{"id":22517,"date":"2026-11-22T02:07:30","date_gmt":"2026-11-22T02:07:30","guid":{"rendered":"https:\/\/imsfund.com\/?p=22517"},"modified":"2026-11-22T02:07:30","modified_gmt":"2026-11-22T02:07:30","slug":"this-is-the-reason-your-ai-actually-remembers-anything","status":"publish","type":"post","link":"https:\/\/imsfund.com\/index.php\/2026\/11\/22\/this-is-the-reason-your-ai-actually-remembers-anything\/","title":{"rendered":"This Is the Reason Your AI Actually Remembers Anything"},"content":{"rendered":"<p> <br \/>\n<\/p>\n<p>\n\t\tOpinions expressed by Entrepreneur contributors are their own.\t<\/p>\n<div>\n<div class=\"tw:border-b tw:border-slate-200 tw:pb-4\">\n<h2 class=\"tw:mt-0 tw:mb-1 tw:text-2xl tw:font-heading\">Key Takeaways<\/h2>\n<ul class=\"tw:font-normal tw:font-serif tw:text-base tw:marker:text-slate-400\">\n<li>It\u2019s usually smarter to start simple. A basic cache is cheap and fast, and it teaches you what people actually want before you build anything more advanced.<\/li>\n<li>A simple cache has one weakness: It only recognizes a question if it\u2019s asked the exact same way. The moment someone rephrases things, which happens constantly, it can miss answers it technically already has.<\/li>\n<li>A smarter memory system costs a little more to run and can feel slightly slower, but that trade is worth it. It understands people however they phrase things, remembers for longer and can be reused across other projects.<\/li>\n<\/ul>\n<\/div>\n<p>During the World Cup this year, I built Rune Score, a sports analytics <a href=\"https:\/\/www.entrepreneur.com\/building-a-business\/i-help-businesses-get-recommended-by-chatbots-heres-how-it-really-works\" target=\"_blank\" rel=\"noreferrer noopener\">chatbot<\/a>. From the outside, it looks like a simple chatbot. Under the hood, it\u2019s a distributed system that has to handle the freshness of sports data, user rephrasing, extraction failures, LLM call costs, streaming, memory, fallbacks and security. What I learned building it applies to almost any AI product that has to stay fast, cheap and accurate at the same time.<\/p>\n<p>Two quick definitions before we go further, since neither tool is exactly a household name.<a href=\"https:\/\/redis.io\/tutorials\/what-is-redis\/\" target=\"_blank\" rel=\"noreferrer noopener\"> Redis<\/a> is \u201can open-source, in-memory data structure store used as a database, cache, message broker and streaming engine,\u201d most commonly used in AI products as a fast cache that stores one exact answer under one exact label.<a href=\"https:\/\/docs.aws.amazon.com\/prescriptive-guidance\/latest\/retrieval-augmented-generation-options\/what-is-rag.html\" target=\"_blank\" rel=\"noreferrer noopener\"> RAG<\/a>, short for retrieval-augmented generation, is defined by AWS as \u201ca technique used to augment a large language model (LLM) with external data, such as a company\u2019s internal documents.\u201d Instead of matching an exact label, it searches for a broader knowledge base using meaning rather than exact wording. Redis is precise and cheap but rigid. RAG is more flexible and durable, but costs more to run.<\/p>\n<h2 class=\"wp-block-heading\" id=\"h-why-i-didn-t-start-with-rag\">Why I didn\u2019t start with RAG<\/h2>\n<p>My first move wasn\u2019t a vector database, it was a Redis cache built around a simple labeling rule: Identical requests get an identical key, something like the football matchup Argentina vs. Spain. The savings were immediate. As I documented in my<a href=\"https:\/\/www.linkedin.com\/pulse\/from-redis-rag-13-production-lessons-building-ai-agents-vijey-tn6xe\/\" target=\"_blank\" rel=\"noreferrer noopener\"> LinkedIn article on Rune Score<\/a>, \u201cwithout the cache, a fresh extraction followed by an analysis can consume up to 23,000 input tokens: scraped content, context, history. With a cache that hits the same key, that number drops to about 5,000 tokens,\u201d a 4.5-fold reduction in the single most expensive part of the pipeline.<\/p>\n<p><a href=\"https:\/\/www.entrepreneur.com\/business-news\/tech\/as-ai-gets-cheaper-the-real-business-opportunity-is-shifting-heres-where-the-value-is-moving\" target=\"_blank\" rel=\"noreferrer noopener\">AI cost<\/a> usually comes from what you repeat unnecessarily. If 10 people ask your product the same question within a reasonable time window, rerunning the full, expensive pipeline 10 times is money burned for no reason.<\/p>\n<h2 class=\"wp-block-heading\" id=\"h-where-redis-breaks-down\">Where Redis breaks down<\/h2>\n<p>People don\u2019t talk in fixed labels. They might write \u201cFC Barcelona vs Real Madrid,\u201d or \u201cBar\u00e7a \u2013 Madrid,\u201d and a strict cache treats each variant as a completely different request. That means duplicate work, a missed answer that was already sitting there or an unnecessary rerun of the whole expensive pipeline just because the wording didn\u2019t match.<\/p>\n<p>This is where RAG earns its place, through a technique called <a href=\"https:\/\/www.entrepreneur.com\/business-news\/tech\/ai-is-exposing-major-gaps-in-traditional-data-backup-strategies-heres-how-to-protect-your-business\" target=\"_blank\" rel=\"noreferrer noopener\">embeddings<\/a>. According to<a href=\"https:\/\/docs.cloud.google.com\/bigquery\/docs\/vector-search-intro\" target=\"_blank\" rel=\"noreferrer noopener\"> Google Cloud\u2019s documentation<\/a>, \u201cthe closer objects are in the embedding space, the more semantically similar they are, and greater distances indicate greater semantic dissimilarity.\u201d In plain terms, a knowledge base built this way can retrieve a past answer even when the question is worded completely differently, because it searches by meaning rather than exact wording.<\/p>\n<h2 class=\"wp-block-heading\" id=\"h-the-part-that-actually-breaks-quality\">The part that actually breaks quality<\/h2>\n<p>Here\u2019s a lesson worth internalizing regardless of what you\u2019re building: The biggest quality bottleneck in a RAG system usually isn\u2019t the retrieval technology itself; it\u2019s what feeds it. This is the classic computing idea that<a href=\"https:\/\/www.techtarget.com\/searchsoftwarequality\/definition\/garbage-in-garbage-out\" target=\"_blank\" rel=\"noreferrer noopener\"> TechTarget defines<\/a> as garbage in, garbage out, or GIGO, \u201cthe idea that in any system, the quality of output is determined by the quality of the input.\u201d In my pipeline, every time I cut corners on how much content got extracted from a source, the resulting analysis got visibly worse downstream.<\/p>\n<p>The same logic applies to how the system ranks relevance. Trusting a single relevance score is fragile.<a href=\"https:\/\/redis.io\/blog\/reciprocal-rank-fusion\/\" target=\"_blank\" rel=\"noreferrer noopener\"> Redis\u2019s own engineering blog<\/a> puts it plainly: \u201cthe consistent takeaway across these studies is that combining retrievers tends to surface more relevant results in the top positions than either method on its own.\u201d Combining exact keyword matches, meaning-based search and how recent the information is, then merging those into one ranking, beats relying on any single signal alone.<\/p>\n<h2 class=\"wp-block-heading\" id=\"h-rag-costs-more-and-that-s-fine\">RAG costs more, and that\u2019s fine<\/h2>\n<p>In my experiment with Rune Score, RAG hit brings a request down from 20,000 to 8,000 tokens \u2014 a 60% reduction. Compared to the 80% reduction from an exact Redis hit, that\u2019s less economical and slightly slower, but that gap isn\u2019t a strike against <a href=\"https:\/\/www.entrepreneur.com\/science-technology\/4-ai-adoption-mistakes-this-engineer-sees-every-company-make\/504110\" target=\"_blank\" rel=\"noreferrer noopener\">RAG<\/a>. What it buys instead is tolerance for how people actually phrase things, memory that outlives a cache\u2019s expiration window and an architecture that\u2019s reusable well beyond the one product you built it for.<\/p>\n<h2 class=\"wp-block-heading\" id=\"h-the-bottom-line\">The bottom line<\/h2>\n<p>Redis and RAG aren\u2019t competing choices; they\u2019re sequential ones. The best refactor isn\u2019t the one that\u2019s theoretically right; it\u2019s the one that lets you roll back in seconds if the real metrics don\u2019t hold up. Build the cheap, fast cache first, let it teach you what your users actually ask for and only then invest in a full retrieval system for the cases a cache structurally can\u2019t handle.<\/p>\n<\/p><\/div>\n<div>\n<div class=\"tw:border-b tw:border-slate-200 tw:pb-4\">\n<h2 class=\"tw:mt-0 tw:mb-1 tw:text-2xl tw:font-heading\">Key Takeaways<\/h2>\n<ul class=\"tw:font-normal tw:font-serif tw:text-base tw:marker:text-slate-400\">\n<li>It\u2019s usually smarter to start simple. A basic cache is cheap and fast, and it teaches you what people actually want before you build anything more advanced.<\/li>\n<li>A simple cache has one weakness: It only recognizes a question if it\u2019s asked the exact same way. The moment someone rephrases things, which happens constantly, it can miss answers it technically already has.<\/li>\n<li>A smarter memory system costs a little more to run and can feel slightly slower, but that trade is worth it. It understands people however they phrase things, remembers for longer and can be reused across other projects.<\/li>\n<\/ul>\n<\/div>\n<p>During the World Cup this year, I built Rune Score, a sports analytics <a href=\"https:\/\/www.entrepreneur.com\/building-a-business\/i-help-businesses-get-recommended-by-chatbots-heres-how-it-really-works\" target=\"_blank\" rel=\"noreferrer noopener\">chatbot<\/a>. From the outside, it looks like a simple chatbot. Under the hood, it\u2019s a distributed system that has to handle the freshness of sports data, user rephrasing, extraction failures, LLM call costs, streaming, memory, fallbacks and security. What I learned building it applies to almost any AI product that has to stay fast, cheap and accurate at the same time.<\/p>\n<p>Two quick definitions before we go further, since neither tool is exactly a household name.<a href=\"https:\/\/redis.io\/tutorials\/what-is-redis\/\" target=\"_blank\" rel=\"noreferrer noopener\"> Redis<\/a> is \u201can open-source, in-memory data structure store used as a database, cache, message broker and streaming engine,\u201d most commonly used in AI products as a fast cache that stores one exact answer under one exact label.<a href=\"https:\/\/docs.aws.amazon.com\/prescriptive-guidance\/latest\/retrieval-augmented-generation-options\/what-is-rag.html\" target=\"_blank\" rel=\"noreferrer noopener\"> RAG<\/a>, short for retrieval-augmented generation, is defined by AWS as \u201ca technique used to augment a large language model (LLM) with external data, such as a company\u2019s internal documents.\u201d Instead of matching an exact label, it searches for a broader knowledge base using meaning rather than exact wording. Redis is precise and cheap but rigid. RAG is more flexible and durable, but costs more to run.<\/p>\n<h2 class=\"wp-block-heading\" id=\"h-why-i-didn-t-start-with-rag\">Why I didn\u2019t start with RAG<\/h2>\n<p>My first move wasn\u2019t a vector database, it was a Redis cache built around a simple labeling rule: Identical requests get an identical key, something like the football matchup Argentina vs. Spain. The savings were immediate. As I documented in my<a href=\"https:\/\/www.linkedin.com\/pulse\/from-redis-rag-13-production-lessons-building-ai-agents-vijey-tn6xe\/\" target=\"_blank\" rel=\"noreferrer noopener\"> LinkedIn article on Rune Score<\/a>, \u201cwithout the cache, a fresh extraction followed by an analysis can consume up to 23,000 input tokens: scraped content, context, history. With a cache that hits the same key, that number drops to about 5,000 tokens,\u201d a 4.5-fold reduction in the single most expensive part of the pipeline.<\/p>\n<\/p><\/div>\n<p><br \/>\n<br \/><a href=\"https:\/\/www.entrepreneur.com\/business-news\/this-is-the-reason-your-ai-tools-actually-remember-anything\" target=\"_blank\" rel=\"noopener\">Source link <\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Opinions expressed by Entrepreneur contributors are their own. Key Takeaways It\u2019s usually smarter to start simple. A basic cache is cheap and fast, and it teaches you what people actually want before you build anything more advanced. A simple cache has one weakness: It only recognizes a question if it\u2019s asked the exact same way. [&hellip;]<\/p>\n","protected":false},"author":5,"featured_media":22518,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"fifu_image_url":"https:\/\/www.entrepreneur.com\/wp-content\/uploads\/sites\/2\/2026\/11\/ai-remembers-1126-g-2264634343.jpg","fifu_image_alt":"","footnotes":""},"categories":[9],"tags":[],"class_list":["post-22517","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blog"],"_links":{"self":[{"href":"https:\/\/imsfund.com\/index.php\/wp-json\/wp\/v2\/posts\/22517","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/imsfund.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/imsfund.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/imsfund.com\/index.php\/wp-json\/wp\/v2\/users\/5"}],"replies":[{"embeddable":true,"href":"https:\/\/imsfund.com\/index.php\/wp-json\/wp\/v2\/comments?post=22517"}],"version-history":[{"count":1,"href":"https:\/\/imsfund.com\/index.php\/wp-json\/wp\/v2\/posts\/22517\/revisions"}],"predecessor-version":[{"id":22519,"href":"https:\/\/imsfund.com\/index.php\/wp-json\/wp\/v2\/posts\/22517\/revisions\/22519"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/imsfund.com\/index.php\/wp-json\/wp\/v2\/media\/22518"}],"wp:attachment":[{"href":"https:\/\/imsfund.com\/index.php\/wp-json\/wp\/v2\/media?parent=22517"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/imsfund.com\/index.php\/wp-json\/wp\/v2\/categories?post=22517"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/imsfund.com\/index.php\/wp-json\/wp\/v2\/tags?post=22517"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}