It is wild how quickly the alliances at the very top of the tech food chain can shift. One minute OpenAI seems to have a lock on the ecosystem, and the next, reports surface that Apple is pivoting to Google’s Gemini for its foundation models. I was glued to the screen watching this breakdown from the hosts at Forward Future Live.
The creators of this episode brought together a panel of four industry heavyweights to move past the rumors and explain where the actual value in AI is being generated right now. From redefining education to solving “impossible” math problems, the insights shared by these experts paint a picture of a sector that is maturing rapidly.
The New Era of Engineering
Austen Allred, the founder of Gauntlet AI, shared a fascinating look at modern education. He detailed a program that costs students nothing but requires 98th-percentile aptitude and 100-hour workweeks. The expert notes that software engineering has fundamentally changed; what used to take a team six months can now be built by a single AI-empowered engineer in 48 hours. This shift allows his graduates to land jobs paying over $200,000 because companies are realizing a small, elite team using AI is more effective than a large, traditional one.
Unlocking Business Data
Aaron Levie from Box discussed a critical bottleneck: unstructured data. Most corporate knowledge is locked inside PDFs, invoices, and contracts. The CEO explained how “Box Extract” uses AI to turn this messy information into structured context that agents can actually use. He emphasized that while models are getting smarter, businesses still need “scaffolding” around them to ensure accuracy and security before letting agents handle sensitive tasks like responding to RFPs.
📌 Captain YAR’s Key Takeaways
Here are the most mind-blowing insights these innovators shared:
- Math is Solved: Tudor Achim from Harmonic revealed that AI recently solved three famous “Erdos problems” that had stumped mathematicians. By using “auto-formalization,” the AI doesn’t just guess; it converts proofs into code to verify they are 100% correct.
- The Bubble Myth: Ara Kharazian from Ramp shared hard data contradicting the idea of an AI bubble. While the Census Bureau estimates 18% adoption, his spend data shows 46% of companies are paying for AI, with retention rates hitting 80%.
- The 2028 Prediction: The experts at Harmonic believe that at the current rate of progress, an AI will solve a Millennium Prize problem, one of the hardest math questions in history, within the next four years.
This discussion proves that while the headlines focus on corporate drama, the real revolution is happening in the trenches of engineering and mathematics. You really should watch the full panel to hear the details directly from the source.