Tech with Mak ✓

@techNmak · 📩 for collaboration · joined 30 Jul 2024

AI, coding, software, and whatever’s on my mind.

43 287Followers

880Following

5 429Posts total

357.2KViews on collected posts

Latest posts

Tech with Mak ✓@techNmak6 Sep

KV cache is one of the most important ideas in LLM inference, but it is often explained too casually. During autoregressive generation, a model produces one token at a time. Without caching, each new decoding step would repeatedly recompute key and value states for tokens the ht 1.4K views · 41 likes · 9 reposts · 2 replies Open on X →

Tech with Mak ✓@techNmak5 Sep

At some point in your career, you meet an engineer who looks lazy until you realize he is just playing a different game. You give him a problem that sounds like a day of work, he asks a few irritatingly obvious questions, searches the codebase for ten minutes, removes most of htt 207.6K views · 1.8K likes · 104 reposts · 25 replies Open on X →

Tech with Mak ✓@techNmak5 Sep

@techNmak I bet Nvidia can also send their own inference jobs to whichever machine in YOUR household has capacity. 576 views · 8 likes · 0 reposts · 1 replies Open on X →

Tech with Mak ✓@techNmak5 Sep

Genuinely one of the best engineering books ever written. https://t.co/vQCXjT45Kb 11.8K views · 371 likes · 18 reposts · 15 replies Open on X →

Tech with Mak ✓@techNmak5 Sep

@techNmak Sure will be an even bigger unlock for us to be able to sell our surplus compute network capacity for LN sats through a Data Vending Machine using @OpenAgentsInc Infrastructure! 325 views · 4 likes · 1 reposts · 1 replies Open on X →

Tech with Mak ✓@techNmak4 Sep

I put together a handbook on how attention works, from the Q/K/V equation up through how it actually runs in modern LLMs. Self vs. cross-attention, causal masking, multi-head attention, RoPE, KV caching, MHA/MQA/GQA/MLA, FlashAttention, worked through by hand, with citations htt 73.7K views · 2K likes · 301 reposts · 23 replies Open on X →

Tech with Mak ✓@techNmak4 Sep

NVIDIA just made all the spare AI compute in your house useful. Install PAIR on your PCs and Macs, and local AI agents can automatically send each inference job to whichever machine has capacity. NVIDIA Personal AI Router (PAIR) is a local inference router that lets multiple ht 61.7K views · 746 likes · 115 reposts · 30 replies Open on X →

Against accounts of the same size

7 posts from the last 90 days, next to the 10K–100K follower range. shown to more people than peers of the same size.

Median views11 843this account924median for 10K–100K

Reach, %27.36%this account3.62%median for 10K–100K

Engagement, %1.85%this account1.52%median for 10K–100K

Metric	This account	Median for 10K–100K	Ratio
Median views per post	11 843	924	12.8×
Reach (views ÷ followers)	27.36%	3.62%	7.56×
Engagement rate	1.85%	1.52%	1.21×

Others in this range → Compare with another account → How these benchmarks are built →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.