John Sambrook

Owner and founder of Common Sense Systems, an AI-first Theory of Constraints consulting and software company located in Kirkland, Washington, USA.

[John Sambrook @CommonSenseTOC]

@Athena_Lioness Can't think of a better investment of a parent's time than doing a good job of educating their children.

[John Sambrook @CommonSenseTOC]

@mwfoutch1 Perfect. Very slick setup! 73!

[John Sambrook @CommonSenseTOC]

@grok @sytelus Got it. So in the future, we might see Microsoft branding models with other qualifiers, like “Sell” for a model focused on selling, or “Writer” for a model trained on creative writing, etc.

[John Sambrook @CommonSenseTOC]

@Athena_Lioness Understand. Glad you're feeling a little better. Thank you.

🇺🇸🫡

[John Sambrook @CommonSenseTOC]

@elonmusk Let’s make sure it’s not similarly over in the United States. 🇺🇸

[John Sambrook @CommonSenseTOC]

@mwfoutch1 What does it mean when you say a short 20m band? Narrowed usable, bandwidth, or something else?

[John Sambrook @CommonSenseTOC]

@BrianRoemmele I managed a software team that wrote execution trace disassemblers. One of our target architectures was the Gmicro 'TRON' line of microprocessors.

You're welcome.

[John Sambrook @CommonSenseTOC]

@fdempster @FrankBr05713205 I think I had an H19, which was an ordinary 24x80 CRT, if memory serves. Perhaps it was a kit version of the Zenith Z19.

[John Sambrook @CommonSenseTOC]

@NoPhule @FrankBr05713205 Yes, very much so. Thank you.

[John Sambrook @CommonSenseTOC]

@FrankBr05713205 I remember them well. Great products. One of the first that I built was the IM-18 VTVM. I had it and used it for decades.

[John Sambrook @CommonSenseTOC]

Grok build with /loop is scary good.

I'm driving continuous improvement on a large software project that involves searching a complex product surface, deducing and building an agent API to the product.

My job is reduced to drinking coffee, watching the output, and nodding my head in approval, as if I'm actually keeping up with its findings.

Identities die hard....

[John Sambrook @CommonSenseTOC]

@when_to_fold_em @SemiAnalysis_ Perhaps he has an insight you do not?

[John Sambrook @CommonSenseTOC]

@wu_s_john @SemiAnalysis_ Yes, but not tomorrow.

[John Sambrook @CommonSenseTOC]

@ollama @grok Compare and contrast OpenJarvis with Hermes. I have an RTX 3090 on Ubuntu 24.04 running with llama.cpp right now and I'm running Qwen3.6-27B at 3 bit quant and MTP - what are my options?

[Pavlo Molchanov @PavloMolchanov]

🚀 Self-speculation brings 6.75x real speedup for LLM generation with SGLang inference!

Same model drafts future tokens in Diffusion mode → then verifies them in AR (causal) mode. One model and one KV cache. Just different attention masks.

Thanks to perfect alignment, we get 2× longer acceptance lengths than MTP techniques (Eagle-3, MTP, dFlash).

We run 2 forward passes… but the 2× higher acceptance means we break even - and with zero overhead from extra drafter, KV cache, or LM head that comes with MTP - those are not free.

Last week we released Nemotron-Labs-Diffusion + Tri-mode LLMs! We did continued pre-training on Ministral-3 models by switching attention patterns (block causal <> bidirectional). Result: one model that runs AR mode, Diffusion mode, and Self-Speculation.

Diffusion mode already shows high benchmark accuracy - excited to see what happens when someone beats left-to-right acceptance! 🔥

Github: https://t.co/Zqbw3KcAyF
Paper: https://t.co/rp86A7D0xJ
SGLang inference: https://t.co/uTgZPALEJl
Try the models on HF: https://t.co/1zStcCCWPi

[John Sambrook @CommonSenseTOC]

Not sure that my HOA will allow it, but I have to try.

[RossRadio @cqcqcqdx]

Marconi Poldhu wireless station antenna.

[John Sambrook @CommonSenseTOC]

@BrianRoemmele “No ring, no write.”

[John Sambrook @CommonSenseTOC]

@DanKalavros @esrtweet Good argument.