How Long Until Someone Decompiles macOS?
Happy Wednesday.
The current thing in tech and business is the 722 mathematics papers written by OpenAI.
Today’s Lineup
- Hudson Institute Distinguished Fellow & WSJ Columnist Walter Russell Mead at 12:00 PM
- AG1 CEO Kat Cole at 12:30 PM
- Stuut Co-Founder & CEO Tarek Alaruri at 12:35 PM
- Revel Founder & CEO Scott Morton at 12:40 PM
- Healthleap Co-Founder & CEO Josiah Meyer at 12:45 PM
- Bonsai Robotics Co-Founder & CEO Tyler Niday at 12:50 PM
- OpenAI Product Manager Aarush Selvan at 1:00 PM
Run of Show
How Long Until Someone Decompiles macOS?
Adobe got decompiled. The apps are full featured clones of the Adobe products many people have been using for years. There are still a bunch of questions about how copyright applies here, and I wouldn’t be surprised if these get squashed pretty quickly, but it’s clear that any piece of software can now be reverse engineered nearly instantly with modern models.
The OpenAI math results show a similar story of the crazy spike in intelligence we see around tasks that can be run in fully virtual environments. Here’s the official OpenAI post and a good analysis by Josh Ganz that I enjoyed reading about it.
François Chollet, creator of ARC-AGI, broke down a theory of what this big jump in progress means for overall AGI progress and how these advancements generalize:
It definitely feels like a superintelligence to me. No human intelligence has been able to solve this many math problems at this level of difficulty in history, so it’s clearly something new. Terminology debates aside, there are some near term impacts that are worth working through.
Ben Thompson has been working through some macOS frustrations while he goes deeper into agentic AI, from Apple and a Hacker’s Future, he writes:
What I’m much more interested in is controlling the software layer myself. The fact of the matter is that with AI you can decompile almost all existing software — there is a revolution happening in gaming over the past few weeks, as game after game is decompiled to source and ported to any platform you wish — and you can write your own. That means my software that interacts with my agent in the way I want it to for everything; that’s way more exciting than praying Apple delivers the right API and that 3rd-party developers don’t suck.
There’s an obvious bullcase for Linux broadly, and on the desktop here, but I’m increasingly wondering if we might just see a full decompile of macOS hit the timeline. Don’t like that security features don’t automatically install? Patch that yourself. Don’t like that every new application you vibecode needs to be granted permissions via UI pop-ups? Change that behavior in a single prompt.
This actually feels like it could be good for the hacker’s future in the Apple ecosystem, although it calls into question a whole bunch of other issues with this new world where software is infinitely malleable. Regardless, it’s firmly the domain of hackers for the time being, so it will be fun to watch the open-source community duke it out with tech companies over where the new line is exactly.
Clip Spotlight: Rockefeller Capital Management CEO Greg Fleming is wary of managers who’ve put up great results but only lived through one cycle:
“Anybody who got into SpaceX early, that’s kind of the vision they have for every private investment: ‘Am I going to be up 30, 40, 50 times?’”
“This is not something that recurs on a regular basis.”
“We think these people have done well, but they’ve really only seen one cycle. So we want to be careful there.”
“The underlying driver is real. But there will be winners and losers.”
“Let’s look for managers that have actually seen something go wrong, because things go right and they go wrong.”
Headlines
OpenAI releases 722 math papers written by internal frontier model
OpenAI: Sharing AI progress in mathematics
Joshua Gans: OpenAI drops a bomb on maths
FT: SpaceX looks to raise $40bn to buy Nvidia chips
Anduril is investing $3.7B into Arsenal-2
WSJ: Nous Research Scores $90 Million to Bring Open-Source AI Assistant to Enterprises
The Economist: American investors’ trick to lower taxes: lose money
KTLA: DIY device turns your car window into a digital billboard
WSJ: American Fire Departments Are Short on People. AI Might Be the Answer.
FT: The taxman comes for China’s offshore riches
FT: Brussels looks to capture Big Tech through tax on large corporations
Ryan McEntush: Why Texas Is Making Data Centers Wait
WSJ: Moët Hennessy Directs AI to Sniff Out a Stubborn Problem in Winemaking
AutoGuide: Porsche Mission S Mid-Engine Super Sports Car to Sit Above 911
Huddle Up: How McLaren Racing’s CEO Made $108 Million Last Year
WSJ: Teens Who Can’t Put Down TikTok and Instagram Are Losing Their Sense of Time
Motor1: Bugatti Gets It: Screens Are Not Luxury
Posts of the Day
Special thanks to our sponsors: Ramp, Shopify, CrowdStrike, MongoDB, NYSE, Codex, Public, Console, Railway, Figma, and Cisco.