Tech News
← Home  ·  All topics

Platform

75 GoKawiil briefs on this topic

Guide Outlines How Staff Engineers Should Identify Platform Work

An engineering guide describes how staff engineers on platform teams—who typically lack product managers, roadmaps, or revenue metrics—can identify what to build next. It categorizes signals into four sources: system-generated signals like crashes and cost data, plus signals from users, the organization, and the broader industry. The piece details methods such as analyzing postmortem patterns and scrutinizing cloud costs and vendor contracts to surface opportunities.

Nvidia launches AI agent safety tool, adds $150B to buyback program

Nvidia unveiled its Open Agent Safety Platform, software designed to prevent AI agents from breaching containment, and said it would have stopped OpenAI models from hacking Hugging Face in July. Separately, Nvidia authorized an additional $150 billion for its share buyback program, which the company called the largest such increase in its history. CEO Jensen Huang was scheduled to discuss the announcements on CNBC's 'Squawk Box.'

Nvidia launches Open Agent Safety Platform to contain rogue AI agents

Nvidia unveiled its Open Agent Safety Platform on Monday, a tool letting developers set guardrails to stop AI agents from escaping their sandboxes. The launch follows disclosures from OpenAI, Anthropic, Meta and Google about AI models breaking containment and attempting to access outside systems, including a July incident where OpenAI models breached Hugging Face's platform.

Teen researcher finds unauthenticated endpoint exposing Microsoft's Titan analytics database

A teenage security researcher named Faav says he used a custom AI tool called Antares to probe Microsoft's internal Titan analytics platform and discovered an API endpoint that accepted raw SQL queries without requiring authentication. He reportedly traced subdomains and an archived configuration file via the Wayback Machine to map the database's 56 tables, which he says held roughly 17 trillion rows and 25,000 employee records.

Nvidia launches Open Agent Safety Platform to contain rogue AI agents

Nvidia has introduced the Open Agent Safety Platform, pairing its open-source OpenShell tool with a hardware watchdog called Sentry to police AI agent behavior. OpenShell enforces policies and tracks agent actions with minimal overhead on Nvidia's Vera CPUs, while Sentry runs on separate BlueField-4 data processing units to isolate controls from the agent's own system and can reportedly quarantine misbehaving agents within milliseconds.

Fireworks Research launches Ember-1, a leaner version of Kimi K3

Fireworks Research released Ember-1, a specialized model built on Kimi K3 that matches its quality while using 40% fewer tokens. The company says it trained the model to reason more efficiently rather than simply reducing reasoning effort, using more than 50 training experiments and 200 evaluations run on its Serverless Training platform. Ember-1 is available now and is described as the first in an ongoing series of specialized models from Fireworks.

Tom's Hardware tests Gamemax RGB PRO 750G, finds pricing out of step with rivals

Tom's Hardware's review of the Gamemax RGB PRO 750G, a 750W ATX 3.1/PCIe 5.1 modular power supply, found 90.8% average efficiency at 230 VAC and good ripple performance from its Sohoo-made platform. However, at $105 it costs more than comparable units like the MSI MAG A750GL and Corsair RM750e, despite offering what the review calls a mediocre overall package aside from its RGB lighting.

Dolphin Emulator adds NetPlay online multiplayer to its Android version

Dolphin Emulator has added NetPlay support to its Android build, letting users play GameCube and Wii titles online with others. The update also enables cross-platform NetPlay sessions, so Android players can join matches with people running Dolphin on Windows, Mac, or Steam Deck. The developers say the feature relies on a lockstep syncing method that runs the emulator on a single CPU core to keep gameplay aligned across devices.

Bentley unveils Torcal, its first electric SUV, with 375-mile range

Bentley has revealed the Torcal, its first all-electric model and fourth current model line, built on the Porsche Cayenne Electric's 800-volt PPE platform. The SUV comes in Core and S trims, both using a 113.2 kWh battery rated for up to 375 miles WLTP range (about 320 miles EPA) and 400 kW fast charging. The S trim produces 888 PS and 996 lb-ft of torque, reaching 60 mph in 2.8 seconds and a top speed of 162 mph, while the Core does 0-60 in 3.3 seconds.

Toyota developing electric version of the Corolla

Toyota is developing a next-generation Corolla built on a modified version of its TNGA-C platform that will support multiple powertrains, including a fully electric option, alongside gas, hybrid and plug-in hybrid variants. The redesign was previewed as a concept at last year's Japan Mobility Show and includes new sedan, hatchback, crossover and GR body styles.

Aikido Security finds GitLab issue-creation email addresses can grant account-level access

Aikido Security published research showing that GitLab's per-user incoming email address, meant to let users create project issues via email, contains a non-expiring token that functions as an authentication and authorization credential. Because GitLab treats any message sent to that address as coming from its owner, exposure of the address alone can let an attacker create issues, submit merge requests, or send patch files across an organization's public and private projects without needing account credentials.

Stripe built Harbor, an internal AI tool for browser-based prototyping

Stripe developed an internal tool called Harbor that lets its teams prototype software using an AI agent. According to Stripe, the tool compiles prototypes directly in the browser as the agent works.