SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingHacker News ·

Researchers Detail Zero-Click Spyware Spread Through Targeted Ads in AdFection Campaign

Citizen Lab has published research describing a method in which targeted advertising was used to deliver spyware without any user interaction. The report, referenced via a Hacker News discussion, examines how ad-tech infrastructure was exploited as a pipeline for distributing the malware. The analysis focuses on the mechanics of the delivery chain rather than the specific victims involved. The discussion thread around the report has so far drawn limited engagement, with 16 points and one comment. Further details are available in the original Citizen Lab publication.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

AdversaryGate patches six releases in four days after fixing fail-open merge-gate bugs

The author of AdversaryGate, a merge gate for coding-agent patches, shipped versions 3.0.0 through 3.0.6 after repeatedly finding ways the gate could wrongly approve changes. The bugs shared a pattern: an input the gate did not measure was silently skipped, letting the rest of the patch decide the verdict. Fixes included detecting tests that read code instead of running it, catching comment-only and no-op edits, refusing binary-marked diffs, and mutating every changed line and branch. Some changes, like handling the newline-at-EOF marker and raising timeouts and memory limits, addressed honest patches wrongly marked INCONCLUSIVE instead. Two releases, 3.0.2 and 3.0.3, were published with an API token, so versions before 3.0.4 were removed and releases now require signed tags and PyPI attestations.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Builds Automated Real Estate Data Pipeline Across 14 Houston Counties

A technical write-up describes how a developer integrated CrawlOS with Supabase to automate property data collection across 14 Houston-area counties. The system uses pg_cron for nightly job scheduling and scrapes Harris County Appraisal District records to produce roughly 125 qualified leads per run. The article covers the database schema design, crawler deployment patterns, and lead scoring logic. It also discusses challenges encountered when moving the prototype to production on self-hosted infrastructure.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Bienestar digital con IA: predecir el exceso de uso antes de que ocurra

The article argues that common screen-time apps like Screen Time, Digital Wellbeing, Freedom and Opal rely on static limits, reactive alerts and easily disabled blocks. It proposes a system built on an XGBoost forecaster that predicts next-day app usage and estimates when a user will hit a limit, along with a binge-risk score explained via SHAP. A transparent controller would gradually reduce usage toward a goal without cutting more than 10% per week, and an LLM coach would intervene at critical moments. The author reports a usage forecast R² of about 0.77 and a binge-risk AUC of about 0.79, which he deliberately left below higher benchmarks rather than manipulating the data.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Competitor prices vary by surface and fees, so one price per product misleads

A product's price is not a single figure, since the same item can cost different amounts on a retailer site, a marketplace, a grocery delivery service, and a delivery app. Each surface is set by different parties, including the seller's own margins, marketplace seller rules, and platform markups or fees. Storing one price field per product either overwrites useful data or averages it into a figure no customer actually paid. The author recommends recording each observation with its surface, location, fee components, and shopper context. Fee stacks can shrink or enlarge the gap between items: in one example, a 21.7% item-price gap narrowed to 9.1% once delivery and service fees were added.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Two Claude-related guides expired on Etsy and dropped from the catalogue

A small digital-products studio found two of its Claude-focused guides had expired on Etsy and disappeared from its catalogue. An alert on October 13 flagged the voice-coding guide, listing #138, after it expired the previous morning, and a similar alert on October 11 flagged the Supabase guide, listing #136. The alerts gave no reason for the expirations and included no sales or price figures, and the reviews of both listings were already waiting in the panel. The studio separately received a critical backup failure alert for its remote copy, though it is linked to the renewals only by timing. The account stresses that an expired Etsy listing leaves the catalogue the same morning, so renewal alerts need to arrive before the expiry date.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Using Pydantic to Get Reliable JSON Output from an LLM in Python

A small shop wants customer emails turned into structured order data for its spreadsheet, but language model outputs can break in several ways. Common failures include stray text around the JSON, wrong types such as writing a quantity as a word, and invented details like order numbers absent from the email. The guide separates format errors, which parsers and schemas catch, from factual errors, which require checking the output against the source text. It recommends a layered approach: clear prompts, schema validation with Pydantic, retries that feed errors back to the model, source checks, and escalation to a human after repeated failures. It also notes that fixed-template inputs may be better served by forms or regular expressions than by a language model.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

DuckDB 2.0 Alpha Triggers: BEFORE Row-Level Support Missing, Row Triggers Cost 3x

A developer tested the DuckDB 2.0 alpha build's trigger feature against the stable 1.5 development build on the same machine. Most BEFORE FOR EACH ROW triggers fail with a 'not yet supported' error, and AFTER FOR EACH ROW triggers work for INSERT and DELETE but not UPDATE. Statement-level triggers functioned for both BEFORE and AFTER across all tested events. Benchmarks on 500,000 inserts showed a row-level AFTER INSERT trigger was about 3.0 times slower than no trigger, while a statement-level equivalent added only about 1.3 times overhead. The tester also found undocumented restrictions, including that statement and row triggers cannot be mixed on one table, and that the DROP TRIGGER syntax in the preview post fails on this build.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why RFID, GPS, and IoT Each Solve Different Asset Tracking Problems

Knowing that an asset exists is simple, but determining its location, movement, and condition is often much harder. RFID uses radio waves to identify tagged items through compatible readers and suits inventory checks, though it does not provide continuous location tracking on its own. GPS is better suited to vehicles and mobile equipment moving across roads or remote sites, but it can struggle indoors and raises power and connectivity considerations. IoT sensors add context such as temperature and humidity, which helps confirm whether conditions like cold-chain requirements are being met. The article argues that the main engineering challenge is integrating these technologies into a single system so users do not have to compare separate dashboards.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

HustleMate AI: Open-Source Local Invoice Assistant for Freelancers

A developer has built HustleMate AI, an open-source tool that helps freelancers and small business owners organize invoices. Users paste invoice text, and the app extracts details such as vendor name, invoice number, dates, and total, then suggests a next action. Records can be saved in the browser and exported as CSV. The backend runs a locally hosted open-weight model via Ollama, so invoice data is not sent to a third-party AI service. The project was created as part of a Hacktoberfest open-source AI challenge and is aimed at making affordable, privacy-conscious AI tools accessible.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why Role-Based Access Controls in Fintech Are Drawing Regulatory Scrutiny

Role-based access control determines who can view customer data, approve payments, and change platform settings at financial firms. Regulators increasingly expect permission structures that match real job responsibilities, prevent conflicting duties, and record sensitive actions. The article gives an example of an investigator granted temporary administrator rights who keeps them after the task ends, showing that the risk lies in what authenticated users can do rather than in login security. It notes that the EU's DORA regulation, applicable since January 2025, requires least-privilege access and segregation of duties, while proposed PSD3 and PSR reforms add broader payment-security context. Institutions must translate these control objectives into their own operating models rather than follow a fixed set of roles.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Three New Serverless Vector Stores on AWS, Compared Through a Starter Project

The author needed a small private document collection to be searchable inside a serverless application, and found that AWS's earlier OpenSearch Serverless required a minimum spend of roughly $200 per month even in dev-test mode. Since then AWS has introduced three services that scale to zero: OpenSearch Serverless NextGen, S3 Vectors, and vector indexing in DynamoDB. The article compares their differences, costs, and tradeoffs, and the author concludes that OpenSearch Serverless NextGen is hard to beat. A CDK TypeScript starter project loads the same document chunks from a nine-page extract of the UK Aeronautical Information Publication into all three stores to test retrieval. The comparison focuses on technical feasibility rather than a formal benchmark, and the approach is presented as applicable to manuals, policies, and other private documentation.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why Python's list.pop(0) slows down and when to use deque instead

Calling list.pop(0) in Python works well for small queues but becomes slow as they grow, because removing the first item forces every remaining item to shift one position. A deque, short for double-ended queue, allows adding or removing items at either end in constant time. In a test on Python 3.10 with 50,000 items, emptying a list with pop(0) took about 0.189 seconds, while using deque.popleft() took about 0.002 seconds. The article recommends deque for queues, breadth-first search and sliding windows, while lists remain suitable for stacks and index-based access in the middle. Exact timings will vary by machine, but the performance gap widens as queues get larger.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Blog post examines the rising cost of finding a job today

A blog post titled 'The cost of getting a job today' was shared on Hacker News, where it received 7 points and no comments at the time of listing. The article, published on the author's personal site weli.dev, addresses the expenses and effort involved in job hunting. The source material provided only includes the link and metadata, so the specific arguments and figures in the post are not described here. Readers interested in the details can consult the original article directly.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Open-source MCP server flags malformed Search Console API requests before Google returns silent empty results

Google's Search Console API often returns empty results without any error when a request is invalid, such as an unsupported country code, a non-RE2 regex, or a date range in the future. This makes it hard to distinguish a broken query from a genuinely quiet period, and AI assistants can mistakenly explain the empty output. The developer built open-gsc-mcp, an MIT-licensed local Model Context Protocol server that validates requests before calling the API and reports problems in plain language. The tool also offers change-explanation and indexing-audit features, and it is read-only by default with no telemetry.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Protecting a Customer-Support LLM Agent from PII Leaks With Logging and Masking

A tutorial describes how to secure an LLM-based customer-support assistant for an e-commerce service like Cdiscount, which can look up orders, issue refunds, and send emails. It warns that attackers could use prompt injection or tool abuse to make the agent exfiltrate customer personal data. The proposed defense is ReskPoints, an AI agent logger that records each action with its confidence, parameters, and results. The tool automatically masks sensitive fields such as tokens, passwords, and contact details before logs leave the application. It also supports sampling rules to limit log volume while keeping full visibility into tool calls, and can export traces to platforms like Datadog, Prometheus, or OpenTelemetry.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

206 Sign-in Starts, Zero Signups: Checking Whether AI Visitors Really Convert

A founder saw 206 sign-in starts and no completed signups in a week, which seemed to point to a broken auth flow. The real cause was bot-like traffic: about 190 of 205 Google sign-in clicks in four days came from Germany, many with impossible browser and OS combinations. After filtering that traffic out, weekly visitors fell from 358 to about 85 and sign-in starts fell from 206 to one. The author argues that AI referral traffic should be judged by downstream conversion, starting with removing junk data before computing any rates. The method also covers building an AI-assistant segment using both referrer and utm_source fields, since most ChatGPT visits in one founder's data arrived with no referrer.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

7 CSS Myths Beginners Should Stop Believing

Beginners learning CSS often receive advice that is incomplete or simply wrong, which can lead to confusion and frustration. A common misconception is that developers must memorize every property, when the more practical skill is understanding core concepts and looking up details as needed. Another myth holds that CSS only changes colours, while in reality it governs layout, spacing, typography, sizing, positioning, responsiveness, and animation. The article also argues that good CSS is not defined by using many visual effects, but by applying styles with a clear purpose that supports clarity and usability.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

mcp-cpp: A C++17 SDK for the Model Context Protocol, Validated Against the Official Conformance Suite

A developer team has released mcp-cpp, a C++17 software development kit for the Model Context Protocol, which targets the 2026-07-28 specification and has no async runtime dependency. The SDK covers server and client roles, stdio and Streamable HTTP transports, OAuth 2.1 client support, and a JSON Schema validator. The authors tested it against the official modelcontextprotocol conformance suite, reporting full passes on the 2026-07-28 server and client checks. Their 2025-11-25 results were lower, with the remaining failures explained as deliberate non-goals or a harness bug. Running the official harness also uncovered five real bugs in the SDK, including dropped notifications over HTTP, which were then fixed.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Meta Launches Muse Wonder Models With Memory and Persistence Features for Agentic Work

Meta has released Muse Wonder and Muse Wonder Full, a two-model family aimed at agentic tasks, on 12 October 2026, with a 1M-token context window. Its Active Memory feature retains user-relevant details across conversations and resurfaces them when they match a current task, raising questions about where that data is stored and how long it persists. A companion feature, Active Focus, is designed to keep multi-step tasks running through difficulties, retry alternative approaches, and seek confirmation before consequential actions. The model card released alongside it includes capability comparisons against frontier rivals and security evaluation results, with Meta recommending human-in-the-loop safeguards for irreversible actions. Muse Wonder is accessible now via the Meta Model API and Muse Code, while Muse Wonder Full is limited to Muse Code subscribers for the time being.

0 comments Read more at DEV Community
← NewerPage 422 of 3108Older →