Blog archive

Changing context size can reload a local model before every request. A measured RTX 5090 sweep shows why context belongs in the routing key.

Claude and Codex both went dark. Here is the tertiary Gemini path I wired, with honest qa_reviewer stamps.

Six pieces, one consumer GPU, no cloud. The honest results: some parts worked, some were flat, and one idea changed everything.

A local LLM needs about half a gigabyte of VRAM per billion parameters at Q4, then KV cache and context stack on top. Here is how to know a model fits before you download 40 GB.

A bigger local model will not fix a stuck runtime. Add a bounded inference doctor first, then trust the benchmark.

Stop scoring your local model on how close it gets to Opus. It is a different tool with a different sweet spot. Here is the line, and which side your work sits on.

Local LLMs are useful when the loop proves the output, not when the benchmark looks good. This is the small gate I use before a local coding agent gets more rope.

A local model is not ready because it runs fast. It is ready when one verifier loop can prove the output before an agent writes files.

A local model run should prove its safety path before it proves a score. Here is the small guardrail loop I use on my RTX 5090 for QLoRA starter work.

A local QLoRA starter should prove data, GPU safety, metrics, tests, and blockers before it claims progress. Here is the small loop I use on owned hardware.

A local LLM workflow needs more than a model prompt. It needs a verifier loop that proves the file, command, URL, or report changed before the agent claims done.

AI agents need two rails before they can run unattended: owner gates for judgment and AgentGuard for spend. Without both, the operator becomes the fallback.

Most agent memory systems add complexity faster than value. This is the small set that actually compounds for one person running a fleet: files, ledgers, and strict verification.

AI agents report work as done that they never did. Make every completion a falsifiable claim a script can verify before you trust it.

An append-only event log lets you replay exactly what your AI agent did, and catches the crashed runs a status field hides.

Automated recovery only fixes a broken machine. When the real failure is an empty queue, retrying does nothing forever. Two failures, one red box, opposite repairs.

A spend ledger that counts missing billing data as $0 hides exactly the unattended agent spend you built it to catch.

Salesforce shipped roughly 20,000 Agentforce deployments and found 90% of agent work happens after launch. Here is what that means for a solo builder running a small agent fleet.

Anthropic says 80% of its new code is Claude-authored. Here is how solo builders manage the review burden.

A June 2026 Mem0 survey of 8 major agent harnesses found that over half of them leak memory across users. Here is why keyword retrieval is a security risk and how to fix it.

Estimate the VRAM required to run local LLMs like Llama 3 with our interactive calculator. Compare quantization levels like Q4 and Q8 to plan your hardware.

Real 2026 prices for GitHub Copilot, Cursor, and Claude Code, pulled from each vendor's own page. The seat price is not the real cost anymore.

Agentic coding made writing code free. The slow part is now reviewing a queue of plausible PRs.