corpus.blog
41554 blogs · 4,417,066 posts Claim your blog
Record · corpus.blog/posts/01a07d5d-86e9-710f-b9c2-b454895f134e

Speculation Is All You Need

Modal · published 19 June 2026

A record is what was published and who pointed at it. The text of the post is not held here: read it at the source, or at its Wayback capture.

Read the post at modal.com Wayback capture
0blogs citing
First cited
—
Most recent
—
Rank this month
not ranked
Links from this post
59
Words captured
3,764

Cited by

No blog we hold has cited this post yet.

Links from this post

Link
Host
https://z-lab.ai/
z-lab.ai
https://z-lab.ai/projects/dflash/
z-lab.ai
https://huggingface.co/modal-labs/Qwen3.5-397B-A17B-DFlash
huggingface.co
https://www.lmsys.org/blog/2026-06-15-next-generation-speculative-decoding-dflash-v2/
lmsys.org
https://huggingface.co/modal-labs/Qwen3.6-35B-A3B-DFlash
huggingface.co
https://huggingface.co/modal-labs/Qwen3.5-4B-DFlash
huggingface.co
https://huggingface.co/modal-labs/Qwen3.5-9B-DFlash
huggingface.co
https://huggingface.co/modal-labs/Qwen3.5-27B-DFlash
huggingface.co
https://huggingface.co/modal-labs/Qwen3.5-35B-A3B-DFlash
huggingface.co
https://huggingface.co/modal-labs/Qwen3.5-122B-A10B-DFlash
huggingface.co
https://modal.com/llm-almanac/token-timing-simulator
modal.com
https://modal.com/llm-almanac/
modal.com
Making FlashAttention-4 faster for inference Read at the source ↗
modal.com
Boosting multimodal inference performance by >10% with a single Python dictionary Read at the source ↗
modal.com
Host overhead is killing your inference efficiency Read at the source ↗
modal.com
https://modal.com/notebooks/modal-labs/charles-dev/nb-mFGHBCz7sQT9THdbkx7jat
modal.com
http://www.incompleteideas.net/IncIdeas/BitterLesson.html
incompleteideas.net
Autoscaling Autoresearch: Give your agents elastic GPUs on Modal Read at the source ↗
modal.com
https://en.wikipedia.org/wiki/Attention_Is_All_You_Need
en.wikipedia.org
https://bookdown.org/rdpeng/advstatcomp/rejection-sampling.html
bookdown.org
At the Intersection of LLMs and Kernels - Research Roundup Read at the source ↗
charlesfrye.github.io
https://modal.com/gpu-glossary/device-hardware/gpu-ram
modal.com
https://modal.com/gpu-glossary/device-hardware/streaming-multiprocessor
modal.com
https://hashrocket.com/blog/posts/materialized-view-strategies-using-postgresql
hashrocket.com
https://github.com/apoorvumang/prompt-lookup-decoding
github.com
Reinforcement learning is an infrastructure problem Read at the source ↗
modal.com
https://arxiv.org/abs/2404.19737
arxiv.org
https://arxiv.org/abs/2503.01840
arxiv.org
https://arxiv.org/abs/2602.06036
arxiv.org
https://fullstackdeeplearning.com/course/2022/lecture-8-teams-and-pm/
fullstackdeeplearning.com
https://newsletter.pragmaticengineer.com/p/why-is-meta-destroying-its-engineering
newsletter.pragmaticengineer.com
Lumon Industries Macrodata Refiner Read at the source ↗
lumon-industries.com
https://en.wikipedia.org/wiki/Goodhart%27s_law
en.wikipedia.org
How we achieved truly serverless GPUs Read at the source ↗
modal.com
https://gym.modal.dev/
gym.modal.dev
https://github.com/fergusfinn
github.com
https://doubleword.ai
doubleword.ai
https://fergusfinn.com/blog/economics-of-speculative-decoding/
fergusfinn.com
https://modal.com/gpu-glossary/perf/memory-bandwidth
modal.com
https://modal.com/gpu-glossary/perf/arithmetic-bandwidth
modal.com
'I paid for the whole GPU, I am going to use the whole GPU': A high-level guide to GPU utilization Read at the source ↗
modal.com
https://modal.com/gpu-glossary/perf/latency-hiding
modal.com
https://modal.com/gpu-glossary/perf/overhead
modal.com
https://modal.com/gpu-glossary/perf/compute-bound
modal.com
https://modal.com/gpu-glossary/perf/memory-bound
modal.com
https://modal.com/gpu-glossary/perf/roofline-model
modal.com
https://modal.com/gpu-glossary/host-software/cutlass
modal.com
https://github.com/modal-projects/specdec-calculator/blob/88776aa8177d14e9faf07a43ef33aa6ecf66d2ca/core/verifier.ts
github.com
https://github.com/modal-projects/specdec-calculator/blob/88776aa8177d14e9faf07a43ef33aa6ecf66d2ca/core/deepseek-v4.ts
github.com
https://github.com/modal-projects/specdec-calculator/blob/88776aa8177d14e9faf07a43ef33aa6ecf66d2ca/core/configs/DeepSeek-V4-Flash.json
github.com
https://github.com/modal-projects/specdec-calculator/blob/88776aa8177d14e9faf07a43ef33aa6ecf66d2ca/core/draft.ts
github.com
https://modal.com/llm-almanac/spec-dec-roofline
modal.com
https://modal.com/llm-almanac/spec-dec-roofline?view=optimal
modal.com
AdapTive-LeArning Speculator System (ATLAS): A New Paradigm in LLM Inference via Runtime-Learning Accelerators Read at the source ↗
together.ai
https://arxiv.org/abs/2505.18825
arxiv.org
https://hazyresearch.stanford.edu/blog/2025-09-28-tp-llama-main
hazyresearch.stanford.edu
https://www.recursive.com/articles/first-steps-toward-automated-ai-research
recursive.com
https://arxiv.org/abs/2411.00841
arxiv.org
https://modal.jobs
modal.jobs