WebSearches Source-led research on search and discovery

Latest research

Newest 25 articles. Each preview is capped for scanning; open the article for the full source trail.

Search & DiscoveryOctober 1, 202613 min

Why AI Speaks Fluently but Doesn't Always Tell the Truth

A deep dive into the honesty gap the space between what artificial intelligence says with confidence and what it can actually verify.

The Gap Between What AI Says and What It Knows There is a moment in every conversation with a large language model when something strange happens. The response arrives instantly, shaped into clean paragraphs, threaded with confident assertions, and delivered with the steady assurance of a well-briefed expert. It reads like truth. It sounds like truth. And yet, somewhere beneath that polished surface, there may be no truth at all just language doing what language has always done: expressing ideas that feel precise...

Education & LearningOctober 1, 202612 min

AI's fluency masks a math problem and a dangerous honesty gap

A lessons-learned digest tracing how the distance between persuasive words and verified truth shows up in both AI systems and education policy and what the convergence reveals about building more trustworthy systems.

Artificial intelligence, despite its impressive ability to generate human-like text, fundamentally lacks mathematical reasoning skills, creating a significant and potentially dangerous honesty gap. This weakness means AI can present incorrect information with convincing fluency, making it difficult for users to discern truth from falsehood. The illusion of understanding, combined with a lack of genuine comprehension, poses a growing risk as AI becomes more integrated into daily life. This is the central tension...

Search & DiscoverySeptember 16, 202610 min

Hersh builds a Google for doctors' brains

How a Portland physician built one of the world's largest biomedical informatics programs and trained thousands to find answers in the language of medicine.

A Professor Who Speaks Fluent Search On a Wednesday morning in Portland, Oregon, a physician types a patient description into a clinical database. She is not browsing she is searching. She needs to know whether anyone in the hospital system matches a rare combination of symptoms, and she needs to know now. The query runs through a layer of structured and unstructured health data, drawing on techniques refined over decades by researchers who understood something fundamental: the body generates text, and that text...

Search & DiscoverySeptember 8, 202611 min

Search lab teams quietly reshape how Google ranks results

A tour through the teams, tools, and frameworks that quietly decide whether your search results are actually working and why that matters more than ever as AI reshapes how we find information.

The Question Nobody Wants to Ask You've built a search engine. The pipeline is running. Results come back in milliseconds. The feature works. But here is the question that haunts the teams behind modern search systems: are the results good ? This is where the real work begins. Not in building the search that, many teams can do. But in measuring whether the search serves the humans typing queries at 2 a.m. or during a lunch break or in the middle of a crisis. The auditors of relevance operate in the space between...

Search & DiscoverySeptember 8, 202610 min

Search engineers' early work fuels today's AI boom

Long before generative AI became a boardroom priority, a small community of information architects was building the controlled vocabularies and taxonomic structures that now form the invisible backbone of enterprise artificial intelligence.

There is a particular kind of document that exists in every large organization, though few people outside the department that created it ever think about its existence. It is usually a spreadsheet, sometimes a database, occasionally a hand-annotated chart pinned to a cubicle wall. It contains the vocabulary that the organization has agreed to use: which terms mean what, which synonyms collapse into a single concept, which broader categories contain narrower ones, and which words, despite sounding similar, carry...

Search & DiscoverySeptember 4, 20269 min

Louis Rosenfeld and Peter Morville Built the Map for Finding Things Online

A 2007 O'Reilly manual on information architecture quietly became one of the most enduring frameworks for how we design websites, build search engines, and help people find what they need.

The website you are reading this on did not organize itself. Somewhere, someone made decisions about where this article lives, how it connects to related pieces, how you found your way here, and what you see first. Those decisions belong to a discipline that Louis Rosenfeld and Peter Morville gave a name and a structure in their 2007 book Information Architecture for the World Wide Web , published by O'Reilly Media. That discipline information architecture is the invisible scaffolding of the modern web. It is the...

Search & DiscoveryAugust 31, 202611 min

Lisbon's linguists built the search algorithms powering AI today

From a 1932 philology center in Lisbon to today's knowledge-graph laboratories, a distinctive line of Portuguese research quietly shaped how machines understand meaning.

In a narrow corridor on the fourth floor of a building in the Alfama district, a researcher in 2025 was debugging an ontology-matching algorithm. The code integrated lexical signals, contextual features, and language-model-based embeddings into what its documentation described as a "unified, interpretable architecture." The repository was called Exact-OM. It lived on GitHub under the account liseda-lab. And its institutional home was a university that, in 2026, had spent nearly a century studying what language is...

Search & DiscoveryAugust 26, 202610 min

Susan T. Dumais and the Quiet Innovation of Latent Semantic Indexing

How a Microsoft Researcher's Mathematical Framework for Understanding Language Through Numbers Shaped the Way We Find Information Online

There is a problem at the heart of every search engine, one so fundamental that most users never notice it. It lives in the gap between what you type and what you mean. You search for 'cheap flights to Seattle' and the engine returns pages about airplane seats. You look for 'jaguar' and receive results about cars, animals, and a football player depending on the day, the algorithm, and what the crowd has been clicking. The machine is matching words, not meanings. And for decades, that seemed like an acceptable...

Search & DiscoveryAugust 20, 20269 min

British duo's probability engine still fuels web search today

Stephen Robertson and Karen Spärck-Jones never set out to shape the modern internet. Their 1970s theory of relevance did it anyway.

There is a formula sitting inside virtually every major search engine in the world. It has no flashy name, no Silicon Valley origin story, and its inventors never sought fame. But the equation that became BM25 a scoring function for ranking documents by relevance has quietly shaped how billions of people find information online for more than three decades. The story begins in London in the early 1970s, in the corridors of City University, where two researchers were asking a question that seems almost obvious in...

Search & DiscoveryJuly 31, 202614 min

Google rankings alone won't cut it in the age of AI search

A deep look at how answer engines actually pick sources reveals why traditional SEO wisdom is quietly leading marketers astray and what the data says you should do instead.

Most marketers are doing Answer Engine Optimization wrong. Not because they lack effort or intelligence, but because they've been handed a map drawn for the wrong territory. The conventional wisdom goes something like this: if your content ranks well on Google, AI engines will naturally cite you. After all, doesn't Google's algorithm already identify quality sources? And aren't these AI systems just synthesizing web content anyway? So the playbook becomes simple: keep doing SEO, maybe add some schema markup, and...

Search & DiscoveryJuly 25, 202615 min

Gerard Salton the forgotten father of Google search

The SMART system, born in a 1960s Cornell lab, quietly invented the math that still runs inside every search engine you use today.

Most people assume the logic of modern search engines began with the rise of the internet. In reality, the foundation wasH wasB was laid in 1968 by Cornell mathematician Gerard Salton, who solved the problem of how machines identify document topics long before the first web browser existed. The answer Salton developed would become the invisible skeleton of every search system built in the decades that followed. His method the vector space model treated documents not as text but as points in space. Queries became...

Search & DiscoveryJuly 22, 202612 min

Eugene Garfield's radical index reshaped scientific research

Before Google, before semantic search, a chemist named Eugene Garfield asked a deceptively simple question about how scientists cite each other's work and built an empire of meaning from the answer.

Eugene Garfield's creation of the Science Citation Index fundamentally changed how scientific research is conducted and evaluated. Before the 1960s, tracking the impact of scholarly work was a laborious and imprecise process; Garfield's index provided the first systematic way to measure a publication's influence through citation analysis. This innovation not only streamlined literature reviews but also established a new metric - the impact factor - that continues to shape academic careers and funding decisions...

Search & DiscoveryJuly 20, 202611 min

Hidden AI powers next-gen college research tools

How ontology-guided knowledge graphs are quietly reshaping how search engines retrieve, reason about, and return academic knowledge and what the latest research reveals about the approach that may finally outpace vector retrieval.

The Problem With Knowing Too Much and Understanding Too Little There is a quiet crisis unfolding in academic search. Modern large language models have demonstrated strong generative and reasoning abilities across a variety of domains, but their reliance on static training data limits their access to unseen and domain-specific knowledge. That limitation becomes especially acute when researchers, students, or practitioners search across specialized corpora where precision matters more than fluency, and where a...

Search & DiscoveryJuly 18, 202613 min

The Recall Architect How Dr. Susan Leavy Built a Framework for Measuring Search Engine Truthfulness

An Irish AI researcher who began detecting bias in political news has quietly shaped how governments and institutions think about what search engines owe their users.

We assume search engines simply return information; the reality, however, is they actively *shape* it. Dr. Susan Leavy spent years building systems to understand how information is prioritized - first on financial trading floors, then in academic research. This unlikely journey led her to a groundbreaking question: can we actually measure the “truthfulness” of a search engine's results, and build a framework to ensure more reliable information access? She was studying bias in political news coverage training...

Search & DiscoveryJuly 15, 202610 min

The Textbook That Defined How We Teach Machines to Find Information

How Ricardo Baeza-Yates and Berthier Ribeiro-Neto built a rigorous computer-science framework for information retrieval that still shapes search technology today.

There is a moment, frozen in the late pages of a book that helped teach an entire generation how to build search engines. The cover shows a simple line drawing a hand reaching toward a globe, fingers almost touching. It is 1999, and the World Wide Web is still young enough that most people access it through dial-up modems, still chaotic enough that nobody quite knows how to organize it. In Santiago, Chile, and Belo Horizonte, Brazil, two computer scientists have just finished a book that will become one of the...

Search & DiscoveryJuly 13, 202610 min

Edinburgh lab's Scrapy powers academic search worldwide

From a Glasgow computing lab in 2004, a team of researchers set out to build a search engine that could handle the world's largest document collections and ended up creating the platform that researchers worldwide still rely on today.

## The Lab Where Search Was Taken Seriously In a computing science department on the banks of the River Clyde, a group of researchers decided that searching large document collections shouldn't require proprietary software or massive institutional budgets. Around 2004, members of the Information Retrieval Research Group at the University of Glasgow began building what would become one of academic search's most enduring open-source tools and they named it with characteristic Scottish humor. Terrier. As in terabyte...

Search & DiscoveryJuly 10, 202610 min

How citation networks are rebuilding academic discovery

A new wave of academic search tools is turning the traditional keyword search inside out following the invisible threads between papers instead of just the words inside them.

Citation networks are replacing keyword searches to fundamentally improve how researchers discover foundational academic work. By mapping the relationships between papers rather than relying on isolated terms, these networks prevent critical research from being buried by search algorithms. This is not a failure of effort. It is a structural limitation of how most academic search works and a small group of researchers have decided to fix it. The Map Behind the Paper A citation network is, at its core, a visual map...

Search & DiscoveryJuly 10, 20268 min

How the Allen Institute Is Rebuilding Academic Search From the Researcher's Mind

An AI research institute founded by Paul Allen has quietly built two tools that rethink how scholars find, compare, and synthesize scientific literature without keyword guesswork or citation dead ends.

The Search That Never Fit the Researcher For decades, academic search has asked researchers to do something counterintuitive: reduce complex questions into a handful of keywords, then scroll through results that may or may not match what you actually meant. You craft a query, scan a list, rephrase, scan again, follow a citation, get lost in related work that turns out to be tangential, and start over. The process works, sort of, but it demands the researcher bend their thinking to match the tool more than the...

Search & DiscoveryJuly 1, 20268 min

Quora's truth arbiter Anant Shar's quality framework revealed

This article explores the intersection of answer quality engineering and truth-decision frameworks at Quora, drawing from available public materials and documented practitioner approaches.

The Question Behind the Question Somewhere in the vast machinery of how knowledge gets organized online, someone has to decide what counts as true. Not in the philosophical abstract philosophers have been doing that for millennia but in the practical, everyday sense of whether an answer to a question is good enough to show to someone who asked. This is the work that sits at the heart of answer engines, and it is harder than it looks. The term "answer quality engineer" sounds almost clinical, like something from a...

Search & DiscoveryJune 28, 202612 min

Inverted index the tech powering Google's search dominance

Tracing the quiet intellectual journey from library card catalogs to Google, through the data structure that still powers every search engine you use today.

Everyone assumes Google's search dominance comes from some impossibly complex algorithm, a secret sauce of artificial intelligence. But the foundation of its speed and power isn't cutting-edge AI at all it's a surprisingly straightforward data structure called the inverted index. This technique doesn't focus on analyzing documents, but rather on flipping the search problem entirely: instead of finding documents *by* their content, it finds them *through* their words. This insight that inversion is the key to speed...

Search & DiscoveryJune 26, 202610 min

The Man Who Taught Computers to Read Between the Lines

How Gerard Salton's Cornell research team built a mathematical framework for understanding text that still powers every search engine you use today.

We often assume the information age began with the internet, but a crisis of knowledge overwhelmed us decades earlier. Even as scientific publishing boomed after World War II, the tools meant to organize it - libraries and their card catalogs - were failing spectacularly. Researchers weren't lacking information, they were drowning in it, hampered by inconsistent indexing and the limitations of keyword searches. This pre-digital deluge sparked a surprisingly modern quest: to teach computers to *understand* what we...

Search & DiscoveryJune 25, 20268 min

The Relevance Engineer From Academic IR Labs to Modern Search Products

The discipline that emerged when search engines stopped matching strings and started understanding meaning now shapes whether your content gets retrieved or buried inside AI-generated answers.

There is a moment in every university information retrieval lab when a researcher realizes that the system they have spent months tuning is solving the wrong problem. The queries that real people type into search bars do not arrive clean and precise. They arrive messy, ambiguous, underspecified fragments of a thought, half a phrase, sometimes just a name spoken aloud and transcribed by accident. The researcher learns to build for this chaos, not against it. That same sensibility building for how people actually...

Search & DiscoveryJune 22, 202612 min

The Patent Decoder Bill Slawski's Twenty-Year Journey Mapping Google's Hidden Logic

For two decades, Bill Slawski turned dense Google patent filings into the intellectual foundation that serious SEO practitioners still work from today.

There is a particular kind of reader who, when handed a 50-page patent filing dense with technical language and legal boilerplate, sees not a wall but a window. Bill Slawski was that reader. For more than twenty years, he sat with documents that most people in the search industry would have scrolled past, and he pulled meaning from them not to win arguments or chase algorithm rumors, but to understand. To map, as he once put it, what Google was actually building beneath the surface of its public statements....

Technology & AIJune 20, 202611 min

The Long Road to Margaret Thatcher's Britain Paul Graham's 1983 Photo Book Returns

How a photographer's journey along Britain's oldest thoroughfare became an accidental portrait of a nation in transformation.

There is a photograph in Paul Graham's book A1: The Great North Road that tells more about Britain in the early 1980s than any policy paper or political speech. Two men in business attire stand against the creamy white stone of the Bank of England in the City of London. One, dressed in a blue suit, smiles broadly as he holds a piece of notepaper toward his younger companion, who wears navy-blue pinstripes and an equally broad grin. The bright-blue necktie of the older man is flopped over his raised arm, its hue...

Technology & AIJune 18, 202612 min

Les Grenadiers and the Quiet Revolution of Haitian Pride

For the first time since 1974, Haiti's national soccer team has qualified for the World Cup and in doing so, it has carried something far larger than a bid for sporting glory into the arena.

The Weight of a Single Goal On the night of November 18, 2025, Louicius Deedson sliced through Nicaragua's defense and buried the winning strike that had eluded Haiti for more than five decades. The goal, scored in the dying moments of the World Cup qualifiers, sent a nation and its far-flung diaspora into a collective release that had been building for generations. That date carried additional resonance: it fell exactly 222 years after Jean-Jacques Dessalines fought his famous battle against French forces on the...