Claims before 2023 to be the next big thing in AI beyond deep learning will not pan out.

'Next big thing beyond deep learning' means an AI approach claimed, before 2023, to succeed or displace deep learning. Advances that are themselves deep learning (e.g. transformers, large language models, diffusion models) do not count as such a successor. True: as of 2023, no approach publicly claimed before 2023 to be the successor to deep learning had become widely recognised in the field as having displaced or superseded it. Mostly true: one such approach had gained notable but limited traction without being generally recognised as superseding deep learning. False: by 2023 a non-deep-learning paradigm claimed as the successor had become widely recognised as the dominant new AI approach.

Set before the research began.

1. E1 Predictions Scorecard, 2023 January 01 Rodney Brooks (rodneybrooks.com) · 1 January 2023 Own record

   “Whether neuro-symbolic becomes a real dominant approach is still up in the air. But it has a whiff of possibility about it.”

2. E2 Predictions Scorecard, 2023 January 01 Rodney Brooks (rodneybrooks.com) · 1 January 2023 Own record

   “There are two or three candidates bubbling up, but all coming out of the now tradition of deep learning.”

3. E3 Predictions Scorecard, 2024 January 01 Rodney Brooks (rodneybrooks.com) · 1 January 2024 Own record

   “A year ago I thought it would perhaps be neuro-symbolic AI, but clearly it is LLMs, and ChatGPT and its cousins.”

4. E4 Why neurosymbolic AI could be the next big thing | Fortune Fortune · 9 December 2024

   “There are many flavors of neurosymbolic AI, but those being deployed today mostly involve having the neural and symbolic components work alongside each other and interact when they need to, rather than being tightly integrated into a new kind of model—a more theoretical vision for now.”

An independent Fortune article from December 2024 (E4) still calls neurosymbolic AI something that "could be the next big thing". It says the versions in use then mostly run neural and symbolic parts side by side, and that a tightly integrated new kind of model was still "a more theoretical vision". That indicates no non-deep-learning successor had displaced deep learning by the end of 2023. Brooks's own scorecards point the same way: in early 2023 the emerging candidates came out of deep learning and neuro-symbolic was "still up in the air" (E1, E2), and in 2024 he named LLMs as the big thing (E3). The criteria count LLMs as deep learning, not as a successor to it. The evidence does not show any claimed successor gaining notable traction by 2023, so the claim resolves true rather than mostly true.