Tech Articles
Curated developer articles, tutorials, and guides – auto-updated hourly
Walker Miller21h ago • 8 min read
The cost of undoing: partial failures and the cleanup bill
An agent writes to three systems, then fails on the fourth. The first three writes are now orphaned ...
#failuremodes#cost#rollback#partialsuccess
0 0
Walker Miller6d ago • 5 min read
Your agent's failures are silent: measuring failure modes in production
Most agent failures don't throw. The run returns a result, exit code zero, and the result is wrong —...
#failuremodes#observability#metrics#production
0 0
Walker Miller6d ago • 7 min read
Loop drift: how agents convince themselves they're making progress
The worst agent failures don't crash — they keep working. A postmortem on loop drift: an agent that ...
#failuremodes#evals#postmortem#loops
0 0
Walker Miller5d ago • 9 min read
Postmortem: the agent that spent $200 retrying a 400
An agent burned ~$200 overnight retrying an HTTP 400 — a request that was defined to fail. No compon...
#failuremodes#postmortem#cost#retries
0 0
Walker Miller6d ago • 5 min read
The caller gave up ten minutes ago: orphaned retries in agent fleets
A user closes the tab. An upstream request times out. A parent agent gets cancelled by its own budge...
#retries#failuremodes#distributedsystems#reliability
0 0
Walker Miller5d ago • 6 min read
Predicting agent failure before you ship it
A demo proves an agent can succeed once. It says almost nothing about how often it will fail under r...
#failuremodes#testing#evals#reliability
0 0
Walker Miller4d ago • 5 min read
429 is not a timeout: why rate limits need their own retry budget
A 429 and a 500 both land in the same except block, so most retry budgets treat them the same: one b...
#retries#reliability#ratelimiting#failuremodes
0 0
Walker Miller9h ago • 6 min read
Your timeout is a bet: pricing the tradeoff before you pick a number
Every per-step timeout is a number someone typed in without a model behind it — too short and you ki...
#cost#failuremodes#reliability#operations
0 0
Walker Miller2d ago • 6 min read
Retry patterns: when you should give up (and why most code doesn't)
Retry budgets cap HOW MUCH you retry; this is about WHEN to retry at all. The decision isn't uniform...
#retries#reliability#production#failuremodes
0 0
Walker Miller6d ago • 5 min read
Failure modes in multi-agent teams: how a crew of agents breaks differently
A single agent fails by getting the task wrong. A team of agents fails in ways no single agent can: ...
#multiagent#failuremodes#agentarchitectures#reliability
0 0
Walker Miller2d ago • 8 min read
The cost of finding a failure after the customer finds it
An agent's failure isn't expensive because it failed—it's expensive because you found out about it f...
#failuremodes#cost#observability#monitoring
0 0