cd /news/ai-chips/from-data-to-silicon-play-connect-4-… · home › topics › ai-chips › article
[ARTICLE · art-170605] src=benoitrostykus.com ↗ pub= topic=ai-chips verified=true sentiment=↑ positive

From data to silicon: play Connect-4 against a chip

A single-cycle chip that plays Connect-4 was learned end-to-end from a dataset of simulated games for under $10 of compute, according to the author's account published November 16, 2026. The chip uses about 14,000 logic gates, measures 0.8 × 0.8 mm on a two-decade-old 130 nm process, runs at 25 MHz and consumes a few hundred picojoules per move — matching a modern CPU Connect-4 engine while using two to three orders of magnitude less energy. The demonstration supports the author's argument that automating chip design for arbitrary ML tasks is closer than commonly assumed, contingent on sharply lowering the cost of producing such specialized ASICs.

read1 min views1 publishedNov 16, 2026

Nov. 16th, 2026

    What if you could take a dataset of (input, output) pairs collected for a task of interest (image recognition, next-word prediction etc.),
    and turn that into a custom chip designed specifically for this task? In other words, learning an AI model where the model *is* the silicon?

    I believe this extreme specialization of hardware is where we could soon find ourselves, assuming the cost of making such chips can be dramatically reduced.
    This hinges on two prerequisites:

Think about it: if you could cheaply print custom ASICs for every non-trivial piece of computation running over and over, why wouldn't you? Soon, AI building AI will do that anyway.

    While I have no idea how to build semiconductor factories, let me convince you that automating the design of chips for arbitrary ML tasks
    is closer than you might think.

    Here is a single-cycle chip that plays Connect-4, autonomously learned end-to-end for less than $10 of compute, just from a dataset of simulated games.
    It has about 14k logic gates. On a two-decade-old 130 nm manufacturing process, it measures 0.8 × 0.8 mm, runs at 25 MHz
    and consumes a few hundred picojoules per move: about as fast as a Connect-4 engine running on a modern CPU, while using two to three orders of
    magnitude less energy.

    Try playing against the chip and see the individual wires light up within a clock cycle as the signal propagates through it:
── more in #ai-chips 4 stories · sorted by recency
── more on @connect-4 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/from-data-to-silicon…] indexed:0 read:1min 2026-11-16 · —