacy. American universities began to receive consistent, large-scale federal funding during World War II with the Manhattan Project, which transformed academic research into a national security asset. Initially, policymakers expressed reluctance toward large-scale government involvement in higher education, but the imperative to develop superior technology overcame these reservations. The Cold War expanded the partnership between the federal government and universities after the 1957 Sputnik launch triggered fears of American technological inferiority. Congress responded with the National Defense Education Act, increasing support for science and engineering research. By the late 1960s, federal funding constituted approximately 73% of all university R&D spending. President Johnson's 1965 executive order further democratized this approach by spreading research funding beyond elite institutions to "build up excellence wherever it is found." A pivotal development came in 1980 when bipartisan legislation allowed universities to patent and profit from innovations developed with federal funds. This market-oriented reform energized the relationship between government funding and practical applications, leading to breakthroughs in medicine, computing, and other fields. By 2023, American universities were spending $60 billion annually in federal research money – thirty times the amount from the early 1950s. This growth stemmed from the expansion of research agencies like NSF and NIH, the space race, periodic funding surges, and the increasing complexity and cost of modern scientific inquiry. But federal funding always hinged on universities demonstrating their ability to convert research and knowledge production into technological and economic primacy.
  
  

Other Reading…

• Jerome Powell’s Outlook for the Economy and Monetary Policy (Federal Reserve Board)

• Paul Atkins Confirmed as New SEC Chair (The Governance Beat)

• Figma Files for U.S. IPO (Reuters)

• Google Makes History With Antitrust Losses (NYTimes)

• YouTube Looks to Creators and Their Data to Win in the AI Era (Bloomberg)

• Marketers Are Putting More Content Creation and Quality Control in the Hands of AI (WSJ)

• A Practical Guide to Building Agents (OpenAI)

• Bank of America, Tether, and Circle Battle to Shape U.S. Stablecoin Rules (The Block)

• Larry Fink’s 2025 Chairman’s Letter to Investors (BlackRock)

• The Most-Cited Papers of the Twenty-First Century (Nature)

On X…