SShortSingh.

Programming

0
ProgrammingDEV Community ·

Claude Code, Codex, Gemini CLI and OpenCode Use Different Names for the Same Commands

A developer guide compares nine everyday slash commands across Claude Code, Codex CLI, Gemini CLI and OpenCode, showing that the same job often has a different name in each tool. For example, context compaction is called /compact in three agents but /compress in Gemini CLI. The comparison, based on each vendor's documentation as of October 15, 2026, also covers commands for switching models, planning, resuming sessions and checking usage. The guide notes that some tools lack a direct equivalent, such as OpenCode, which uses a separate Plan agent instead of a /plan command. It also points out that launch flags for the same tasks differ further, with Codex using a subcommand for resuming sessions where the others use flags.

0
ProgrammingDEV Community ·

Zod .refine() Rules Are Silently Dropped From JSON Schema Sent to LLMs

Passing a Zod 4 schema to structured-output helpers such as the OpenAI Node SDK or Vercel AI SDK silently discards every .refine() rule. The SDKs convert schemas with z.toJSONSchema(), and the converter only maps checks it recognizes, while custom refinements are stored under an unrecognized name and skipped. The refinements still run when the response is parsed, but the model is never told about them. The author illustrates this with a refund extraction schema, showing that regex, positive-number, enum and date rules survive while a field-level refinement and a cross-field date check do not. Transforms are also lost in input mode, leaving only a plain string.

0
ProgrammingDEV Community ·

Tenen Argues AI Is Externalized Human Intellect Built on Collective Labor

An analysis of Dennis Yi Tenen's book examines artificial intelligence as a form of intellectual externalization rather than an autonomous entity. The author argues that AI systems are the product of collective labor, drawing on the efforts of engineers, researchers, and largely invisible data workers. The piece draws on theories of distributed cognition and challenges the tendency to anthropomorphize AI. It stresses that these systems are socio-technical, shaped by the historical and social structures through which knowledge is produced. Overall, the analysis promotes a model of human-technology collaboration over the myth of independent machine intelligence.

0
ProgrammingDEV Community ·

Empty litigation searches must be kept separate from unmeasured ones

A litigation search table with no case rows can reflect a completed empty search, a failed request, a partial jurisdiction response, or a missing field, and collapsing these into "no litigation" hides what analysts most need to know. The Company Litigation Coverage Dossier organizes supplied observations for an explicit company query, recording cases, jurisdictions, contradictions, and evidence without running a new court search or assigning a legal risk score. Cases are deduplicated by jurisdiction plus exact source URL or court and case number rather than by title or date, and the query text is treated as a search identity rather than a verified legal entity. The guidance stresses that a completed search with an explicit empty list counts as a measured zero, while errors or absent lists do not, and that a full-text mention is not evidence of party status. Integration testing is recommended across four fixtures, covering an empty success, an error, a partial result, and duplicate observations, with outputs checked against source records.

0
ProgrammingDEV Community ·

The Hidden Costs of Go Interfaces: itabs, Escape Analysis and Heap Allocation

A hexagonal architecture port in Go is an interface, and its runtime cost is not simply the dispatch itself. An interface value is two machine words, an itab pointer and a data pointer, and calls go through an indirect function pointer. The main penalty is that the compiler cannot see through the interface, which blocks inlining and makes escape analysis conservatively send pointer arguments to the heap. Storing non-pointer values in interfaces can also trigger boxing allocations, as with integers above 255. The article notes several of these issues remain open in the Go compiler and suggests measuring the effect yourself with a small benchmark and -gcflags=-m.

0
ProgrammingHacker News ·

New Withings Users Face $99 Annual Fee to Sync with Apple Health

Withings, the maker of connected health devices, has reportedly begun requiring new accounts to pay $99 per year to sync data with Apple Health. The change was flagged in a post on Withings' support community, where a user described the paywall as unacceptable. The source material is limited, with only a brief community post and a small Hacker News discussion thread available. Details such as the exact effective date and whether existing users are affected were not confirmed in the available information.

0
ProgrammingDEV Community ·

AI coding agents add config flags to dodge product decisions, and that costs teams later

A developer observes that AI coding assistants often respond to feature requests by adding multiple configuration flags instead of making a decision. The author argues this happens because training data treats flexibility as a virtue, so agents expose uncertainty as knobs rather than committing to defaults. Each unnecessary flag creates untested combinations, pushes product choices to runtime where on-call engineers must resolve them, and gradually produces a sprawling, poorly understood configuration surface. To catch this in review, the author asks whether hardcoding would break anything, whether any real deployment needs a different value, and whether the choice is tenant-level or a runtime engineering default.

0
ProgrammingDEV Community ·

Empty Video Rooms Are Cleaned Up Through Leases and Periodic Reconciliation

A technical guide argues that video rooms should not be deleted the moment the last participant leaves. Instead, the room is marked empty and held through a bounded grace period, followed by a second authoritative occupancy check before teardown. The article contends that deletion should be modeled as a state transition rather than a callback, with leave events starting suspicion and later observations confirming absence. It recommends expiring presence leases stored in an authoritative system, separating ephemeral signals like typing indicators and read receipts from connection state, and running a periodic reconciler as the repair path. The author also suggests measuring how long rooms stay empty rather than counting successful callbacks.

0
ProgrammingHacker News ·

White House Press Secretary's 'I'm a Lawyer' Remark Draws Florida's Attention

A Substack post by Chaz Stevens addresses a claim by the White House Press Secretary that they are a lawyer. The piece suggests that Florida officials may want to weigh in on the matter. The source material provided is limited to the headline, a link, and basic Hacker News metadata, so specific details of the claim and any Florida response are not fully available. The discussion on Hacker News had drawn 20 points with no comments at the time of capture. Readers would need to consult the original article for the full context and substance of the remarks.

0
ProgrammingGitHub Blog ·

GitHub Says Ambient Intelligence Will Be the Next Developer Interface

GitHub's blog argues that software development is moving beyond traditional tools toward ambient intelligence. Artificial intelligence, agents, voice input, and physical controls are being combined to change how developers work. In this model, developers express their intent rather than operating each tool manually. The shift is intended to help developers stay focused on building software. The source text is brief, so it offers little detail on specific products or timelines.

0
ProgrammingHacker News ·

Blogger Details Repeated Tampering With His Sonarr Setup by Torrent Site User

A developer published a blog post describing how someone repeatedly interfered with his Sonarr media-automation setup through torrent sites. The account is shared via a Hacker News link with only limited metadata available, so the specific methods and timeline are not confirmed here. The post has drawn a small amount of attention, with one comment so far. Readers wanting the full details of the incident are directed to the original blog post.

0
ProgrammingDEV Community ·

Short link creator treats slug collisions as negotiations, not errors

The developer of a short link creator built custom slugs for AI agents, but initially returned a hard error whenever a slug was already taken. A collision with an abandoned link caused retry storms that saturated the rate limiter, prompting a redesign. Collisions now return the existing link with a flag, letting callers accept an auto-suffixed variant or reclaim a slug only if it has had no clicks in 90 days. The developer also found that agents losing their one-time manage secret create duplicate links for the same destination, and proposed making creation idempotent on destination and owner hint. The overall lesson is that link APIs should treat slug collisions, lost secrets, and moved destinations as normal operation.

0
ProgrammingDEV Community ·

Six documented ways to simplify a cluttered Zoho CRM

Zoho CRM can feel overwhelming when every user sees the same fields and records, so the platform offers several built-in ways to tailor the view for each role. Page layouts let a single module serve different processes, with distinct fields, sections and pick-list values assigned to profiles, though enabling this requires the Module Customization permission and edition checks. Layout rules can show or hide fields and sections or make them mandatory based on other entries, but they cannot remove fields and do not apply to imported, web-form or lead-convert records. Field permissions per profile let administrators set fields to read and write, read only, or hidden, with hidden fields blocked from viewing and editing rather than deleted. Custom list views, Kanban views and Canvas views then present records in task-focused formats, with the advice to keep columns under ten and build one view per daily need.

0
ProgrammingDEV Community ·

NTP Clock Step Cost a Batch Job 40 Seconds of Its Budget, Field Notes Show

A developer's 20-minute batch job kept stopping early on a host where every API call was fast. The cause was deadline accounting that read the wall clock, which shifted during the run and made the job think more time had passed than it had. Switching to a monotonic clock would have prevented the early exits, and the author also points to a separate retry bug in parsing the Retry-After header. The author also notes the difficulty of confirming the clock adjustment's cause, since the affected server was not theirs to inspect. The post closes by describing how injecting the clock makes such failures testable on a local machine.

0
ProgrammingDEV Community ·

Chrome extension swaps DOM scraping for site's internal API, cutting AI Studio export from 150s to under 1s

A developer built AI Studio Exporter, an open source Chrome extension, to export Google AI Studio conversations, which the site does not support natively. The first version scrolled the page and read messages from the DOM, which was slow, could miss messages, and captured only thumbnails of uploaded files. The second version instead calls the same undocumented internal API the AI Studio frontend uses, reducing exports of a 200-message conversation from about 150 seconds to under one second. Because the API's responses are unlabeled nested arrays, the developer built a decoder and tolerant helpers that degrade gracefully if Google changes the format. The extension runs a small script in the page's main context to get around a CORS restriction that blocks requests from extension content scripts.

0
ProgrammingDEV Community ·

Python Script Generates Adobe, Vecteezy and Dreamstime Stock Metadata CSVs

A developer describes writing a standard-library Python script that produces correctly formatted metadata CSVs for Adobe Stock, Vecteezy and Dreamstime from a single asset folder. The author explains that each platform requires a different column layout, and that an incorrect header causes silent rejection of the whole file. The script fills in default titles from filenames, though the author stresses that real titles and keywords should be supplied via a JSON file before submission. A companion ledger script tracks completed uploads to prevent the duplicate submissions that drew reviewer scrutiny. The whole pipeline reportedly runs on a single-vCPU, 2 GB server without a GPU.

0
ProgrammingDEV Community ·

Fixing Stock Platform CSV Rejections and Duplicate Uploads With Two Python Scripts

A developer lost a Sunday to a Dreamstime rejection caused by a malformed CSV header, which prompted a metadata generator tailored to each platform's exact schema. Adobe Stock, Vecteezy, and Dreamstime each require different column counts and strict header naming, so a single universal CSV failed everywhere. The solution writes only the header each platform expects, and the author notes that auto-generated titles from filenames should be replaced with real ones to avoid quality-based rejections. To prevent double-posting when batches are re-run, a second script keeps a plain-text ledger that marks each asset as done immediately after upload. The tools use only the standard library to run comfortably on a small, low-memory server.

0
ProgrammingDEV Community ·

Gemini Canvas Shows Parameter-Driven 3D Design with STL Export

Google Gemini Canvas has demonstrated a workflow that moves from conversational design to 3D-printable output. In the showcase, an app built in Canvas lets users adjust mathematical parameters for an object such as a vase, view the updated 3D model, and export it as an STL file. The STL export comes from the app itself rather than being a native feature available for every Canvas project. Analysts note this is not a general-purpose CAD replacement, and it works best for repeatable object families with clear variables. The app's code must define the geometry and handle the export, so output quality depends on its design logic.

0
ProgrammingDEV Community ·

Why a model script that passed locally can fail on a remote server

A developer found that the same code, prompt, and environment file can behave differently on a laptop versus a remote workspace. The article argues these failures usually stem from hidden assumptions about time, state, and identity rather than outright bugs. Hosted model endpoints are unpinned dependencies: providers can change aliases, retire endpoints, or alter routing behind an unchanged URL, so a version string does not guarantee stable behavior. The author recommends recording baseline outputs, hashing them, and running a drift-check script repeatedly across time and environments to detect changes. Remote sessions can also lose in-memory state, so jobs should use append-only checkpoints to be restart-safe.

0
ProgrammingDEV Community ·

Apple Intelligence reportedly set for China in December with Alibaba models

Apple plans to bring Apple Intelligence to mainland China in December 2026 with the iOS 27.3 update, according to a Bloomberg report. The China version would use Alibaba's AI models rather than the Google Gemini models used in Western markets. The features have been absent from iPhones in China for about two years due to regulatory delays. Reports indicate Chinese regulators have granted network access permits for Alibaba's Qwen and Tongyi Wanxiang models, though outlets differ on timing and feature details. Apple has not confirmed the plan, and Siri AI and Image Playground are reportedly excluded.

← NewerPage 413 of 3107Older →