12-Year-Old Dev Launches KODA Code Jam Challenging Beginners to Build Niche Websites

Harun, a 12-year-old solo developer, has announced the first-ever KODA Code Jam, a seven-day coding challenge running from August 27 to September 2, 2026. Participants must build a single-page website using HTML, CSS, and JavaScript targeting a niche audience, such as dog walkers, gamers, or students. A key requirement is that at least 50% of the code must be written with assistance from KODA, Harun's AI coding mentor tool. The winner will receive exclusive in-app rewards including an enhanced AI memory feature and a featured interview on Harun's DEV Community blog, which has over 577 followers. Results are scheduled to be announced on September 4, 2026, with submissions accepted via a live deployment link or GitHub repository.

How a company's switch to local AI models quietly broke its dev workflow

A software firm of around twenty developers initially found success using AI tools collaboratively, with developers retaining control over architecture and using AI for targeted assistance. Leadership then shifted to a one-developer-per-project model after the rise of vibe coding, expecting dramatic speed gains, but results fell short of expectations while AI subscription costs rose sharply. To cut costs, management mandated a switch to a single in-house local AI model running on one dedicated machine shared across all twenty developers. The setup proved inadequate for agentic workflows — where AI plans, edits multiple files, runs tests, and iterates — which the team had come to rely on, as one machine could not realistically serve twenty concurrent users at that capacity. Developers quietly resorted to using external AI tools for planning and reasoning, then feeding outputs into the approved local model, effectively circumventing the policy to get work done.

House price growth slows to 3.6% in Q1 FY27, says RBI; what’s driving the moderation

Meghalaya Trinamool Chief Asserts Legitimacy Amid Internal Party Crisis

The Meghalaya unit of the Trinamool Congress is facing a significant internal crisis, with questions raised about its support base and organisational strength. The state party chief has asserted that their faction represents the real Trinamool Congress amid the ongoing dispute. The controversy carries added weight given the approaching 2028 Meghalaya Assembly elections. The party's prospects in the state are being closely watched as the leadership struggle plays out.

Israel's Most Dangerous Mob Boss Yanis Yushvaev Killed in Point-Blank Shooting

Yanis Yushvaev, widely regarded as Israel's most dangerous criminal, has been shot dead at point-blank range. He had a long and notorious reputation in the country's criminal underworld. Yushvaev was known to have been involved in prolonged and violent blood feuds. His killing marks the end of one of Israel's most feared organized crime figures.

Alex Apodaca stuns Bella Mir in biggest upset in Contender Series history

Alex Apodaca defeated Bella Mir by unanimous decision at Dana White's Contender Series Week 3, in what is considered the biggest upset in the show's history. Mir entered the bout as a massive -6000 favorite but lost for the first time in her professional MMA career, dropping to 3-1. All three judges scored the contest 29-28 in favor of Apodaca, who improved his record to 8-1. The victory earned Apodaca a coveted UFC contract. Four other fighters also secured UFC deals on the same night.

Loan API Orchestration vs. Aggregation: Why Choosing Wrong Breaks Your Lending Stack

Software architects building loan origination platforms face a critical design choice between API aggregation and API orchestration, with significant consequences for compliance and functionality. API aggregation fans out parallel, stateless calls to multiple services and merges results, making it suitable for dashboards and reporting views where data sources are independent. API orchestration, by contrast, manages a strict sequence of dependent steps — identity verification, credit bureau pulls, underwriting, lender submission, disclosures, and e-signature — where each stage's output feeds the next. Lending workflows are governed by regulatory requirements such as Regulation Z, which mandates that TILA disclosures cannot be generated before a lender offer exists, making parallel execution legally problematic. Choosing aggregation for a workflow that demands sequential, stateful coordination results in an integration layer unable to enforce compliance timing, track state, or apply conditional logic across multi-lender submissions.

Chinese Artist Gao Zhen Sentenced to 3 Years for Satirical Artwork

Chinese artist Gao Zhen has been handed a three-year prison sentence by authorities in China. He was charged with "infringing upon the reputation of heroes" based on his satirical artworks. His wife described the conviction as deeply disappointing and unexpected. Supporters contend that the works in question were created before the law used to prosecute him even existed. The case is seen as a stark indicator of shrinking freedoms for artistic and political expression in China.

How to Evaluate AI Agents on Process and Tool Use, Not Just Final Output

A software developer has published an open-source framework for evaluating AI agents by assessing their step-by-step trajectories rather than only their final answers. Unlike standard LLM evaluation, which grades a single text output, agentic systems involve multi-step reasoning and tool calls where early errors can silently corrupt results even when the final response appears correct. To test this approach, the author built a bug-fixing agent, deliberately introduced 10 bugs, and constructed a custom evaluation harness around it. The framework distinguishes between capability evaluations, which probe the limits of an agent's abilities, and regression evaluations, which guard against breakage when prompts or models are updated. The author argues that generic benchmarks like MMLU or SWE-bench measure base model intelligence but cannot assess whether a custom agent wired to specific tools is performing reliably.

UK Developer Launches Buslens Tool to Map Bus-Accessible Destinations

A UK-based developer named Rupert Linacre has created a web tool called Buslens that helps users discover where they can travel by bus. The tool was shared on Hacker News as a community project under the 'Show HN' category. Buslens appears designed to visualize bus connectivity, allowing users to explore reachable destinations from a given location using public bus routes in the UK. The project received modest early engagement on Hacker News, garnering five points and one comment at the time of posting.

Dev Team Builds Tool to Auto-Generate Airflow DAGs from DataHub Metadata

Two developers created a tool called datahub-dag-generator during DataHub's Build with DataHub Agent Hackathon that automates the conversion of DataHub metadata into Airflow DAGs. The tool accepts a table name as input, traces the lineage graph backwards to identify all upstream dependencies, and generates corresponding Airflow task code. It operates in two modes: an LLM-powered agent mode using models like Claude via OpenRouter, and a script mode designed for CI pipelines that bypasses the LLM entirely. A key design choice keeps the LLM from writing code directly — it produces only a structured plan, which a separate deterministic renderer then translates into actual task code. The project is still in early stages, lacking dbt and Spark operator support, and has primarily been tested on the DataHub NYC Taxi sample dataset.

Why Bias Detection and Ethical Design Are Critical to Trustworthy AI

As AI systems increasingly influence high-stakes decisions in areas like hiring, lending, and healthcare, concerns about embedded bias and fairness have grown significantly. AI models learn from historical data, and if that data reflects societal inequalities, the resulting systems can perpetuate or amplify discrimination against protected groups. Researchers and developers identify several competing definitions of fairness — including demographic parity, equalized odds, and predictive parity — each with different implications depending on the use case. Bias detection involves systematically auditing AI systems to uncover hidden prejudices before they cause real-world harm. Ethical AI frameworks aim to make these systems not only more powerful but also more transparent, accountable, and equitable for all users.

Deloitte Agrees To Pay $21.5 Million In US Hiring Case: What Went Wrong?

Deloitte has denied wrongdoing. The company said it was pleased to resolve the matter to "avoid the cost and distraction of protracted litigation".

Boy in Surat survives after iron rod pierces shoulder and exits through mouth

A boy named Nazish survived a serious accident in Surat after an iron rod pierced his shoulder and came out through his mouth. The incident occurred when Nazish went to play with friends on the third floor of a nearby mosque. While attempting to jump from the mosque's upper level to an adjacent under-construction building, he lost his balance and fell to the ground floor. The fall resulted in an iron rod from the construction site impaling him in a near-fatal manner. He was rushed to medical care and survived the severe injury.

HDFC Bank, Federal Bank revise FD rates

OpenAI Uncovers and Disrupts Covert Russian AI-Driven Influence Operation

OpenAI has identified and disrupted a covert influence campaign linked to Russia that was exploiting its AI tools. The operation involved the misuse of OpenAI's technology to generate content intended to manipulate public opinion. OpenAI detailed its findings and the steps taken to shut down the activity in a published report. The disruption is part of the company's broader effort to prevent malicious actors from weaponizing its platforms for disinformation purposes.

JSW MG Motor India Launches Hector Tomahawk SUV with 517 km EV Range

JSW MG Motor India has introduced the Hector Tomahawk, a new seven-seater SUV available in both pure-electric and plug-in hybrid variants. The electric version claims a driving range of 517 km on a single charge. The vehicle is underpinned by MG's new ADAPT platform, which supports multiple electrified powertrain configurations. It comes equipped with advanced connected technology, including what the company describes as India's first generative AI system with a karaoke feature.

GitNexus (Akon Labs)

Discussion | Link

Hyundai's 26-car blitz is coming to India

BJP Plans Anti-Drug Yatras Across All 117 Punjab Constituencies Ahead of 2027 Polls

BJP national general secretary BL Santosh addressed party leaders and workers at a meeting in Ludhiana, Punjab. He urged the party to organize anti-drug yatras across all 117 Assembly constituencies in September. The initiative appears to be part of BJP's early groundwork for the 2027 Punjab Assembly elections. The drug issue is being positioned as a central campaign theme to mobilize voters at the booth level.