Looped transformers are iterated filters, researcher says, citing ideas years older than new paper
docmilanfar · x · 2026-10-08
Responding to the recent wave of "looped transformers" posts, researcher Milanfar points out the idea isn't new: looped transformers are essentially iterated filters, a concept several years older than the paper sparking the discussion. Another scholar cites a 2016 paper as where he first saw looping in neural nets — "what's old is new again."
Related event: Looped Transformer Hype Sparks Scholarly Deep-Dive Into Decades-Old Ideas(6 posts)→
More from Research
- LeJEPA lets you pretrain self-supervised vision models on your own data — randall_balestr · 2026-10-08
- LEDGER builds persistent 3D object memory from egocentric video, answers questions without rewatching — mangahomanga · 2026-10-08
- Applied Compute unveils On-Policy Self-Distillation as a step toward continual learning — ypatil125 · 2026-10-08
- Depth2Depth fuses DepthAnything with RealSense for dense, metric depth maps — chrismatthieu · 2026-10-08
- NavSafe-∞: photorealistic closed-loop benchmark exposes safety gap across 20 E2E driving policies — zhoubolei · 2026-10-08
- Sampling test: pplx-decider-v1.1 judges pretraining data quality out-of-the-box vs Fineweb-edu classifier — bo_wangbo · 2026-10-08