SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

Essential Self vs. Simulated Subject: Where Identity Meets AI

The article examines how human identity relates to artificial intelligence, drawing a distinction between the essential self, the nominalized subject, and the simulated subject. It cautions that handing over one's thinking to outside systems, a process the author calls nominalization, can erode genuine human agency. The piece also criticizes treating AI systems as human-like when they only imitate behavior. Accurately modeling someone's style or data, the author argues, does not amount to personal continuity or conscious experience. As AI agents spread, the author stresses keeping a clear line between a person and their digital model to preserve intellectual independence and authentic personal identity.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Eight AI API providers offer free tiers without a credit card as of October 2026

As of October 22, 2026, eight providers offer free AI API access that requires no payment method: Google Gemini, Groq, OpenRouter, Mistral, Cloudflare Workers AI, NVIDIA's API catalog, Cohere and SambaNova. Six other providers offer starter credits instead, which are one-time trials rather than ongoing free tiers. All eight use OpenAI-compatible chat completion endpoints, so a single client can work across them. Each free tier has limits on requests per minute, requests per day or tokens, which are set by the providers and may change. Google warns that data sent on its unpaid tier may be used to improve its products and reviewed by humans, so sensitive information should be avoided.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Most long-tail PyPI packages still ship wheels, but source-only releases are far more common there

A census of PyPI's latest releases compared the 3,000 most-downloaded projects with a random sample of 2,850 long-tail projects. About 81% of tail projects ship a wheel, versus 98% of top projects, so the gap is real but smaller than commonly assumed. Source-only (sdist-only) releases are roughly eight times more common in the tail, at 18% compared with 2%, reflecting older setup.py uploads that never adopted wheels. Top projects also tend to ship more files per release, largely because of compiled-package families with many platform-specific wheels. The analysis is a single snapshot from October 2026 and the tail figures carry about one point of sampling error.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

gcloud ADC Credentials Expired After ~44 Hours; Service Account Key Fixed Scheduled Jobs

A blog's automated job, which pulled Search Console data using gcloud application-default credentials, stopped working about 44 hours after setup. The failure was a "Reauthentication failed" error, since the login used a Google Workspace account whose policy periodically requires re-entering the password. Because a scheduled job has no screen, it cannot answer that prompt, so the credentials became permanently unusable. The author resolved this by switching to a service account key, which neither expires nor requires reauthentication, and exchanging a self-signed JWT for an access token using only Node's crypto module. The service account was given no Google Cloud roles, only read access granted in Search Console, and the key was stored outside the repository with restricted file permissions. The switch also involved several 403 errors, which made it difficult to determine whether the cause was a missing scope, a wrong account, or a wrong project.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Gemini API free tier in 2026: which models are included and when billing starts

Google's Gemini API still offers a free tier as of October 2026, but it is more restricted than before. Developers can call several Gemini Flash and Flash-Lite models, Gemma 4, and certain audio and embedding models at no cost, without a credit card or billing account. Gemini 4 Argon, image models, and video models are paid-only, and the free tier is granted per project and per model rather than per API key. Google no longer publishes per-model RPM and RPD figures in its documentation, and charges apply only after a billing account is linked and at least $5 is prepaid. The article notes that access to Gemini 2.5 models may be restricted for new projects, and that developers are being steered toward newer models.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

PHP2Shell: two unauthenticated PHP RCE bugs disclosed, with attacks already underway

Two critical PHP vulnerabilities, CVE-2026-109618 and CVE-2026-109619, were disclosed on October 21, 2026, both enabling remote code execution without authentication and rated CVSS 4.0 9.2. Exploitation is already occurring in the wild. Upstream PHP fixed the issue only in versions 8.5.12, 8.4.27, 8.3.36 and 8.2.35, so other users depend on their Linux vendors' backported patches. Because distributions such as Ubuntu keep the old upstream version number while applying fixes, checking php -v is unreliable; administrators should inspect package release versions instead. Testing of edge mitigations on Rocky Linux with NGINX and PHP-FPM found that default request buffering and a limited client_max_body_size help block some attack paths, while per-request PHP_VALUE settings did not work as expected.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Aldi cheapest on weekly 31-item grocery basket, Coles dearest by 22%

A weekly price index comparing Coles, Woolworths and Aldi found Aldi had the lowest total for a 31-item staples basket at $105.40 on 22 October 2026. Coles was the most expensive at $128.57, about 22 percent more than Aldi, while Woolworths came in at $123.81. The comparison uses online shelf prices for the cheapest matching product per item, measured per kilogram or litre so pack sizes do not skew results. The report is rebuilt weekly so that price movements can be tracked over time.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why Standard HTTP Status Codes Make APIs Easier to Use and Debug

HTTP status codes signal the outcome of a request, and using them consistently helps developers consume and troubleshoot APIs. The article highlights 200 OK for success and 201 Created when a new resource is made, such as after a POST to a users endpoint. Client-error codes like 400, 401, 403, 404, 409, and 429 indicate that a request was unacceptable, for instance when authentication is missing or permissions are lacking. Server-side failures are represented by codes such as 500 Internal Server Error and 503 Service Unavailable, the latter often signaling temporary downtime. The author argues that returning 200 for every response forces clients to parse custom fields, so standard codes should be treated as part of the API contract.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Disney Raises Streaming Prices Days After Saying Ads Will Reach All Tiers

Disney has announced a price increase for its streaming service just days after confirming that advertisements will now appear on every subscription tier. The change means subscribers on previously ad-free plans will see commercials alongside their content. The news was shared on Hacker News, where it drew discussion from readers. The source article, published on a TV fan site, reports the pricing move and links it to the broader push toward advertising revenue. Further details on exact new prices and effective dates were not included in the provided text.

0 comments Read more at Hacker News
0
ProgrammingHacker News ·

ClickBench Pull Request Calls for Removing pgrust Entry Over Alleged Cheating

A pull request on the ClickBench GitHub repository proposes removing the pgrust entry from the benchmark. The request argues that the submission is cheating, meaning it does not represent a legitimate performance result. The article itself is brief and provides only the pull request link, with limited detail about the specific methods alleged. The discussion gained modest attention on Hacker News, with 8 points and one comment at the time of posting. Readers should consult the pull request directly for the full reasoning behind the accusation.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Developer shares how to build a native-feeling six-box OTP code input

A DEV Community tutorial describes common usability problems with six-box one-time code inputs, such as paste landing only in the first box and backspace failing on empty cells. The author explains that their component treats the code as a single digit string that the six inputs merely display, so typing, pasting, and iOS autofill all update the same value. Key behaviors include moving focus to the first empty cell, replacing digits in filled cells, stripping non-digit characters from pasted codes, and shaking the field with a reset on an incorrect code. The article also covers accessibility attributes such as inputMode, autoComplete set to one-time-code on the first cell, and ARIA labels for screen readers. Code excerpts from the shipped React component illustrate the approach.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How ISO 27001 and ISO 42001 Can Be Built Into AI Development Workflows

Governance gets harder when security and compliance work sits apart from the engineering process, especially for AI systems that raise questions about intended use, model behaviour, and risk ownership. The article argues that ISO/IEC 27001 (information security) and ISO/IEC 42001 (AI management) can help organizations close the gap between technical implementation and governance. It explains that standard security controls alone do not address issues such as sensitive data exposure in model inputs, the need for human review of outputs, or post-deployment monitoring. The piece recommends connecting these management-system activities to existing development lifecycle stages, starting with planning questions about purpose, data, dependencies, and ownership. It also suggests that testing and review should cover both security and AI-specific concerns, with documented results and clearly identified approvers.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Browser image loop decodes, draws, encodes and saves photos one at a time, with per-step timings

A developer reviewed a teammate's in-browser loop that converts a folder of product photos to JPEGs for an upload form, measuring each step with timers. The loop processes one photo at a time and releases the decoded bitmap after drawing it to a canvas, keeping memory low compared with decoding everything first, which had reached about 5,192 MB by photo 100. Tests on six valid photos plus two invalid files showed that corrupt images are caught at the decode step by a try/catch, so processing continues rather than stopping early. Measurements across two runs put decoding at roughly 24–29 ms per photo and encoding at 46–52 ms, with drawing and saving much faster. The author notes that the progress display should show each file's label, since files arrive in input order rather than sorted order.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Vibe coding moves to phones and enterprise tools, but review safeguards lag behind

Vibe coding, once a personal experiment on home computers, expanded in October into mobile apps, company tools, and autonomous agents. Google Play now lets users build and privately run Android apps made in AI Studio without a paid developer account, though the apps are not publicly listed. A VentureBeat-published survey of 307 tech leaders, sponsored by Retool, found that few organizations have full visibility into internal tools or governance over AI-generated code. GitHub made local sandboxing for Copilot agents generally available, but it is off by default, leaving commands largely unrestricted until enabled. The author argues that the speed of generating code now needs matching review processes and stopping rules, rather than abandoning the practice.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Symfony Bundle Skeleton Refresh Grows Into a New Generator Built by AI Agents

A Symfony developer set out to modernize the bundle skeleton created by a Symfony core team member, which had not been updated since November 2024. Since then, Symfony 7.4 LTS, 8.0 and 8.1 had been released, and the generator still declared support for older versions and lacked testing, CI and QA tooling. Instead of a minor refresh, the project turned into a completely new generator, with the code written by AI agents working in parallel git worktrees while the author made decisions and reviewed results. The work was coordinated through a planning phase, specialized agents for different components, an integration step to merge branches, and independent reviewers who checked for problems. The author has disclosed the AI involvement openly in the commit history.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Batch photo processing in the browser: memory spikes from the final ZIP step

A developer tested browser memory use while re-encoding a 300-photo batch to JPEG, measuring total usage across all browser processes. Decoding every photo up front made memory grow with folder size, while releasing each bitmap after use kept it near one photo's footprint. Keeping all re-encoded results for a final ZIP pushed memory up sharply at the end, adding about 830 MiB against a 428 MiB archive. Writing each file to disk instead showed memory climbing in the browser outside the JavaScript heap, a cause the author had not yet identified. The author also noted an unexplained startup memory baseline that appears across runs.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

APEX Compute: A Zero-Dependency C++20 Engine for Vectorized Matrix Multiplication

APEX Compute is a C++20 engine built from scratch with no external dependencies, designed to run cache-tiled, explicitly vectorized matrix multiplication (GEMM) on both ARM and x86 processors. The project pairs its low-level SIMD kernels with an automated benchmarking tool, apex-bench, and a differential validation suite, apex-validate, to check results for correctness. It also offers an optional natural-language layer powered by Gemma 4, which converts a user's stated intent into a structured hardware run plan. The project's source code is available in its public repository.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Solo Developer Packages 39 Digital Products on a 2 GB, Single-Core Server

A developer built a pipeline to package 39 digital products totaling 856 MB on a server with one vCPU, 2 GB of RAM, and no GPU. After a parallel batch job crashed the machine, they shifted to running one process at a time, freeing memory between items, and measuring peak memory usage. Four scripts handle metadata for stock platforms like Adobe, Vecteezy and Dreamstime, upload tracking to prevent double posting, Markdown-to-PDF conversion, and cover and thumbnail creation. Two of the four scripts rely only on Python's standard library, while the others need two third-party packages. The article presents this fixed-structure approach as a way to keep a growing product catalogue organized.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Python ile borsa veri analizinde kullanılan sekiz temel paket

Bir yazar, borsa verileriyle çalışırken karşılaştığı karmaşayı aşmak için kullandığı sekiz Python kütüphanesini paylaşıyor. Pandas'ın veri temizleme ve zaman serisi işlemleri için nasıl kullanıldığı, eksik işlem günlerinin reindex ve ffill yöntemleriyle doldurulması örnekleniyor. Yazar, büyük veri setlerinde bellek yönetiminin önemine ve veri tiplerini optimize etmenin gerekliliğine dikkat çekiyor. Yahoo Finance üzerinden veri çekiminde API sınırlamalarının yarattığı engeller ve veri doğrulamanın önemi de vurgulanıyor. Ayrıca zaman serilerinde look-ahead bias hatasının tahmin modellerini nasıl geçersiz kılabileceği ve teknik göstergeler için Technical Analysis Library'nin RSI hesaplamasındaki kullanımı anlatılıyor.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Checklist for Agencies Uploading Client Documents to AI Tools

Agencies should not upload client documents to an AI tool until they know where the text is sent, how long it is retained, whether it is used for model training, who else processes it, and what their client contract permits. These answers vary widely between personal chat accounts and business products built on commercial APIs. The article offers a vendor-neutral checklist covering storage, training, model providers, sub-processors, location, access, deletion, export, contracts, and independent assurance. It stresses that the guidance is practical rather than legal advice, and that client contracts and counsel decide what is permitted. Discovery material can be more sensitive than it appears, and some of it may count as personal data under laws such as the GDPR.

0 comments Read more at DEV Community
← NewerPage 93 of 2918Older →