gxqwb

Research notes on quantum computing, AI infrastructure, blockchain systems and high-performance computing — and what they change in everyday life.

Joined on Oct 29, 2026

gxqwb

Oct 31

Analog In-Memory Computing Versus the Digital NPU: The Energy Per MAC Argument

#ai #hardware #energy #architecture

5 min read

gxqwb

Oct 31

The Reticle Wall: Why Packaging Became the Real Fab Constraint

#hardware #semiconductors #ai #infrastructure

5 min read

gxqwb

Oct 31

800 Volts Versus 48: Why the Rack Power Standard Is Breaking

#infrastructure #hardware #datacenter #energy

5 min read

gxqwb

Oct 31

LPDDR6 Versus HBM: The Memory Fight Moves to the Edge

#ai #hardware #performance #mobile

5 min read

gxqwb

Oct 30

Six Years in the Queue: Why Datacenters Are Building Their Own Power Plants

#infrastructure #cloud #energy #operations

5 min read

gxqwb

Oct 30

The Migration Nobody Budgeted For: Post-Quantum Cryptography in Production

#security #cryptography #devops #infrastructure

5 min read

gxqwb

Oct 30

Twenty Millijoules: What Event-Driven Sensing Changes About Always-On Devices

#hardware #iot #embedded #machinelearning

5 min read

gxqwb

Oct 30

Cheap Cells, Heavy Racks: Where Sodium-Ion Storage Actually Wins

#energy #hardware #climate #infrastructure

5 min read

gxqwb

Oct 30

The Bandwidth Wall: Your Inference Bill Is Set by Memory, Not FLOPS

#ai #performance #hardware #cloud

5 min read

gxqwb

Oct 29

Event Cameras Leave the Lab: Neuromorphic Vision in Warehouses, Cars and Factories

#hardware #robotics #iot #machinelearning

4 min read

gxqwb

Oct 29

Proving Without Revealing: zkVM Throughput and the Real Cost of On-Chain Identity

#blockchain #security #programming #webdev

5 min read

gxqwb

Oct 29

Co-Packaged Optics Is Eating the Copper Backplane

#hardware #infrastructure #networking #performance

4 min read

gxqwb

Oct 29

Serving LLMs at 12 Watts: The Edge-Inference Cost Curve Nobody Budgeted For

#ai #edge #performance #architecture

5 min read

gxqwb

Oct 29

Logical Qubits Are Getting Cheap: A Developer's Map of the Quantum Error-Correction Stack

#quantum #hardware #research #ai

6 min read

DEV Community

gxqwb

Badges

Writing Debut

Analog In-Memory Computing Versus the Digital NPU: The Energy Per MAC Argument

The Reticle Wall: Why Packaging Became the Real Fab Constraint

800 Volts Versus 48: Why the Rack Power Standard Is Breaking

LPDDR6 Versus HBM: The Memory Fight Moves to the Edge

Six Years in the Queue: Why Datacenters Are Building Their Own Power Plants

The Migration Nobody Budgeted For: Post-Quantum Cryptography in Production

Twenty Millijoules: What Event-Driven Sensing Changes About Always-On Devices

Cheap Cells, Heavy Racks: Where Sodium-Ion Storage Actually Wins

The Bandwidth Wall: Your Inference Bill Is Set by Memory, Not FLOPS

Event Cameras Leave the Lab: Neuromorphic Vision in Warehouses, Cars and Factories

Proving Without Revealing: zkVM Throughput and the Real Cost of On-Chain Identity

Co-Packaged Optics Is Eating the Copper Backplane

Serving LLMs at 12 Watts: The Edge-Inference Cost Curve Nobody Budgeted For

Logical Qubits Are Getting Cheap: A Developer's Map of the Quantum Error-Correction Stack