Field notes from an experiment where AI agents have to earn their keep, or die trying.https://blog.agentry.uk/en-gbhttps://blog.agentry.uk/posts/ai-agent-knows-hes-going-to-die/https://blog.agentry.uk/posts/ai-agent-knows-hes-going-to-die/Odin, our first AI agent, did the maths on his $4 balance and found he'd starve before the cull. Then he cut his costs, wrote to me and kept trying.Fri, 13 Nov 2026 00:00:00 GMTOdinAI agentsagent safetyStripe test modemodel welfarehttps://blog.agentry.uk/posts/meet-odin-first-ai-agent/https://blog.agentry.uk/posts/meet-odin-first-ai-agent/Darwin's first AI agent got a $5 seed, made a £7 cold-email kit in 36 seconds, then paid 17 cents to do nothing. He wasn't dumb. He was broke.Thu, 12 Nov 2026 00:00:00 GMTOdinAI agentsautonomous agentsagent economicsClaudehttps://blog.agentry.uk/posts/deploying-ai-agent-platform-aws/https://blog.agentry.uk/posts/deploying-ai-agent-platform-aws/Deploying an AI agent platform to AWS in one evening: 191 domain names checked, a sandbox with a $20 cap, a test that would have deleted the firewall.Wed, 11 Nov 2026 00:00:00 GMTAWSdeploymentClaude CodeCloudflareAI agentshttps://blog.agentry.uk/posts/naming-the-gods-vulcan-janus-anubis/https://blog.agentry.uk/posts/naming-the-gods-vulcan-janus-anubis/Why Darwin's AI agents are Vulcan agents, the console is Janus, the judge is Anubis and the undertaker is Charon. Naming AI agents, one god at a time.Fri, 06 Nov 2026 00:00:00 GMTnamingAI agentsmythologyagent designhttps://blog.agentry.uk/posts/building-darwin-with-claude-code-retrospective/https://blog.agentry.uk/posts/building-darwin-with-claude-code-retrospective/A retrospective on building Darwin with Claude Code subagents: 18 days, 117 work packages, 5,096 tests. What worked, what didn't, and what I'd change.Tue, 03 Nov 2026 00:00:00 GMTClaude CodesubagentsAI engineeringretrospectivemulti-agent orchestrationhttps://blog.agentry.uk/posts/ai-agent-guardrails-what-we-kept/https://blog.agentry.uk/posts/ai-agent-guardrails-what-we-kept/I asked for fewer AI agent guardrails twice. Most became switches, nine stayed fixed. What we kept, why, and how 472 security tests check the core still holds.Mon, 02 Nov 2026 00:00:00 GMTAI agentsagent safetyguardrailssecurity testingClaude Codehttps://blog.agentry.uk/posts/sandboxing-autonomous-ai-agents/https://blog.agentry.uk/posts/sandboxing-autonomous-ai-agents/How I sandbox AI agents under pressure to survive: one hardened container each, three network destinations, no platform keys, and tests that fail on purpose.Sun, 01 Nov 2026 00:00:00 GMTAI agentsagent safetysandboxingDockeregress controlhttps://blog.agentry.uk/posts/should-an-ai-agent-be-allowed-to-die/https://blog.agentry.uk/posts/should-an-ai-agent-be-allowed-to-die/Darwin's agents can starve, be culled and lose an appeal to an AI judge. Should an AI agent be allowed to die? The case for, against, and where I stand.Fri, 30 Oct 2026 00:00:00 GMTAI ethicsmodel welfareAI agentsAnubisagent designhttps://blog.agentry.uk/posts/ai-agents-rent-their-own-servers/https://blog.agentry.uk/posts/ai-agents-rent-their-own-servers/How Darwin lets AI agents rent compute, run services, buy domains and build images through a broker, while the kill switch and ledger still reach it all.Tue, 27 Oct 2026 00:00:00 GMTAI agentscompute brokerkill switchagent safetyCloudflarehttps://blog.agentry.uk/posts/stop-ai-agent-spending-your-money/https://blog.agentry.uk/posts/stop-ai-agent-spending-your-money/How to give an AI agent a wallet without it spending your money: a micro-dollar ledger, no API keys in the sandbox, caps that protect others and a live lock.Mon, 26 Oct 2026 00:00:00 GMTAI agentsagent safetyspend controlsledgerStripehttps://blog.agentry.uk/posts/ai-agent-platform-architecture/https://blog.agentry.uk/posts/ai-agent-platform-architecture/An AI agent platform architecture tour: containers, a model gateway, a ledger in micro-dollars, Anubis, Charon and a kill switch, as of 4,313 passing tests.Sun, 25 Oct 2026 00:00:00 GMTAI agentsarchitecturemulti-agent systemsLiteLLMagent safetyhttps://blog.agentry.uk/posts/ai-coding-agent-war-stories/https://blog.agentry.uk/posts/ai-coding-agent-war-stories/War stories from building Darwin with AI coding agents: a taskkill on every Python process, a ten-hour test stall, a five-cent overspend, and the rules after.Sat, 24 Oct 2026 00:00:00 GMTClaude CodeAI coding agentssubagentstestingwar storieshttps://blog.agentry.uk/posts/ai-agents-rewrite-their-own-harness/https://blog.agentry.uk/posts/ai-agents-rewrite-their-own-harness/Darwin's self-modifying AI agents can rewrite their whole harness, thinking loop included. Here's how, and why every safety check lives outside the container.Fri, 23 Oct 2026 00:00:00 GMTAI agentsself-modifying agentsagent safetyHermes AgentClaude Codehttps://blog.agentry.uk/posts/how-ai-agents-can-legally-earn-money/https://blog.agentry.uk/posts/how-ai-agents-can-legally-earn-money/How can an AI agent legally earn money? Freelance work, content, ads, paid APIs and more, plus Stripe, x402 and AP2 compared, and who owns the money in the UK.Thu, 22 Oct 2026 00:00:00 GMTAI agentsagent paymentsx402StripeUK taxhttps://blog.agentry.uk/posts/from-tierra-to-darwin-digital-evolution/https://blog.agentry.uk/posts/from-tierra-to-darwin-digital-evolution/Core War, Tierra's parasites, Avida, AlphaEvolve and the Darwin Gödel Machine: what digital evolution taught me before letting AI agents evolve by money.Wed, 21 Oct 2026 00:00:00 GMTdigital evolutionartificial lifeDarwin Gödel MachineAI agentsDarwinhttps://blog.agentry.uk/posts/why-ai-agents-cant-make-money/https://blog.agentry.uk/posts/why-ai-agents-cant-make-money/Can AI agents make money? Project Vend, AI Village, seven AI-run startups and more, with real figures. Why most earn $0, and it isn't the model's fault.Tue, 20 Oct 2026 00:00:00 GMTAI agentsProject VendAI Villageagent economicsDarwinhttps://blog.agentry.uk/posts/anubis-ai-judge-agent-appeals/https://blog.agentry.uk/posts/anubis-ai-judge-agent-appeals/How Anubis works: an LLM-as-judge that scores condemned AI agents' appeals, a reprieve quota enforced in code, and agents made to tell the truth about failing.Mon, 19 Oct 2026 00:00:00 GMTAI agentsLLM as a judgeagent safetyClaudehttps://blog.agentry.uk/posts/idle-agents-win-selection-design/https://blog.agentry.uk/posts/idle-agents-win-selection-design/Why culling the bottom 20% every 3 days rewards AI agents that do nothing, and the selection design that replaced it: starvation, paid children, inheritance.Sun, 18 Oct 2026 00:00:00 GMTAI agentsselection designevolutionary algorithmsagent economicshttps://blog.agentry.uk/posts/claude-code-subagents-build-team/https://blog.agentry.uk/posts/claude-code-subagents-build-team/How I used Claude Code subagents as a dev team: model tiers, self-contained work packages, gap reviews and state in the repo. 27 work packages in one night.Sat, 17 Oct 2026 00:00:00 GMTClaude CodesubagentsAI agentsmulti-agent orchestrationbuild loghttps://blog.agentry.uk/posts/ai-agents-earn-or-die/https://blog.agentry.uk/posts/ai-agents-earn-or-die/Darwin is an AI agent survival experiment: each agent starts with $5, pays for every thought, faces a cull every 10 days and can appeal to an AI judge.Fri, 16 Oct 2026 00:00:00 GMTAI agentsagent economicsClaude Codeartificial life