<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>The Quanta Space</title><description>Deep, practical writing on the technologies shaping what comes next: AI, agents, systems, cloud and the science underneath.</description><link>https://thequantaspace.com/</link><item><title>Agents need state machines, not longer prompts</title><link>https://thequantaspace.com/blog/agents-need-state-machines/</link><guid isPermaLink="true">https://thequantaspace.com/blog/agents-need-state-machines/</guid><description>When an LLM agent misbehaves, the reflex is to add another paragraph to the system prompt. A better fix is to make the agent&apos;s control flow explicit, so the model chooses between moves instead of inventing them.</description><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>Agents</category><category>llm</category><category>agents</category><category>architecture</category></item><item><title>The KV cache, explained from first principles</title><link>https://thequantaspace.com/blog/kv-cache-explained/</link><guid isPermaLink="true">https://thequantaspace.com/blog/kv-cache-explained/</guid><description>Why transformer inference gets faster after the first token, why long contexts eat GPU memory, and the back-of-envelope math you need to size a deployment.</description><pubDate>Tue, 10 Nov 2026 00:00:00 GMT</pubDate><category>AI &amp; LLMs</category><category>llm</category><category>inference</category><category>gpu</category></item><item><title>An edge-first reference architecture for a small SaaS</title><link>https://thequantaspace.com/blog/edge-first-saas-architecture/</link><guid isPermaLink="true">https://thequantaspace.com/blog/edge-first-saas-architecture/</guid><description>How far can you get with a TypeScript monorepo, serverless functions at the edge, and a managed database? Further than you&apos;d think. Here&apos;s the layout we&apos;re using, and where it bends.</description><pubDate>Mon, 02 Nov 2026 00:00:00 GMT</pubDate><category>Cloud &amp; Edge</category><category>saas</category><category>cloudflare</category><category>architecture</category><category>typescript</category></item><item><title>Idempotency keys: the API feature you&apos;ll wish you&apos;d shipped on day one</title><link>https://thequantaspace.com/blog/idempotency-keys/</link><guid isPermaLink="true">https://thequantaspace.com/blog/idempotency-keys/</guid><description>Networks fail in the worst possible place, after the server did the work but before the client heard back. Idempotency keys turn &quot;did my payment go through?&quot; into a non-question.</description><pubDate>Tue, 27 Oct 2026 00:00:00 GMT</pubDate><category>Systems</category><category>api</category><category>distributed-systems</category><category>reliability</category></item><item><title>Passkeys for SaaS: what actually changes in your auth stack</title><link>https://thequantaspace.com/blog/passkeys-for-saas/</link><guid isPermaLink="true">https://thequantaspace.com/blog/passkeys-for-saas/</guid><description>Passkeys remove the password, but not the account-recovery problem, the session layer, or your need for good audit logs. A practical tour of what to build and what to keep.</description><pubDate>Mon, 19 Oct 2026 00:00:00 GMT</pubDate><category>Security</category><category>auth</category><category>webauthn</category><category>passkeys</category></item><item><title>Qubits for software engineers</title><link>https://thequantaspace.com/blog/qubits-for-software-engineers/</link><guid isPermaLink="true">https://thequantaspace.com/blog/qubits-for-software-engineers/</guid><description>Skip the mysticism. A qubit is a vector, a gate is a matrix, and measurement is a weighted coin flip. Here&apos;s the mental model, with enough linear algebra to run it yourself.</description><pubDate>Thu, 08 Oct 2026 00:00:00 GMT</pubDate><category>Quantum</category><category>quantum-computing</category><category>fundamentals</category></item></channel></rss>