🥳🎉 "TiDAR: Think in Diffusion, Talk in Autoregression"—because why just talk to yourself when you can overcomplicate
#communication with a sprinkle of computational gobbledygook? Thanks to this paper, we can now awkwardly combine two processes that nobody asked for, giving a whole new meaning to "crossed wires" in the tech world. 🤷♂️💻
https://arxiv.org/abs/2511.08923 #TiDAR #ComputationalGobbledygook #CrossedWires #TechInnovation #Autoregression #HackerNews #ngated
TiDAR: Think in Diffusion, Talk in Autoregression
Diffusion language models hold the promise of fast parallel generation, while autoregressive (AR) models typically excel in quality due to their causal structure aligning naturally with language modeling. This raises a fundamental question: can we achieve a synergy with high throughput, higher GPU utilization, and AR level quality? Existing methods fail to effectively balance these two aspects, either prioritizing AR using a weaker model for sequential drafting (speculative decoding), leading to lower drafting efficiency, or using some form of left-to-right (AR-like) decoding logic for diffusion, which still suffers from quality degradation and forfeits its potential parallelizability. We introduce TiDAR, a sequence-level hybrid architecture that drafts tokens (Thinking) in Diffusion and samples final outputs (Talking) AutoRegressively - all within a single forward pass using specially designed structured attention masks. This design exploits the free GPU compute density, achieving a strong balance between drafting and verification capacity. Moreover, TiDAR is designed to be serving-friendly (low overhead) as a standalone model. We extensively evaluate TiDAR against AR models, speculative decoding, and diffusion variants across generative and likelihood tasks at 1.5B and 8B scales. Thanks to the parallel drafting and sampling as well as exact KV cache support, TiDAR outperforms speculative decoding in measured throughput and surpasses diffusion models like Dream and Llada in both efficiency and quality. Most notably, TiDAR is the first architecture to close the quality gap with AR models while delivering 4.71x to 5.91x more tokens per second.
arXiv.org
Top UK Radio Host Greg James Becomes Creative Director Of Podcast Festival
Greg James has joined the Crossed Wires Podcast Festival with Alice Levine, Dino Sofos.
Deadline
Lucius - Two of Us on the Run (The New Recording) (Official Lyric Video)
YouTube
Crossed Wires è ora disponibile - Gamernews.it
Crossed Wires è scaricabile da Steam
Gamernews.it
Annunciata la data d'uscita di Crossed Wires - Gamernews.it
Crossed Wires uscirà il 4 novembre su PC
Gamernews.it
‘My Dad Wrote A Porno’ Host Alice Levine Among Organizers Of New British Podcast Festival
Alice Levine & Dino Sofos are behind a Crossed Wires podcast festival with the organizer of Tramlines
DeadlineCROSSED WIRES – a comic by Iris Jay

Iris Jay | Patreon
creating Comics & Art
Patreon