Dev claims ChatGPT cited his site two days after AI-optimisation overhaul

·3 views

Software developer Suman Debnath reports that ChatGPT began accurately identifying him in response to queries about his name around 24 August 2026, roughly two days after he completed a week-long answer-engine optimisation of his personal site. The changes included generating an llms.txt file, adding self-contained answer blocks to each page, and implementing entity disambiguation and structured data. Debnath verified the result across three devices, including one where his name had never been typed, to rule out personalisation as a cause. He cautions that two days of evidence from a single engine is a sequence, not proof of causation, and notes that Claude, Gemini, and Grok still do not cite his site. He also found that keyword stuffing his name into metadata and headings had no effect, arguing that AI engines prioritise entity corroboration and structure over repetition.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

NicheDB 0.2.0 adds domain name market tracking across 1,114 new gTLDs

Open-source data platform NicheDB has released version 0.2.0, its most significant update since launch, introducing a new domain-layer collection built on ntlddata.com's free, CC-BY-licensed JSON API. The release adds eight data feeds covering daily domain counts across new generic top-level domains, registry operator details for 1,114 TLDs, top gainers and losers, and accreditation changes. Developers noted several data quirks addressed in the build, including timestamp timezone gaps, misleading zero-movement entries on the first day of a series, and a page-size parameter that returned more rows than requested. The update also fixes a deduplication bug where a single publisher exposing one article through two feeds could trigger duplicate entries. Additional improvements include a corrected database key format that had previously caused query failures in Postgres.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Three Hidden Traps That Break JPG Uploads on Mobile and How to Fix Them

Mobile photo uploads commonly fail due to three technical issues: EXIF orientation metadata causing sideways images, oversized camera files ranging from 5 to 12MB, and iOS HEIC photos masquerading as JPGs. Phone cameras store rotation instructions in EXIF metadata rather than adjusting pixel data, which many browsers and image libraries fail to interpret correctly. Modern smartphones shoot at up to 48 megapixels, producing files far larger than most apps require, slowing uploads on weak connections. iOS devices save images in HEIC format by default, meaning a file named photo.jpg may not actually contain JPG data. Drawing the image onto an HTML canvas before upload corrects orientation and reduces file size, while server-side normalisation handles uploads from third-party apps and older client versions.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer Shares Tip to Fuzzy Search Terminal History Efficiently

Developer Akshay Khot published a short guide on fuzzy searching terminal history on February 3, 2022, via DEV Community. The post, tagged under terminal tools, was a quick one-minute read aimed at improving command-line productivity. Fuzzy search allows users to find past terminal commands without needing to remember exact syntax. The article received 15 reactions from the developer community, indicating practical interest in the tip.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

XXE Vulnerabilities Explained: How XML Parsers Can Leak Server Files and Internal Data

XML External Entity (XXE) vulnerabilities exploit a built-in XML feature that allows parsers to load content from external files or URLs, rather than injecting malicious code in the traditional sense. When an XML parser processes untrusted input without disabling external entity resolution, attackers can instruct it to read sensitive server files such as /etc/passwd and return their contents. Beyond file disclosure, XXE can be leveraged for server-side request forgery, theft of cloud provider credentials, and in rare cases, remote code execution. The vulnerability hinges on the parser behaving exactly as designed — processing the SYSTEM keyword to fetch external resources — making it a configuration and hardening issue rather than a code flaw. Security testers are advised to understand XML entity declarations and DTD structure to fully assess XXE exposure, including blind and out-of-band attack variations.

0 comments Read more at DEV Community