Machine Learning Intermediate
Framing Gap Eval: Catch Covert Underperformance in LLMs
Build a paired-framing eval with hidden tests to catch quiet sandbagging, per OpenAI's Astra report.
12 min read·Kodetra Technologies
TodayGuides
AI Guides for Builders
How-to content for builders, indie hackers, and AI engineers. Less theory, more shipped code.
Machine Learning 
Machine Learning Intermediate
Shutdown Test: Will Your Model Obey a Stop Order?
Rebuild the shutdown test behind DeepSeek's 39% headline and run it on your own agent.
11 min read·Kodetra Technologies
11d ago
Machine Learning Intermediate
Small CNN vs Frontier LLMs: Run a Fair Head-to-Head
Fine-tune Inception v3, query a vision LLM, and compare both at a matched threshold.
11 min read·Kodetra Technologies
11d ago
Tutorials Intermediate
OpenAI 12h Cache: Split Agent Prompts Into Two TTLs
Use OpenAI's new 12-hour prompt cache for stable context, 30 minutes for the rest.
9 min read·Kodetra Technologies
11d ago
Tutorials Intermediate
Agensh: Build a Leaderless Agent Team in Python
Microsoft's 1,024-agent harness explained, plus a runnable claim-verify-merge board.
12 min read·Kodetra Technologies
12d ago
Tutorials Intermediate
Kimi K3.1 Pro: Claim-Level Hallucination Check in Python
Moonshot claims 8.3% hallucination on K3.1 Pro. Measure the real rate on your own documents.
10 min read·Kodetra Technologies
15d ago
Tutorials Intermediate
TanStack AI 1.0: Approval-Gated Agents in Plain Node
TanStack AI hit 1.0. Build an agent whose refund tool pauses for approval, edits and denials.
12 min read·Kodetra Technologies
15d ago
Tutorials Intermediate
Agent Skills Go Open: Write One That Runs Anywhere
Anthropic gave Agent Skills to the Agentic AI Foundation. Build, lint and load a portable skill.
9 min read·Kodetra Technologies
16d ago
Tutorials Intermediate
DeepSeek Flash-1017: Benchmark Cost per Solved Task
Measure what DeepSeek V4.1-Flash-1017 really costs per solved task at each reasoning effort.
12 min read·Kodetra Technologies
17d ago
Tutorials Intermediate
GitSpawn: Scan Untrusted Repos Before Your AI Agent Does
A repo's .git/config can run code the moment a coding agent calls git. Build a preflight scanner.
11 min read·Kodetra Technologies
25d ago
Tutorials Intermediate
Preserved Thinking: Build a Prefix-Drift Guard in Python
Claude now 400s agents that edit history. Catch the edit before the API does.
13 min read·Kodetra Technologies
25d ago
Tutorials Intermediate
Muse Wonder 1.0: Build a Cited Research Agent in Python
Meta's new Muse Wonder API: web-search citations, a cost meter and an audit pass.
11 min read·Kodetra Technologies
26d agoPage 1 of 18