BLOG
Writing about what actually happens when you try to give AI agents long-term memory. The wins, the unsolved problems, and the parts nobody warns you about.
YOUR AI MEMORY CAN'T TELL A RIVER BANK FROM A SAVINGS ACCOUNT
Regex-based YMYL classification flags 'bank of the river' as financial and misses 'my chest hurts' entirely. How we fixed both problems with a two-stage pipeline at zero additional API cost.
WHY YOUR AI CODING AGENT KEEPS FORGETTING YOUR PROJECT CONVENTIONS
CLAUDE.md, .cursorrules, and Aider conventions files don't scale past 200 lines. Here's how to give Claude Code, Cursor, and Aider a real memory layer with widemem, plus three cross-session saves from this week.
THE REAL COST OF AI MEMORY: EVERY PROVIDER, COMPARED
How much does AI memory actually cost? Mem0, Zep, Letta, Cognee, LangMem, widemem. Platform fees, LLM calls, embeddings, infrastructure. Real numbers from $0 to $475/mo.
WHY CONTEXT WINDOWS AREN'T MEMORY (AND WHY IT MATTERS)
Context windows keep getting bigger. 128K, 1M, 10M tokens. But bigger context is not memory. Three research papers, real cost math, and benchmark data explain why.
THE COMPLETE GUIDE TO LLM MEMORY
A deep dive into how memory works for LLMs. The core concepts, how OpenAI, Claude, Grok, DeepSeek, and Perplexity handle memory, what memory layers are, and where this is all heading.
YOUR AI SHOULD KNOW WHEN IT DOESN'T KNOW
Why memory systems that always answer are worse than ones that sometimes say 'I'm not sure.' Confidence scoring, uncertainty modes, and the case for honest AI memory.
I BUILT A MEMORY LAYER THAT FORGETS ONLY WHAT DOESN'T MATTER
Why forgetting is harder than remembering, what cognitive science says about selective memory, and how batch conflict resolution, importance decay, and YMYL safety work under the hood.
YOUR AI FORGOT SOMEONE'S MEDICATION. NOW WHAT?
YMYL safety for AI memory: why some facts should never decay, how two-tier classification works, and the edge cases that still need solving.
THE CONTRADICTION PROBLEM IN AI MEMORY
What happens when AI agents accumulate conflicting facts over months. Why vector similarity alone cannot detect contradictions, and what batch conflict resolution looks like in practice.
EVERY LLM MEMORY PROJECT, RATED
A comprehensive directory of open-source projects, commercial platforms, and LLM provider solutions for AI memory. What each solves, how they compare, and how to choose.