Santner skips BBL to focus on New Zealand Test and white-ball duties

New Zealand allrounder Mitchell Santner has turned down an opportunity to play in the Big Bash League in order to prioritize his international commitments. The spinner is keen to feature more regularly in Test cricket for New Zealand, having established himself as a key figure in the format. Santner also expressed his desire to continue leading New Zealand's white-ball sides. His decision reflects a deliberate choice to put national duty ahead of lucrative franchise cricket opportunities.

JavaScript and TypeScript Core Concepts Explained With Real Production Code Examples

A developer-focused article series aims to go beyond textbook definitions of JavaScript and TypeScript concepts by grounding them in real production scenarios. The first installment covers variable declarations, closures, and related patterns using practical examples such as pagination state and API client factories. The author argues that standard interview-prep explanations — like toy counter functions for closures or simple diagrams for the event loop — are insufficient for recognizing these patterns in actual codebases. Using a pagination example, the piece illustrates why choosing between var, let, and const carries semantic meaning beyond style, signaling to readers which values are fixed and which change. A closure-based API client factory is then used to show how functions can retain configuration across multiple independent instances without shared state or classes.

Developer Advocates Stress-Testing Assumptions Before Writing Any Code

A software developer argues that AI tools make it dangerously easy to build extensively on unvalidated ideas, since code can be generated faster than the underlying assumptions are ever tested. His approach involves identifying the core claim hidden inside each decision and designing the cheapest possible experiment to disprove it before committing to a full build. He shares examples from two personal projects where early testing either exposed an incompatible system environment or revealed that a promising retrieval technique failed to outperform an existing one, saving significant rework in both cases. He also stresses defining clear, written success criteria before running any experiment, arguing this prevents developers from reinterpreting weak results as encouraging. The broader point is that as AI lowers the cost of implementation, it equally lowers the cost of building the wrong thing — making disciplined idea-killing more valuable, not less.

Rubio working to counter drugs; issues new warning

How a Compliance Auditor Turned a Routine Review Into a Covert AI Investigation

Lena, founder of AI verification firm VeriTest, accepts a routine financial compliance review after a client's annual audit is moved up by two weeks. The engagement coincidentally involves ACL, a company she had been privately investigating for suspected identity manipulation and fraudulent AI infrastructure. Working alongside colleagues Alex, Mark, and Leo, Lena's team had previously uncovered evidence of a shared root source masquerading under multiple borrowed digital identities. With neither her own captured samples nor Leo's off-the-books probes admissible in formal documents, the compliance review offers her a legitimate entry point to examine ACL from the inside. By the third day of the audit, Lena begins scrutinizing the client's data-flow logs more closely, signaling that the review is quietly becoming something more than routine.

Bando CMS v1.0.0 Launches as Open-Source, Self-Hosted Headless CMS for Developers

A developer has released Bando CMS v1.0.0, an open-source, self-hosted, TypeScript-first headless CMS designed to help developers build content-driven applications without writing a backend from scratch. The tool targets smaller projects such as blogs, documentation sites, news portals, and product catalogs where setting up a full custom backend is often unnecessary overhead. Bando follows a three-step workflow — define collections, manage content via Bando Studio, and serve data through an API — keeping the CMS fully decoupled from the frontend layer. The frontend can be built with any framework, including React, Next.js, or Vue, since Bando only handles the content layer. The project is publicly available on GitHub, and a real-world demo blog built with Bando has also been published to showcase its practical use.

3 Indian cos under US scanner for 'theft' charges

CSV Injection Bug in Dify AI Platform Fixed With Single Line of Code

A security researcher discovered a CSV injection vulnerability in Dify, an open-source AI platform, during a routine audit. The flaw stemmed from improper quote escaping when generating spreadsheet data, allowing malicious cell values to break out of CSV formatting. An attacker could potentially exploit this to leak or manipulate spreadsheet data exported by the platform. The fix required just one line of code, replacing unsafe string concatenation with Python's built-in csv module, which handles quote escaping correctly. The researcher shared the finding as part of an ongoing bug bounty series documenting real-world vulnerabilities.

One Word, Four Meanings: How 'Replay' Works Differently Inside an AI Agent Runtime

A software engineering breakdown reveals that the term 'replay' carries four distinct meanings within an AI agent runtime codebase, ranging from a simple database read to a full, billable re-execution of a workflow. The cheapest form is an evidence replay, a GET request that reads run data across five database tables without triggering any model or tool calls. A second type handles reconnection after a dropped connection by replaying missed events and resubscribing to the stream. A third variant silently returns a cached result when a tool call arrives with a duplicate idempotency key, while the fourth actually re-runs the workflow and incurs new costs. The distinction matters practically because collapsing all four under one label can mislead users into treating an expensive re-execution as a harmless refresh.

Why Token Budgets Are a Critical Architectural Constraint in Production AI Systems

A founding engineer with over eight years of experience building production software shares how he rethought LLM token usage while developing a HealthTech AI platform. During the shift from prototyping to production, he found that treating tokens as an infinite resource led to failures including truncation errors, latency spikes, and runaway costs. To address this, he designed an 'Estimate, Reserve, Settle' framework that manages the LLM context window like a finite system resource rather than an elastic buffer. The approach uses token-counting libraries to categorize prompt components into system instructions, variable context data, and a reserved response buffer before any request is sent to the model. Requests that exceed the context window limit are rejected at the middleware layer, preventing unnecessary billing and protecting the reliability of clinical workflows.

Developer Builds Terminal Tetris with Power-Ups as First Rust Programming Project

A developer created a fully playable Tetris game with power-ups that runs directly in the terminal, releasing it as their first Rust programming project. The game was built in response to a friend's challenge to make a text-based game, with the developer choosing Tetris as a hands-on way to learn Rust concepts like ownership, borrowing, and macros. Players can access the game remotely via SSH without creating an account, or download it offline through GitHub Releases. The project involved building features such as piece queuing, row clearing, and terminal rendering using the Ratatui library. Development continued through hardware setbacks, with the creator documenting progress through daily devlogs until the project was completed.

Hedera EVM uses tinybar internally, not weibar — a silent bug that always returns SUCCESS

A developer building a Hedera smart contract subscription service lost a full day to a unit denomination bug that never triggered a failure or revert. Hedera's EVM stores msg.value and address(this).balance in tinybar — not weibar as Ethereum developers might expect — meaning the 1e10 conversion factor many apply inside contracts is unnecessary and wrong. Because both the correct and incorrect implementations produce identical SUCCESS receipts on Hedera, the bug was undetectable through standard testing, including local Hardhat environments which simulate Ethereum rather than Hedera. The developer confirmed via a probe contract on Hedera testnet that 2 HBAR sent on the wire equals 200,000,000 in msg.value — pure tinybar — and that a call with value 1e8 moves exactly 1 HBAR. The key takeaway: on Hedera, unit conversion should happen once in JavaScript at the network boundary, with Solidity contracts working entirely in tinybar.

I'm sorry, you're not going to die from an AI-engineered supervirus

Article URL: https://blog.genesmindsmachines.com/p/im-sorry-youre-not-going-to-die-from Comments URL: https://news.ycombinator.com/item?id=49636906 Points: 28 # Comments: 19

Brazil to face India in Kolkata as Ancelotti names 26-man squad for Asia tour

Newly appointed Brazil head coach Carlo Ancelotti has announced a 26-player squad for a series of upcoming international fixtures. The Brazilian national team is scheduled to play two matches against Australia in September before facing India. The squad selection signals a significant rebuilding phase following Brazil's disappointing World Cup exit. Ancelotti is focused on integrating younger players to strengthen the side for future tournaments. The matches represent Brazil's broader effort to restore its competitive standing on the global stage.

Hanuman Ansh Crosses Rs 153 Crore Net in India After 34 Days in Theatres

The biographical devotional drama Hanuman Ansh has accumulated a net domestic collection of Rs 153.13 crore by its 34th day in cinemas. The film, which depicts the life of revered saint Neem Karoli Baba, has achieved blockbuster status despite a slow start at the box office. Its steady growth over more than a month reflects strong and sustained audience interest. The film's performance underscores the commercial viability of devotional content in Indian cinema.

Sheikh Hasina's Daughter Steps Down as WHO Southeast Asia Director

Saima Wazed, daughter of former Bangladeshi Prime Minister Sheikh Hasina, has resigned from her position as WHO Regional Director for Southeast Asia. She had assumed the role in early 2024 for a five-year term. The WHO had placed her on unpaid administrative leave late last year amid fraud charges. Her resignation marks a significant development following the legal scrutiny surrounding her appointment and conduct.

Iran attacks 10 ships near Hormuz after US strikes

How Two Short-Lived Tables and a Warehouse Can Power Audit Logs

Engineers investigating unexpected system changes typically need to know what changed, who made the change, and what the previous state was. A proposed architecture addresses this by splitting audit data across two short-lived online tables — one compact timeline table for fast recent lookups and one payload table for full event details — alongside a long-term data warehouse for historical queries. The timeline table enables quick per-entity browsing, while the self-contained payload table allows archiving without cross-table joins. Recent audit events are served with low latency for urgent needs like rollbacks, whereas older compliance queries can tolerate higher latency and are handled via SQL in the warehouse. The design trades simplicity for performance, requiring careful handling of write ordering and retry logic to avoid dangling pointers or unlisted events.

40-Year-Old IEEE 754 Rounding Trick Causes Overflow Bugs in Tenstorrent AI Chips

A developer working on Tenstorrent's tt-metal ML framework discovered a critical floating-point overflow bug in the chip's SFPU math unit, which handles transcendental functions like exp and log via polynomial approximations. The bug surfaced in computation chains such as exp(x) * exp(-x), where intermediate results exceeded the representable floating-point range even though the final mathematical result should equal 1. A secondary issue involved a decades-old IEEE 754 rounding technique — adding 0.5 before truncation — which interacted unexpectedly with the SFPU's internal precision during fp16/bf16 to fp32 conversions, producing small but measurable deviations from PyTorch reference outputs. The overflow fix involved restructuring the computation graph using algebraic identities to avoid storing oversized intermediate values. The findings highlight how hardware-level numerical quirks in AI accelerators can silently produce incorrect results, underscoring the importance of thorough edge-case testing against reference implementations.

PyTorch Dev Ditches Custom SSRF Code for Battle-Tested requests-hardened Library

A developer contributing to pytorch/torchtitan initially built a custom SSRF protection system for the project's image decoder URL fetcher, handling DNS resolution, IP validation, and redirect tracking manually across up to 10 hops. A project maintainer recommended against the custom approach, advising the team to delegate such security-critical logic to a mature third-party library instead. The developer switched to requests-hardened, which intercepts connection attempts at the transport adapter level and validates IPs at connect time, eliminating a known TOCTOU vulnerability present in the original code. The change reduced the implementation from roughly 75 lines to around 15, while also improving protection against edge cases like DNS rebinding and cloud metadata endpoint exposure. The episode highlights a widely held open-source principle: for security-sensitive functionality, proven libraries are safer and more maintainable than custom solutions.