标签：#search

Short answer: for an ask-your-docs feature in a multi-tenant SaaS help center, start with embeddings over document chunks, retain keyword search for exact identifiers, and add reranking only when retrieval evaluation shows that the first-stage ordering is weak. The architecture is simple: ingest tenant-scoped chunks, embed them, store the vectors in a managed index, retrieve a small candidate set, and give only those matches to the answer model. The important marketplace constraint is less glamorous: every retrieval and model call must carry a tenant identifier into metering, or the team will know the total bill while remaining unable to explain which storefront created it. Don't begin with a vendor. Begin with the miss you can tolerate. How should a SaaS help center combine semantic search, embeddings, and keyword search? Semantic retrieval handles the normal language mismatch between a customer's question and the documentation. A user may ask how to “change the shop owner,” while the source chunk says “transfer account administration.” Keyword matching sees different tokens; embeddings map both query and chunks into vectors and can retrieve text with related meaning. That is the decisive reason to use embeddings for support questions, not fashion and not an assumption that vectors make every search problem better. Keyword search still earns a narrow, valuable lane. Error codes, plan names, API fields, invoice identifiers, and product-specific phrases often need literal matching. PAYMENT_1042 is not a semantic concept that should be softened into something approximately related. For a beginner implementation, run vector retrieval as the default and merge an exact-match result when the query contains one of those identifiers; don't build a many-stage ranking system before the corpus supplies evidence that you need one. Chunk boundaries matter because retrieval returns chunks, not abstract documents. Split by meaningful document structure, retain the page title and s

As research and AI-assisted papers surge, volunteer reviewers struggle to keep up.

The Stable Liar The dashboard was green for eight quarters The most dangerous number on a dashboard is the one that has stayed green the longest, and the way it fails has a shape you have probably watched up close. For eight straight quarters the dashboard holds green. Revenue up and to the right. Retention flat and healthy. NPS in the fifties. Every board meeting opens on the same slide and closes on the same nod. The plan is working. Then, six months after the eighth green quarter, the business the dashboard was supposed to describe nearly falls over. Pull the post-mortem apart and the easy story is that the numbers lied. They did not. Every quarter the dashboard reports something true: customers are still paying, logins are still happening, the survey scores are still fine. All of it accurate. The failure is quieter and worse than a lie. The words behind the numbers change meaning while the numbers stand still. “Retention” still counts the same logins, but a login has stopped predicting a customer who will renew. The metric keeps its shape long after the thing it measured has walked out of the room. Anyone who has run a team has felt a smaller version of this. The number you trusted most became the number that surprised you most. You were not lied to. You were tracking something that used to mean one thing and quietly came to mean another, and the dashboard had no way to tell you the meaning had moved. This is the stable liar: a number that goes on looking right long after it stopped being right. It is a structural property of measurement under pressure, and it has a law underneath it. Why every optimised metric drifts A metric is a substitution: you replace the thing you care about with something you can count, and the gap between them is where the trouble lives. Start with the substitution. You cannot measure value, loyalty, insight, or health directly, so you pick a proxy you can count. Revenue stands in for value. NPS stands in for loyalty. Citations stand in

A few notes and takeaways from the 2026 State of CSS survey results, including a nice CSS-Tricks cameo! 2026 State of CSS, Devs Surveys originally handwritten and published with love on CSS-Tricks . You should really get the newsletter as well.

Shopify says AI isn’t cannibalizing search traffic the way it has for publishers. Instead, AI-driven traffic and orders to Shopify stores tripled year over year in Q2.

A portfolio can look good on the usual scorecard and still answer the wrong question. One line says return was high. Another says risk-adjusted performance was acceptable. A third says drawdown stayed inside a tolerable range. Then the market turns, the benchmark starts recovering, and the thing I actually care about is different: how efficiently did the portfolio catch up? That is where a new metric can fool its own author. If I build a recovery measure and it moves almost exactly like an existing ratio, I have created a longer name for the same signal. The right acceptance test is geometric: a useful metric should cast a different shadow. This is the rule I used while validating Hyperlogarithmic Benchmark Catch-Up Ratio (HBCR): orthogonality to existing measures is a first-class test, not a chart for the appendix. 1. A new metric has to earn its axis HBCR was built to measure benchmark-relative recovery dynamics. The research page states the motivation plainly: traditional benchmark-relative metrics often fail to capture the true dynamics of investment performance, especially during market recoveries [ A New Metric for Private Equity Risk Adjusted Returns , Calibration of Risk and Correlation in Private Equity ]. That framing matters because the obvious validation path is tempting and weak. You compare the new number with familiar performance measures, find a comforting relationship, and declare victory. But a high correlation with a well-known score can be a warning. If HBCR strongly tracked Sharpe Ratio, it would probably be an expensive synonym for risk-adjusted return. The acceptance test I wanted was sharper. HBCR should have some relationship with performance, because recovery has economic content. It should also avoid collapsing into the same direction as Sharpe Ratio, Beta, Volatility, Alpha, Total Return, or Max Drawdown. Written as a predicate, the test has two sides. Let $\mathcal{T}$ be the set of metrics already on the scorecard, $\rho_{n,m}$ the corr

Bottom line: for a beginner ask-your-docs feature in a SaaS help center, I would start with embeddings-based semantic retrieval over document chunks, retain keyword search as a fallback, and add reranking only after I can measure weak top results. It is the least complicated architecture that handles the natural-language questions support teams actually receive while still giving an operator clear levers for relevance, cost, and SLOs. I learned to treat retrieval as a production dependency after a token bill landed at $8,742 for a help-center experiment I had estimated at under $1,000. The expensive part wasn't one dramatic model call; it was sending whole articles, navigation chrome, and duplicate chunks to the answering model for every vaguely phrased question. That mistake changed my order of operations: retrieve a small, attributable set first, inspect it, then generate. A chat model is a poor index. Small index. Big difference. How should a SaaS help center use semantic search, embeddings, and keyword search? Semantic search turns both a question and each document chunk into vectors, then retrieves chunks that are close in that vector space. For an ask-your-docs semantic search feature, that means a customer asking "Why can't I invite another teammate?" can reach a passage titled "Adding users to a workspace" even when the words do not line up. Keyword search remains useful for exact error identifiers, product SKUs, and freshly published terms, but by itself it is a thin answer to the way people phrase support questions. My beginner architecture is deliberately boring: export approved help-center content, strip templates and repeated navigation, split the remaining text into stable chunks, attach the page URL and heading as metadata, create embeddings, and put the vectors in a managed vector database. At question time, retrieve a modest candidate set, optionally rerank it, and pass only the best cited chunks to the chat model. Node.js can own the ingestion job

I was on a call last month with a startup CTO who had just gotten their AWS bill. They had built a beautiful RAG application: semantic search, conversational AI, the works. Their vector index was humming along with about 50 million embeddings. Then they hit product-market fit. Within six weeks, they scaled to 500 million vectors. Their monthly infrastructure costs went from $2,000 to $20,000. The real kicker? When we looked at the access patterns, over 80% of those vectors were queried less than once a week. They were paying hot-storage prices for data that was, by any honest measure, cold. The standard advice here is "just use a cheaper vector database." The more interesting question is: why are you storing all your vectors at the same temperature in the first place? The Cost-Recall-Latency Triangle Vector search forces a three-way tradeoff. You can optimize for cost, recall, and latency, but you only get to pick two. Want high recall and low latency? That costs money (in-memory HNSW graphs with full-precision vectors eating RAM). Want high recall at low cost? Latency goes up. Want cheap and fast? Recall suffers. Most teams pick a single point on this triangle and apply it uniformly to every vector in their index. That decision made sense when vector databases offered a single storage tier. It makes the same amount of sense as storing your entire filesystem on NVMe SSDs because some files need fast access. The conventional wisdom says you pick your point on the triangle and live with it. But the conventional wisdom was written before vector storage got interesting. The better approach: tier your vectors the same way you already tier your storage. Different access patterns deserve different economics. The same embedding that costs $0.12/month in RAM might cost $0.004/month on disk and $0.0002/month in object storage. When you have 500 million of them, those decimals matter. The Hot Tier: In-Memory HNSW and Exact k-NN For vectors that get hit constantly (your user-fa

Reddit advances lawsuit accusing Perplexity AI of conspiring with web scraper.

How a branch-free loop and byte-space arithmetic let GitHub case-fold every byte of code search at >45 GiB/s on a single core. The post Don’t stop early: Case-folding source code at memory speed appeared first on The GitHub Blog .

Manticore Search 28.6.6 has been released. The headline additions are UUID document IDs for real-time tables and ordered, limited GROUP_CONCAT() for grouped queries. The release also includes 16 fixes for backups, replication, query processing, SQL compatibility, and secondary indexes. This post covers everything shipped from 28.4.5 through 28.6.6 . Upgrade notes There are no new mandatory data migrations in this release. UUID IDs are an opt-in table definition: existing numeric-ID tables keep working as they are. If you want UUID identifiers, create a real-time table with id uuid ; ALTER TABLE cannot convert an existing table between numeric and UUID IDs. Two fixes are particularly useful for production installations. Successful backups now always unfreeze real-time tables when they finish (previously in rare cases they didn't), rather than leaving writes blocked. And authenticated replication can again add an existing populated RT table with ALTER CLUSTER ... ADD . UUID document IDs for real-time tables Applications often already have UUID identifiers from the system of record. Until now, using them with Manticore Search meant maintaining a separate numeric ID mapping. Real-time tables can now use UUID document IDs directly: CREATE TABLE products_uuid ( id uuid , title text , price int ); Manticore accepts an explicit UUID string, or generates one when id is omitted from an insert or replace. UUID equality and IN filters work in queries, and UUID IDs can be used with REPLACE , UPDATE , and DELETE . This is currently a real-time-table capability, including columnar and replicated RT tables. Plain, percolate, and sharded tables continue to use their existing ID models. Ordered and limited GROUP_CONCAT() Grouped results often need a compact preview of the most relevant values in each group. GROUP_CONCAT() can now sort values and retain only the requested number of them in explicit SQL GROUP BY queries: SELECT category , GROUP_CONCAT ( title ORDER BY price DESC SEPARA

AI is an amplifier; strategic focus on the organizational system brings the greatest returns. DORA's 2025 research on AI in software development provides team profiles and success capabilities that can be used to put the research into practice. By Ben Linders

How two API settings improved GPT-5.6 performance on ARC-AGI-3, boosting scores and efficiency by retaining reasoning and enabling compaction.

The screwup shows how tricky it can be to stop web crawlers from making ostensibly private conversations with AI chatbots entirely too public.

Google’s AI Overviews now appear in 43% of searches, underscoring how quickly AI-generated answers are becoming the default way people discover information online.

Researchers say TikTok, X, and Meta aren't providing data they're legally required to.

If you’re tired of sponsored and AI-summarized search results, Kagi gives you a curated—and human-centric—way to find what you’re looking for.

The European Commission claims that Google boosted its own apps and products to the top of search rankings to the detriment of its competitors.

Most bibliography failures show up the night before arXiv or the journal deadline: placeholder DOIs, year pasted into volume= , inverted page ranges, invented case reporters. The fix is a paper repo that fails closed from day one . One command pip install https://github.com/SybilGambleyyu/citesure/releases/download/v0.5.68/citesure-0.5.68-py3-none-any.whl citesure init my-paper cd my-paper citesure gate . --preset ci citesure gate . --preset arxiv citesure init writes refs.bib , pre-commit hooks ( gate --preset ci + soft-lint), .github/workflows/citesure.yml , and a short CITESURE.md for coauthors. Empty bibliographies skip hard-ID floors until entries appear. What the gate checks Soft-lint — placeholder number/issue, inverted pages, year-like volume/month/edition, unsafe keys, all-caps titles, missing venues, duplicate DOIs/titles Health — hard-ID coverage floors Promote dry-run — DOIs still buried in url= Live verify — Crossref, doi.org, arXiv, PubMed, Europe PMC, DataCite, OpenAlex, CourtListener Domain packs Fifty-five live-clean packs (demography, sociology, political science, anthropology, ML, law, ecology, …): citesure packs --gate-all citesure packs --run anthropology-classics Evidence: 256/256 integrity · 209/209 claim pairs · 55 packs. Source: github.com/SybilGambleyyu/citesure · Demo: citesure.sybilgambleyyu.workers.dev

LLM-written bibliographies do not stop at arXiv preprints. Law review drafts invent reporter cites; multilingual papers mangle Chinese titles. citesure 0.2 extends the integrity gate into those failure modes. US case law via CourtListener References that look like court cases — @jurisdiction entries, Plaintiff v. Defendant titles, or reporter strings such as 347 U.S. 483 — are resolved against Free Law Project CourtListener. Ranking prefers an exact reporter cite over companion orders, so Brown lands on 347 U.S. 483 rather than a later procedural listing. @jurisdiction { brown1954 , title = {Brown v. Board of Education} , year = {1954} , howpublished = {347 U.S. 483} , } citesure check examples/packs/us-case-law.bib citesure warm-cache cases.bib Optional COURTLISTENER_TOKEN for higher rate limits. Law-review CI: templates/journal/law-review.yml . CJK-aware matching NFKC + fullwidth folding; character-level similarity for CJK-heavy titles; CJK bigrams in claim scoring so Chinese claims are not silently empty. Evidence Integrity bench 242/242 (US cases + Chinese titles + multi-domain set) Claims mini-bench 29/29 Eight domain packs including us-case-law Install pip install "git+https://github.com/SybilGambleyyu/citesure.git[pdf]" Source: github.com/SybilGambleyyu/citesure · Demo: workers.dev