Uploaded March 2026 | Updated September 2026, 2 weeks ago
Have we hit the “data wall?"
According to SandboxAQ CEO Jack Hidary, today’s major language models have already trained on most of the publicly available internet.
But the next generation of AI breakthroughs won’t come from web data.
Defense systems, advanced batteries, materials science, and biopharma require proprietary, physics-based datasets built from scratch — not what's left to scrape from the internet. That’s the shift from large language models (LLMs) to large quantitative models (LQMs).
This goes beyond language: AI that's built on new data, new science, and deep partnerships with industry and the Pentagon.
Read Jack's Op-Ed about Quantitative AI on the Wall Street Journal or our website:
wsj.com/opinion/america-needs-ai-that-can-do-math-392b223b
sandboxaq.com/post/america-needs-ai-for-the-real-world
#sandboxaq #lqms
Have we hit the “data wall?"
According to SandboxAQ CEO Jack Hidary, today’s major language models have already trained on most of the publicly available internet.
But the next generation of AI breakthroughs won’t come from web data.
Defense systems, advanced batteries, materials science, and biopharma require proprietary, physics-based datasets built from scratch — not what's left to scrape from the internet. That’s the shift from large language models (LLMs) to large quantitative models (LQMs).
This goes beyond language: AI that's built on new data, new science, and deep partnerships with industry and the Pentagon.
Read Jack's Op-Ed about Quantitative AI on the Wall Street Journal or our website:
wsj.com/opinion/america-needs-ai-that-can-do-math-392b223b
sandboxaq.com/post/america-needs-ai-for-the-real-world
#sandboxaq #lqms
