Testingthewaters: “I believe that within 6 months this line of research [online in-sequence learning] will produce a small natural-language capable model that will perform at the level of a model like GPT-4, but with improved persistence and effectively no “context limit” since it is constantly learning and updating weights.”
Hey, just to clarify that this prediction has not expired yet. Thanks for collecting these predictions though! From the original post:
Concretely, these are the developments I am predicting within the next six months (i.e. before Feb 1st 2026) with ~75% probability:
And for those interested these are recent papers that continue to chart the trend:
Hey, just to clarify that this prediction has not expired yet. Thanks for collecting these predictions though! From the original post:
And for those interested these are recent papers that continue to chart the trend:
[test time training family]
https://test-time-training.github.io/e2e.pdf (probably closest to the prediction but unsure how it scales to GPT-4 levels)
[tiny recursive model/hierarchical reasoning model family]
https://arxiv.org/pdf/2510.25741
https://arxiv.org/pdf/2512.14693
[other similar things]
https://arxiv.org/pdf/2512.24695
Ah sorry. Edited.