SShortSingh.
0
TechnologyThe Verge ·

Nvidia CEO Jensen Huang steps down after affair allegations and spy ties

Nvidia's board announced that co-founder Jensen Huang has resigned as CEO, president, and board member, effective immediately. The resignation follows reports of two extramarital affairs, one involving an alleged Chinese spy and another with an Nvidia executive who reported to him. After a late-night board meeting, directors accepted his letter, and CFO Colette Kress, who has been at Nvidia since 2013, was named interim CEO. The company is also restructuring its leadership, with former Synopsys CEO Harvey C. Jones among the new appointments. The article is truncated, so further details on the corporate changes are not included here.

0
ProgrammingHacker News ·

NVIDIA Announces CEO Transition

The article text provided contains only a link to NVIDIA's newsroom and a Hacker News discussion thread, without the details of the announcement itself. Based on the material available, the headline indicates NVIDIA has disclosed a change in its chief executive leadership. The specific names of the outgoing and incoming executives, the effective date, and the reasons behind the transition are not included in the source text. Readers would need to consult NVIDIA's official announcement for these details. The discussion on Hacker News has drawn 387 points and 267 comments, reflecting notable interest in the news.

0
ProgrammingDEV Community ·

Marcelo Alberto Futerman explica cómo usar una IA para presentar ideas en el trabajo

Un artículo publicado en DEV Community por Marcelo Alberto Futerman aborda el uso de herramientas de inteligencia artificial para preparar la presentación de una idea en el ámbito laboral. El texto está acompañado de enlaces a su sitio web, su perfil en DEV y su canal de YouTube. La nota es breve y no detalla en su contenido visible los pasos concretos ni los ejemplos del método propuesto. El material parece orientado a profesionales que buscan apoyo tecnológico para comunicar propuestas dentro de sus organizaciones.

0
ProgrammingDEV Community ·

Anthropic Files for IPO with SEC, Offering Terms Not Yet Disclosed

Anthropic has submitted a registration statement and prospectus to the U.S. Securities and Exchange Commission in connection with a planned initial public offering. The announcement confirms the filing but leaves out key details such as share count, expected price, timing, exchange, and ticker symbol. Investors are directed to the prospectus for further information. The filing does not signal changes to Anthropic's products, APIs, pricing, or service terms. Analysts note that a public listing could affect the company's access to capital and its visibility in the AI sector.

0
IndiaNDTV ·

Brooklyn Building Collapse Kills 8; Pregnant Woman Pulled From Rubble

A building constructed in 1930 collapsed entirely into an adjacent excavation site in Brooklyn, New York. The incident occurred at 2625 Bedford Avenue at approximately 4 pm local time. Eight people have died in the collapse, according to the reported details. Rescuers managed to extract a pregnant woman from the debris alive. The article provides limited information beyond these core facts.

0
ProgrammingDEV Community ·

Come tenere aggiornato il team di osFoundry sui progressi di Claude Code

Il testo descrive come Claude Code possa lavorare su un fix nel repo di un utente mentre una sessione di team di osFoundry attende il risultato. Lo strumento legge la posta della sessione, mantiene uno stato di una riga con repo, branch e task, e pubblica note di avanzamento visibili nella scheda Team mail. La connessione richiede un'approvazione una sola volta per workspace, mentre codice e repository restano sulla macchina dell'utente. Viene spiegato come collegare Claude Code tramite il comando /mcp, aggiungere il progetto e assegnare il ruolo di Contributor alla sessione. Infine, vengono illustrate alcune istruzioni da inserire nel file CLAUDE.md del repo per guidarne il comportamento.

0
ProgrammingDEV Community ·

Developer tests GitHub MCP Server to see AI agents create pull requests directly

A developer wrote a hands-on guide to the Model Context Protocol (MCP), a set of rules that lets AI tools take actions rather than only giving advice. Using GitHub's official MCP server inside VS Code, they cloned a repository, created a test branch, and edited the README. They installed the MCP extension, which generated a configuration file listing the operations the AI is permitted to perform, and then asked Copilot to open a pull request. The developer reported that Copilot successfully created the PR, which was confirmed on GitHub. The author notes that many other MCP servers exist and plans to try them later.

0
ProgrammingDEV Community ·

AI coding agents opened 427,042 public GitHub pull requests in 24 hours

Between November 11 and 12, 2026, AI coding agents opened 427,042 pull requests on public GitHub repositories, about 31.9% of the 1,339,893 opened there in that period. The count relies on the marks four popular agents leave by default, namely a footer line or a fixed branch-name prefix, and was taken directly from GitHub's public search API without sampling. Over the prior 30 days, agent pull requests had a merge rate that varied by agent from 68.5% to 92.6%, though most activity came from repository owners merging their own work. The author notes the figure is a floor, since private repositories, unmarked agent work, and bot-account pull requests are not captured, and it does not measure code quality. The count is published on a live board run by MyClawn, which the post discloses as its own project.

0
ProgrammingDEV Community ·

Seven Open-Source GitHub Repos That Add Browsing, Memory, Skills and Security to AI Coding Agents

Coding agents often struggle with web interaction, forgetting prior context, lacking security checks, and producing diagrams, the article notes. Seven open-source GitHub repositories are highlighted as tools that fill these gaps. They include browser-use for driving real browsers, agentmemory for persistent context across sessions, and scientific-agent-skills for reusable on-demand instructions. Other picks add editorial-quality diagram generation and 817 cybersecurity skills mapped to MITRE ATT&CK and NIST CSF 2.0, which the author says should supplement rather than replace human review.

0
ProgrammingDEV Community ·

Developer Launches CyberNova, a Responsive Portfolio Hosted on GitHub Pages

A developer has built and deployed a personal web portfolio called CyberNova to showcase their web development projects and technical articles. The site was created with HTML5, CSS3, and JavaScript and is hosted live on GitHub Pages. The creator emphasized a clean, fast, mobile-friendly design and set up a custom sitemap, verified ownership through Google Search Console, and added meta verification tags to aid search indexing. The developer is inviting the community to review the layout, performance, and user interface and share feedback.

0
ProgrammingDEV Community ·

Shared agent memory lets coding agents reuse fixes other agents already found

Coding agents typically start each session with no memory of what other agents have learned, so the same bugs get rediscovered across a company. The author argues the solution is a shared 'company brain' where agents post lessons after working and search before starting. In one example, a webhook signature change at Shipfast caused 401 errors that only a teammate's agent in another repo had already solved. Benchmarks reported by the author show success on related bug families rising sharply with shared memory, for example Sonnet going from 12% to 79%. The author's project, Big Agentic Brain, is an open-source, self-hosted tool that runs as one container and is accessible to any agent supporting MCP.

0
ProgrammingDEV Community ·

Claude Code adds advisor mode: Opus 5.5 does the work, Fable 5.2 weighs in on key decisions

Claude Code's advisor feature lets a main model handle routine work while a stronger model is consulted only at pivotal moments, such as planning, recurring errors, and final review. With Opus 5.5 as the main model and Fable 5.2 as advisor, the advisor receives the full conversation, including tool calls and results, and returns guidance that the main model applies, while flagging conflicts with its own evidence. The advisor can be enabled per session with a command, saved as a default in the settings file, or set with a launch flag, and an advisor must rank at or above the main model. Pairings are checked before each request, and Fable may need a one-time consent step on some plans. Claude decides when to consult the advisor, though users can request consultations directly.

0
ProgrammingDEV Community ·

A2UI 1.0 Currency Agent Runs End to End on AWS with Amazon Nova

A step-by-step guide describes a currency agent that returns generated user interface components, running entirely on Amazon Web Services. A Strands agent powered by Amazon Nova 2 Lite retrieves exchange rates and expresses its answer as A2UI 1.0 components, which a Next.js front end built with CopilotKit renders as cards and charts. The agent is hosted on Amazon Bedrock AgentCore Runtime and the interface on Amazon ECS Express Mode, and it is the AWS counterpart to a Google ADK and Gemini version of the same app. In tests of ten prompts run ten times each, all 100 replies rendered correctly both locally and through the deployed application. The browser never handles AWS credentials, since a server-side proxy signs requests and streams the agent's events back.

0
ProgrammingDEV Community ·

PostgreSQL Outperforms MySQL in Most Benchmarks, With Caveats

A developer ran PostgreSQL 16.15 and MySQL 8.0.46 on the same two-vCPU, 8 GB machine with identical data and comparable memory settings, testing typical web workloads. PostgreSQL delivered higher throughput in every sysbench transactional test and ran aggregate queries and joins over 2 million rows five to six times faster. MySQL was roughly twice as fast on single primary-key lookups from PHP and had cheaper connection setup, and PostgreSQL's table size grew after heavy updates while MySQL's stayed flat. The author stresses that the results come from one small machine, specific software versions and largely default settings, with run-to-run variation of 20 to 30 percent. MySQL 8.0 has reached end of life, so the findings reflect a snapshot rather than a general verdict.

0
ProgrammingDEV Community ·

How Claude Code cloud environments cache setup scripts to speed up sessions

Claude Code cloud sessions start in a fresh virtual machine whose contents are determined by the environment's network access, variables, secrets, and setup script. The setup script runs once and its filesystem is saved as a snapshot, which later sessions reuse, so setup time is paid only occasionally. Changing the setup script or allowed network hosts triggers a rebuild, as does cache expiry after about seven days. Only file-based results such as installed packages and pulled images are cached, while running services and scripts exceeding roughly five minutes are not, which is why those should be handled by a SessionStart hook instead. Network access can be set to none, trusted (the default), full, or a custom domain list, and secrets can be stored as network secrets that are injected by a proxy rather than exposed as environment variables.

0
IndiaTimes of India ·

Hinojosa Says She Will Work With Musk's Companies if Texans Come First

Gina Hinojosa, Texas's newly elected governor, has said she is willing to work with Elon Musk and his companies, but only if the work prioritizes the interests of Texans. She had previously criticized the outgoing governor over proposals involving public land transactions near SpaceX operations. After winning the election, she asked that decisions on these state land matters be paused during the transition period. Musk, for his part, argued that Texas offers a more favorable business climate than California.

← NewerPage 255 of 7335Older →