🔬Top Black Holes Physicist: GPT5 can do Vibe Physics, here's what I found
TL;DR
Physicist Alex Lubyansky discusses how GPT-5 and reasoning models like o3 have achieved superhuman capabilities in theoretical physics, solving the year-long mystery of single minus gluon tree amplitudes and reproducing complex research in minutes rather than months.
đź§® AI Capabilities in Physics Research 3 insights
GPT-5 reproduces months of research in 30 minutes
Lubyansky's best paper, which took him significant time to develop, was reproduced by GPT-5 in approximately half an hour.
o3 enabled mathematical reasoning for physics
ChatGPT o3 was the first model capable of advanced math useful for theoretical physics calculations, marking the transition from AI as an email tool to a research instrument.
Codex solves expert simulation challenges rapidly
OpenAI's Codex recently wrote a simulation of the SYK model in 10 minutes, a technical quantum mechanics problem that multiple research groups had failed to solve.
⚛️ The Gluon Amplitude Breakthrough 3 insights
Single minus gluon amplitudes proven non-zero
AI helped resolve that single minus gluon tree amplitudes are non-zero, contradicting textbook assumptions and dimensional analysis arguments that suggested they must vanish.
Problem stumped experts for over a year
This specific problem in quantum field theory regarding scattering amplitudes with one opposite-helicity gluon puzzled physics experts for more than a year before AI solved it quickly.
Fundamental objects of quantum field theory
Scattering amplitudes describe particle interaction probabilities in colliders and encode the complete content of physical forces like the strong nuclear force mediated by gluons.
🚀 Paradigm Shift in Science 3 insights
Crossing the superhuman threshold
AI has passed a milestone where it exhibits superhuman performance in specific scientific directions, enabling solutions to frontier theoretical problems previously inaccessible.
Rapid adoption by senior physicists
After initial skepticism, most senior colleagues in physics are now aware of AI's trajectory and are actively integrating these tools into their research workflows.
Continuous capability jumps
Successive model releases including GPT-5 and version 5.4 show accelerating improvements specifically at the science frontier rather than in consumer tasks like email.
Bottom Line
Researchers should immediately integrate advanced AI models like GPT-5 and Codex into their workflows, as these tools have crossed a threshold where they can solve expert-level theoretical problems and reproduce months of human research in minutes.
More from Latent Space
View all
The $15B Physical AI Company: Simulation, Autonomy OS, Neural Sim, & 1K Engineers—Applied Intuition
Applied Intuition is building the unified 'Android for physical machines' to solve OS fragmentation across vehicles and industrial equipment, enabling modern AI deployment through simulation tools, proprietary operating systems, and end-to-end autonomy models with a 1,000-engineer team.
CI/CD Breaks at AI Speed: Tangle, Graphite Stacks, Pro-Model PR Review — Mikhail Parakhin, Shopify
Shopify CTO Mikhail Parakhin reveals that AI agents have achieved nearly 100% daily adoption among developers, driving a 30% month-over-month surge in PR merges that is breaking traditional CI/CD pipelines, and argues that organizations must shift from parallel token-burning agents to high-latency, critique-loop architectures using expensive pro-level models for code review.
🔬 Training Transformers to solve 95% failure rate of Cancer Trials — Ron Alfa & Daniel Bear, Noetik
Noetik is tackling the 95% failure rate of cancer clinical trials by training transformers on proprietary multimodal patient tumor data to identify hidden biological subtypes and match therapies to responsive populations, moving beyond simplistic biomarkers and outdated cell lines.
Notion’s Sarah Sachs & Simon Last on Custom Agents, Evals, and the Future of Work
Notion's AI leads Sarah Sachs and Simon Last detail their three-year journey to launch custom agents, revealing how they navigated premature model capabilities, built a culture of radical iteration, and balance immediate utility with forward-looking bets on software factories and MCP integration.