Apple: Tim Cook, CEO of Apple for almost 15 years, is stepping down. The current senior vice president of Hardware Engineering, John Ternus, will be replacing him on Sept. 1, 2026. Under Cook, Apple became the first trillion-dollar company, launched several new products (the Apple Watch, HomePod, AirPods), started several climate initiatives and continued to innovate in digital accessibility. Cook will transition to Apple’s board of directors as executive chairman.
Sprinting Robot: In China, a sprinting robot named Lightning recently beat the human record for the fastest half marathon. Racing in an annual human-robot race, Lightning completed the course in 50:26, 6 minutes and 54 seconds faster than the record set by Jacob Kiplimo last month. Approximately 5’5” tall, Lightning is autonomous and uses liquid cooling to ensure optimal performance. Around 60% of the robots in the race were remotely operated.
Framework: Laptop company Framework has announced the latest in its laptop lineup: the Framework Laptop 13 Pro. This new computer is configurable with the Intel Core Ultra Series 3 chips, a touchscreen, a haptic touchpad, and speakers that support Dolby Atmos. Framework is a company known for creating laptops with parts that can easily be swapped out in ‘modules’, allowing users to repair and upgrade their own computers.
Meta: United States-based employees at Meta will soon be training AI while completing daily tasks. The company has said that it will be using a tool known as Model Capability Initiative (MCI) to track employees’ digital actions across apps. With this additional data, Meta hopes to improve its models’ capabilities in navigating a computer. CTO Andrew Bosworth has claimed that the company is working towards a future where “agents primarily do the work and our role is to direct, review and help them improve.” The company has no plans to integrate the collected data into performance reviews.
Google: At Google Cloud Next 2026, Google announced the eighth generation of their Tensor Processing Unit (TPU): TPU8t and TPU8i. The 8t is intended for model training, whereas the 8i is designed for inference (what happens after a user sends a prompt). Compared to its previous generation of chips,offers up to 80% better performance per dollar.
