A Belgian paediatric oncology challenge asked a team to make lengthy clinical trial protocols easier for doctors to use. Rather than put a chatbot in front of the PDFs, the team designed a structured viewer: clinicians still need to verify treatment decisions against the original protocol, and a chatbot would add cost while potentially requiring patient details. The viewer instead surfaces diagnostic checklists, risk groups, treatment pathways, start-of-treatment thresholds and monitoring schedules, with each item linked directly to its source page. Search handles clinical synonyms, abbreviations, typos and French and Dutch; an optional assistant provides brief, page-cited answers but is not the main interface.
The central technical challenge was extracting reliable information from inconsistent PDFs. A pipeline analyzes page layout and uses Gemini to label sections and treatment logic, while code copies clinical wording directly from the documents and checks quotations and numbers. Translated summaries undergo checks for changed values or meaning. Automated tests, visual comparison, hand-labelled examples and clinician questions drove iterative fixes; the demo scored 97% on the supplied questions. Processing costs roughly €5-€10 per protocol and is repeated when protocols change, while ordinary searches and viewer features require no model calls. Patient information stays out of the system: browsing does not send data to a model, and case progress is stored in the doctor’s browser. The team’s goal is a reusable pipeline whose checks improve across protocols.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
Richard Sutton argues that brute-force computation alone cannot produce intelligence and that AI needs clever new methods. He says the science of AI is far from finished. (twitter.com)
Anthropic is accused of charging Russian Claude users before identity checks, then cutting off access after verification and retaining biometric data. Critics say account bans can affect linked GitHub projects and Russian citizens living abroad, and accuse the company of failing to answer public questions. (twitter.com)
LLM coding agents boost workplace productivity but can prevent developers from learning through hands-on problem-solving and weaken their technical skills. Developers should preserve their expertise by building projects manually, studying fundamentals, and practicing without AI tools in their own time. (ducktape.blot.im)
A fabricated Newtype interview led The Diplomat to retract its article about a Korean startup’s AI war test in Ukraine. The retraction followed a diplomatic dispute. (thediplomat.com)
Microsoft Research’s AI Frontiers group brought together about 50 researchers to develop agent systems and foundational AI advances. After the team joined Microsoft AI in September, its research-first approach clashed with leadership’s emphasis on scaling. (hunch.net)
Microsoft researcher John Langford has left the company. Microsoft’s AI Frontiers group has been declared finished. (businessexaminer.ca)
Today's best Hacker News stories, summarized and screenshotted, one email a day.