Hackers hit AI coding tool OpenCode, running up charges on users’ AI accounts

techcrunch.com Oct 19, 2026

Hackers hit AI coding tool OpenCode, running up charges on users’ AI accounts

Curated by Shubham · TechCapsules · Oct 19, 2026

AI-summarised brief · reviewed before publication

OpenCode, an open‑source AI coding agent, disclosed a security breach that overnight compromised an older system, affecting 1.4% of users on its OpenCode Console platform where API keys are stored. The attack, described as a large‑scale distillation attempt, led to unauthorized charges on users’ AI accounts, with some exceeding $1,000. OpenCode notified impacted users, blocked exposed keys, and is processing refunds. The incident highlighted challenges for companies lacking vetted, low‑restriction AI models to defend against AI‑driven cyberattacks.

💡 Why It Matters

· The breach exposes a growing vulnerability in AI‑centric development tools, forcing developers to confront the limits of current safety guardrails and the need for more robust, vetted AI models to secure sensitive credentials.

AI Capsules Cybersecurity

Read full article on techcrunch.com

Replika’s founder built the AI assistant she pitched at Disrupt 12 years…

Twelve years ago, Eugenia Kuyda introduced a conversational chatbot called IO at TechCrunch Disrupt, a precursor to today’s large‑language‑model assistants. The startup Luka later pivoted into Replika, a popular AI companion. After stepping down as Replika CEO, Kuyda founded Wabi, which launched a multiplayer personal AI assistant that organizes tasks into distinct threads and auto‑builds custom apps. Wabi 2.0 lets the agent create and manage these threads, aiming to reduce context drift and enable shared group chats.

💡 Wabi’s multithreaded design tackles a key limitation of current AI assistants—context loss over long conversations—by keeping tasks compartmentalized and shareable. This approach could set a new standard for personal AI, blending individual productivity with collaborative interaction.

9to5mac.com · Oct 19

Apple strikes deal with AI startup that lets musicians rework recorded music

Apple announced it will acquire assets of Swiss AI audio startup Samplab, adding its music‑editing technology to the company’s portfolio. The deal, reported on the EU DMA website, includes hiring selected Samplab staff. Samplab’s platform lets musicians edit audio like MIDI, detect and adjust chords, match tempo and key, and separate stems for granular control. The startup’s website now redirects to Apple, and its social media has been cleared of prior content. Apple has not yet disclosed specific integration plans.

💡 By securing Samplab’s AI‑driven editing tools, Apple positions itself to enhance its creative software suite, potentially redefining how musicians and producers manipulate audio on iOS and macOS platforms.

techcrunch.com · Oct 19

Tolan snags $30M for its AI companion that helps users navigate life

Tolan, an AI companion app that assists users with daily tasks, relationships, and emotions, has secured an additional $30 million in funding from Union Square Ventures, Khosla Ventures, and Homebrew, doubling its total capital raised. The investment pushes its valuation past $120 million, with PitchBook estimating it above $268 million. Tolan now reports double‑digit million annual recurring revenue and boasts over 250,000 active users, 70 % women and 40 % parents, engaging in more than 70 million conversations. The company’s cartoonish avatar differentiates it from other AI companions that faced backlash.

💡 The influx of capital positions Tolan to scale its emotionally supportive AI, potentially reshaping how consumers seek digital mental health and productivity tools.

ign.com · Oct 19

Steam Fans Are Begging for a No-AI Filter as Steam Next Fest…

Valve’s Steam Next Fest opened with an unusually high volume of AI‑generated indie games, prompting backlash from users who feel the event is cluttered with low‑effort titles. Fans on the Steam subreddit demanded a filter to exclude AI‑generated content and other tags, citing the platform’s existing disclosure requirements. Valve’s current tagging system would need an “AI” tag to support such filtering, but the company has previously reduced user‑defined tags. The influx of AI‑driven demos reflects a broader industry shift toward AI‑first development.

💡 The surge of AI‑generated games threatens to dilute the quality of indie showcases, pressuring platforms like Steam to rethink curation standards and transparency.

therobotreport.com · Oct 19

Daimon Robotics building the infrastructure layer for tactile intelligence

Daimon Robotics is constructing an end‑to‑end infrastructure for tactile intelligence, linking high‑frequency tactile sensing, large‑scale data collection, calibrated simulation, and AI models. The company’s pipeline captures physical interaction data at 30 Hz, aims to amass over 10,000 hours of tactile recordings—targeting a million‑hour “Daimon‑Infinity” dataset—and uses these measurements to close the sim‑to‑real gap by calibrating virtual environments against real‑world forces, friction and slip. Rather than producing a single foundation model, Daimon positions itself as a provider of tools, sensors, and datasets that downstream robotics firms can leverage for tasks such as assembly, packaging, and delicate manipulation. Demonstrations at IROS 2026 showcased applications ranging from heat‑transfer tote‑bag handling to broken‑glass packaging, and partnerships with Lenovo Capital and Inovance Technology support its move from proof‑of‑concept to industrial lines.

💡 By supplying the data and simulation backbone for tactile perception, Daimon enables robots to reliably handle nuanced physical tasks that vision alone cannot manage, accelerating the deployment of truly dexterous automation in manufacturing and logistics.

therobotreport.com · Oct 19

Unleash AI Innovation: The Power of NVIDIA RTX PRO 6000 Blackwell Workstation…

PNY Technologies announced the availability of NVIDIA RTX PRO 6000 Blackwell Workstation Editions, including a Max‑Q variant that can house up to four GPUs delivering 14.2 petaflops of AI compute and 384 GB of GPU memory. The systems target AI training, inference, data‑science preprocessing, and physical‑AI workloads such as robotics simulation, offering local processing power comparable to data‑center resources. Integrated with NVIDIA’s software stack—CUDA‑X, RAPIDS, NGC, AI Workbench, and Blueprints—the workstations enable rapid prototyping, fine‑tuning of generative models, and seamless scaling to cloud or on‑premise clusters. Enterprise users can embed the machines into existing IT environments, supporting high‑resolution creative tools like Adobe Photoshop Neural Filters and DaVinci Resolve face tracking while reducing cloud latency and costs.

💡 By delivering data‑center‑grade AI performance on a desktop, the RTX PRO 6000 Blackwell workstations let organizations iterate locally, cutting reliance on expensive cloud compute and accelerating time‑to‑innovation.