So far this year, I have taken a particular interest in Eric Weinstein's Geometric Unity. At this point it's probably the kingpin of unified theories from outside the academy. I have a whole blog about it.
I am also at an early stage of studying Jenny Nielsen's "Topological Unified Field Theory".
Finally, there is GIFT. This is the most interesting thing I've seen come out of the world of amateurs (non-physicists) using AI. Just to be clear, as of this writing, it does not hold together. But it's like a simulacrum of someone studying a string theory vacuum and getting predictions. The author even has machine learning code trying to approximate compactification metrics! Given the improvement in AI chronicled here over the past few years, it's possible that GIFT could become something fully rigorous, if the author keeps at it.
One way that could happen, is if via incremental change, it evolved into genuine string theory. "What part of the phenomenologically relevant string landscape is closest to GIFT?" is a legitimate question. Formally it might be closest to something downstream of this paper. However, the author also started with some "numerological" or pure-number intuitions, and it's not so clear what the upshot of that aspect will be.
The multitude of independent researchers now using AI is a step forward for that whole demimonde, the way that Vixra was, when it first appeared. At the same time, in the West it's the era of DOGE cuts to state-funded research, and populist skepticism regarding experts of all sorts. From the perspective of the academy, it might seem like an intellectual apocalypse in which actual science is being drowned out by AI-generated nonsense. (On X, "String King" suggests that China and Europe will be the centers of good science for the foreseeable future.)
However... At one point, when she was still alive, Marni Sheppeard wrote that she and various physics dissidents that she knew, seemed to be converging on some common ideas. This was dubbed a "crackpot condensate". In today's world in which millions of diverse independent experiments in human-AI collaboration are being carried out, there's a serious chance that AI will play a role in finally elevating something true out of the babel of would-be theories of everything.
P.S. I am also engaged in some pure-math studies with T.L., which I will post about when we're ready.
