Hardware and Digital Design

This section is about what happens below the abstraction boundary — the level where software meets silicon, where digital becomes analog, and where understanding the machine deeply enough produces capabilities that no amount of high-level programming can match.

Open Hardware and Trust

Open Hardware traces the revolution: Project IceStorm cracking Lattice FPGAs, formal verification via SymbiYosys catching bugs that simulation never would, hobby RISC-V cores that compete with commercial designs. The Precursor security device makes the strongest argument: an FPGA-based system where every line of the hardware description is auditable, because the epistemological relationship between user and device matters. Silicon Reverse Engineering approaches from the opposite direction — dissolving chips in acid and tracing transistors to discover what datasheets don't tell you, including bunnie's warning about SoC "dark matter" that's present, powered, and undocumented.

Gongkai And Shanzhai reveals a parallel innovation ecosystem in Shenzhen where IP sharing operates through social norms rather than legal frameworks, producing $12 phones and seasonal product cycles that Western startups can't match. The bridge between gongkai and Western open source — legal reverse engineering, custom scripting languages to avoid plagiarism — shows how different cultures solve the same problem of making technology accessible.

The Hardware Design Space

FPGA Design is the section's most technically rich article: thinking in parallel (no instruction pointer, everything runs simultaneously), formal verification proving designs correct for all possible inputs, ring oscillators as thermometers, time-to-digital converters with 10-picosecond resolution, and tiny GPUs that ray-trace at 60fps on $5 chips. The connection to Demoscene — designing the hardware itself as the ultimate creative constraint — is explicit.

Semiconductor Scaling debunks node numbers ("7nm" has no physical referent) and proposes replacements: GMT (gate pitch, metal pitch, tiers) and LMC (logic density, memory density, connectivity). The key insight: progress continues through 3D integration and new materials, but the marketing metric makes it look like a countdown to a wall. Analog Computing argues for revival: modern CMOS can build single-chip analog computers that solve differential equations in milliseconds at microjoule energy costs, and the hybrid analog-digital architecture may be the next frontier for problems with the right structure.

What Connects Outward

The hardware section bridges to Compiler Bootstrapping and Supply Chain Trust through the question of what you can trust when you can't inspect the silicon. To GPU Pipeline Architecture in the Graphics section through the shared understanding of how parallel hardware actually works. To Information And Computation through Landauer's principle (the thermodynamic cost of computation lives at the hardware level). And to The Material Basis Of Civilisation through the titanium paradox — process economics as the binding constraint on what technology makes possible.