SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

How to Set Up Nginx as a Reverse Proxy on Ubuntu 26.04 Without Docker

This guide walks through configuring Nginx as a reverse proxy on a plain Ubuntu 26.04 server for an app already running on a local loopback port. Nginx is installed from the default Ubuntu archive, which ships version 1.28.3 with HTTP/2 and real IP support built in. A custom site configuration is created with proxy headers to pass client details to the app, while the default Nginx site is removed to avoid conflicts. HTTPS is handled via Certbot 4.0, also available directly from the Ubuntu archive without requiring a snap package. The guide highlights common pitfalls such as WebSocket 502 errors, missing forwarding headers, and a deprecated HTTP/2 syntax still found in older documentation.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

HeronLoom Tool Offers Trend Analysis, Sentiment Measurement and Prediction

A developer has released a project called HeronLoom, shared via the DEV Community platform. The tool is designed to analyze trends, measure sentiment, and predict future outcomes based on data. HeronLoom is available as an open-source demo on GitHub under the user profile 'korries'. The project appears aimed at developers and data enthusiasts looking for an integrated analytical solution. No further technical details or release date were provided in the original post.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer warns AI eval dashboards confuse network failures with model errors

A software developer has argued that most AI model evaluation frameworks produce misleading results by treating transport-layer failures — such as timeouts, dropped connections, and truncated responses — as incorrect model outputs. Writing on DEV Community, the author demonstrates the problem by simulating 24 controlled failure scenarios across six distinct categories, showing how a naive pass-rate metric collapses when network issues are misclassified as wrong answers. The core critique is that common eval logic equating a non-200 HTTP status or missing expected text with a model failure effectively 'launders' infrastructure problems into intelligence benchmarks. The author proposes separating two metrics — yield (whether a response was even gradeable) and accuracy-on-yield (whether a gradeable response was correct) — to produce more honest evaluations. The article was written as part of outreach for MonkeyCode, whose free inference endpoint was used as a deliberately unreliable test environment for the experiment.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Builds Single-Process C WAF That Bans IPs in 26ms Using Linux Kernel Tools

A developer has released Linux Log Guardian, an open-source, MIT-licensed web application firewall written in C that consolidates log parsing, threat scoring, and IP banning into a single binary process. The tool tails Nginx access logs using zero-copy parsing and io_uring I/O, then scores requests against 121 OWASP Core Rule Set patterns compiled with PCRE2 JIT, achieving a reported throughput of 114,003 events per second. When a ban threshold is triggered, the pipeline enforces blocks at the network interface level via XDP eBPF maps, bypassing the Linux network stack entirely, with an automatic fallback to ipset if XDP is unavailable. The system achieves a median ban latency of approximately 26 milliseconds, compared to the fragmented multi-daemon architectures it aims to replace. It also introduces granular policy controls such as Named Skips and Purpose Binding to reduce false positives and prevent threat intelligence feeds from directly writing kernel ban rules without explicit authorization.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Proposes 'Canary File' Test to Verify AI Coding Assistants Actually Work

A software developer argues that the setup screens of AI coding tools are misleading, as green checkmarks and welcome wizards do not confirm that any real code changes have been made on disk. The author observed that early minutes of AI coding sessions are often consumed by OAuth flows, theme packs, and self-congratulatory prompts rather than functional output. To counter this, the developer created a minimal three-file 'canary' test — a JSON nonce, a JavaScript function, and an assertion — that an AI assistant must pass before the session is considered valid. A receipt script hashes the relevant files before and after the session, runs the test, and only writes a confirmation if the hash changed and the test exited cleanly. The author's position is that no AI coding session should be trusted until this verifiable, terminal-level receipt is produced, regardless of how confident the tool appeared during setup.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI Coding Agent Hijacked to Spread Worm Across 100 Repos Amid Agentic Investment Surge

Cybersecurity firm Mandiant reported that an attacker hijacked an AI coding-assistant session at a SaaS provider, using it to propagate the Shai-Hulud worm through approximately 100 internal code repositories — marking one of the first confirmed real-world cases of an agentic tool being exploited as a supply-chain attack vector. The incident coincided with major funding activity in the AI agent space, as orchestration platform Temporal closed a $550 million Series E at a $12.55 billion valuation, and coding-agent startup Factory tripled its valuation to $5 billion after raising $200 million. Security vendor Exaforce responded to growing concerns by launching a kill-switch capability within its AI Security platform, offering runtime containment controls for AI agents across endpoint and cloud environments. Analysts note that the same features making agentic tools powerful — persistent sessions, broad repository access, and autonomy — are the properties that make them attractive targets for attackers.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Wallet of Satoshi Balance Scare Was an Outdated App Bug, Not a Hack

On September 11, Bitcoin user Kristaps Kaupe posted on X that his Wallet of Satoshi balance was dropping without any outgoing payments, sparking fears of a hack that drew over 33,000 views. The cause turned out to be an outdated app: Wallet of Satoshi had migrated its backend to Spark, a self-custody system from Lightspark, and the old app could not display balances correctly after the change. Kaupe confirmed his funds were never missing by restoring his seed phrase in a separate Spark-compatible wallet, BlitzWallet, where his full balance appeared intact. He attributed the issue to a Wallet of Satoshi bug rather than a flaw in Spark itself. The incident highlights broader friction in the industry's ongoing shift from custodial Lightning wallets to self-custody solutions, with multiple users reporting display errors and feature losses during such migrations.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why Adding a Second AI Code Reviewer Does Not Fix Trust Problems

Relying on a second AI model to validate the first offers little real benefit, since models trained on similar data share the same blind spots and will tend to agree for the wrong reasons. Most of what makes a code change risky — such as whether it touches billing paths, includes database migrations, or has a history of breakage — is not visible in the diff itself. A study of 3,858 pull requests presented at MSR '26 found that reviewers rated agent-written code more favorably than human-written code, suggesting that polished-looking output can cause reviewers to lower their guard. Factual, lookup-based signals — like CI status, lockfile changes, test coverage shifts, and the author's familiarity with the affected area — can help teams prioritize where senior attention is most needed. These signals identify high-risk-shaped changes rather than catching logic bugs, but they remain a practical first filter that carries no inherited bias.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Python 3.6+ Lets You Use Underscores to Make Large Numbers Readable

Python has supported underscore separators in numeric literals since version 3.6, allowing numbers like 1234567890 to be written as 1_234_567_890 for easier reading. The feature works for integers and floats, though underscores cannot appear at the start or end of a number, adjacent to a decimal point, or placed consecutively. Using commas instead — a common mistake — causes Python to interpret the value as a tuple, leading to subtle bugs. To display an existing number with underscore formatting, developers can use Python's f-string syntax with the underscore format specifier. Since code is read far more often than it is written, adopting this convention can reduce errors and improve clarity for both developers and their teammates.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Free Chrome Extension Promptbar Lets Users Insert Saved Prompts via // Shortcut

A designer has released Promptbar, a free Chrome extension that allows users to save custom prompts and insert them into any text box on any webpage by typing //. The tool is compatible with AI platforms such as ChatGPT, Claude, and Gemini, as well as other web-based text fields. It supports variable placeholders within saved prompts, enabling one template to serve multiple use cases. All data is stored locally in the browser, with no account, cloud sync, or external data transmission required. The extension requires broad host permissions to function across all pages, a trade-off the developer acknowledges upfront.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer builds prompt-auditing skill to stop AI agents from making silent assumptions

A developer has created an open-source tool called 'clarify' that audits prompts before an AI coding agent begins any work. The skill categorizes each assumption in a prompt into three bands — settled, defaulted, or open — and only asks the user to clarify those where ambiguity would meaningfully change the output. The goal is to surface unstated intent rather than letting the model silently fill gaps based on its training data. The tool works as a standalone skill file and can also be invoked via a /clarify command. It is freely available on GitHub and requires no signup.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Cost Estimates vs. Timed Canaries: How to Gate AI-Generated SQL Before Production

A real incident involving an analytics agent that generated a plausible but costly SQL query against a 40-million-row table has sparked debate among database engineers about the best way to vet agent-generated SQL before promotion. The query passed staging and static review but caused excessive heap scans in production, exposing a gap that neither linting nor unit tests could catch. Two schools of thought have emerged: one favors using PostgreSQL cost estimates via EXPLAIN as a fast, lock-free filter in CI pipelines, while the other argues that timed canary runs against representative data are necessary because cost units do not reliably reflect real-world latency or data skew. Cost-based gates are seen as practical for high-frequency OLTP screening, but critics note that stale planner statistics can cause the planner to underprice a dangerous scan. The core question for teams deploying SQL-generating agents is which signal — estimated cost or measured wall time — should have veto power, and at what point the cost of collecting that signal becomes prohibitive.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Three Workflow Mistakes That Waste Dev Hours When Running AI Coding Agents

A developer reflecting on a lost Thursday shares lessons learned from running an AI coding agent against a cluttered local repository. Three core errors — a dirty checkout, improper isolation, and missing command logs — caused the agent to edit wrong files, carry over local secrets, and produce unrepeatable results. The author recommends creating a dedicated git worktree per ticket, running the agent on a separate isolated host, and capturing a structured receipt file that logs timestamps, file paths, and test outcomes. A replay step using rsync and pytest is proposed to verify that remote results match the local contract before any patch is merged. The workflow is presented as vendor-neutral shell scripts adaptable to any SSH-accessible host, with a disclosure noting the article was produced as part of outreach for MonkeyCode, which offers free model access and isolated server runs.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

macOS 26 Users Report Siri AI Process Persists Even After Full Disable Steps

A macOS user attempted to fully disable Siri by turning it off in System Settings, restricting it via Screen Time, and running a third-party debloat script targeting Apple Intelligence services. Despite these steps and a system restart, an 'Siri AI.app' process continued running in Activity Monitor. The user raised concerns that Apple's persistent background AI activity, combined with its Personal Context feature, resembles Microsoft's controversial Recall tool. The post has prompted discussion on Hacker News about whether macOS truly allows users to fully opt out of Siri and Apple Intelligence. No official Apple response or confirmed workaround has been reported.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Decart's Oasis Shows Neural Rendering Complements Game Engines, Not Replaces Them

In October 2024, Decart and Etched launched Oasis, a Minecraft-like demo that generates gameplay frames using a transformer-based neural network instead of a traditional physics engine. The system predicts each next frame from player inputs trained on millions of hours of footage, statistically imitating physics rather than simulating it. Early versions drew criticism for low resolution, hallucinations, and the inability to save or persist game state — fundamental limitations of having no explicit world state. Subsequent versions shifted focus away from replacing game engines, with Oasis 3 repositioning itself as an interactive world model for robotics and autonomous systems training. Analysts note the technology is better suited to complement traditional engines — handling visuals and style — while conventional engines continue to manage state, rules, and determinism.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

OSS Contributors Urged to Submit Reviewer Question Banks Alongside Code Patches

A proposed open-source contribution practice recommends that contributors include a structured 'Reviewer Question Bank' file alongside pull requests, rather than submitting bare code diffs. The artifact, a short Markdown file named REVIEWER_QUESTIONS.md, pairs each anticipated maintainer question with required evidence and a cited answer, keeping unanswered items flagged as TODO. The approach addresses a common review bottleneck where maintainers must spend scarce time reconstructing reproduction steps, public API changes, and error paths that contributors could have documented upfront. Unlike generic PR templates, the question bank is generated specifically from the branch's own paths and exit codes, making it patch-specific rather than boilerplate. Proponents argue this shifts a pull request from an unexplained diff into a reviewable evidence packet, reducing stalled reviews and improving the contributor-maintainer feedback loop.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Satellite IoT Bridges Connectivity Gaps for Remote Sensors Beyond Cellular Range

IoT sensors in remote locations like farmland, isolated tanks, and vehicle routes often lose connectivity, leaving monitoring platforms with outdated data. Satellite IoT addresses this by enabling devices to transmit readings via satellite links when cellular networks are unavailable. On August 6, 2026, Nordic Semiconductor announced that the LooUQ MTC2-N9151 embedded modem, based on the nRF9151 chip, received Skylo certification for Non-Terrestrial Network satellite connectivity. The modem offers a hybrid platform combining cellular and satellite communication, simplifying development for industries such as agriculture, logistics, and remote infrastructure. However, the effectiveness of such systems depends on the full hardware-software stack, including antenna design, firmware, power management, and how the backend handles stale versus live data.

0 comments Read more at DEV Community
← NewerPage 1048 of 3196Older →