ExplainerAI in production
How AI got from rules to reasoning models
Follow the changes in learning methods, data and compute that led to today’s models, and see why newer does not always mean better for a task.
Articles
Guides, news and explanations about AI and reliable systems.
ExplainerPlatform engineering
Linux history: the kernel, the tools and the distributions
Unix, GNU, the Linux kernel and distributions explain the cooperating layers you operate today, including what a container actually shares.
ExplainerAI in production
Gemini 4 Argon: can it actually finish the job?
Google’s new reasoning model promises to stay with difficult work. Here is what that means for coding, research and investigations, and how to check the result.
ExplainerAI in production
Microsoft UserRepGen makes inputs for a simulated user
UserRepGen turns a conversation into task and behavior descriptions. Follow the intermediate artifact through a synthetic support-assistant evaluation.
ExplainerAI in production
PulsarLM keeps context small by dropping old token pages
PulsarLM removes low-attention pages rather than summarizing them. A synthetic archive-retention case explains what a local agent’s recall trial must inspect.
ExplainerAI in production
KoboldCpp 1.123 makes agent approvals depend on the tool and workdir
The global auto mode is only part of the permission decision. Follow per-tool overrides, directory scope and restored settings through a copied-runbook exercise.
ExplainerAI in production
Anthropic Scion puts source control behind a cache hierarchy
Scion reduces repeated repository work for online agents. Its cache hierarchy explains why warm checkout speed and fleet recovery need different tests.
ExplainerAI in production
AMD Helion separates agent execution from model inference
Helion places CPU agent loops, token routing, GPU inference and supporting services in separate capacity zones that still share one task outcome.
ExplainerAI in production
Muse Wonder: what useful agent memory needs to remember
A corrected maintenance window shows how Muse Wonder’s memory and persistence claims can be tested through recall, interrupted work and a verified result.
ExplainerAI in production
Scheduled AI work needs an execution contract
Follow a reliability report from its scheduled start through calculation, saving and notification, including recovery when a write times out.
ExplainerAI in production
Bit-exact decoding turns inference speedups into testable changes
Compare reference and speculative paths for token-level equivalence, then run a separate capacity trial so speed cannot conceal a serving-contract change.
ExplainerAI in production
Coding agent benchmarks: what actually won?
A coding agent's tools and testing policy help determine its benchmark result. Understand what was compared before choosing a setup for your team.