Imagine a scene where a robot sprints faster than the world’s fastest human. This week, that happened in Beijing when a Chinese-made robot clocked 9.39 seconds over 100 meters—beating Usain Bolt's world record of 9.58 from 2009. But don't get too excited; the video shows the machine plowing into a padded wall and needing help to be carried off. It’s not just about the time, but what it means for robotics.

The event took place at the second World Humanoid Robot Games in Beijing, where robots competed in various tasks like running, football, and delicate operations such as picking beans or connecting cables. Tiangong Ultra from Beijing's research center set a new benchmark by finishing 100 meters in under ten seconds, while another robot called Lightning came close at 9.47.

Just last year, the same robot clocked 21.50 seconds over that distance—a huge improvement showing how fast technology is advancing. Experts attribute this to better cloud-based AI connections allowing robots to compute faster as they run. More teams are now involved too—this year saw 666 teams compared to 280 last year, mostly from China.

But while the sprinting achievements are impressive, other tasks showed where robotics still lags behind human skills. Robots struggled with simple tasks like handling beans or opening jars. For such challenges, human assistance is allowed, and full points go only when robots work independently. This reflects how far we are from having truly autonomous machines that can manage everyday life's complexities.

Behind these competitions lies a broader race in robotics. China sees humanoid robots as strategic industries and has poured state funds into the sector. They announced plans to invest around $138 billion over twenty years for AI and robotics. In contrast, while the US focuses on software development, it's becoming wary of foreign-made robots due to cybersecurity concerns.

So why is this important? The gap between what these machines can do now and what they'll need to manage in real-world scenarios is vast. Spectacular feats like sprinting fast are one thing, but dealing with everyday chaos another. It’s a reminder of how much work lies ahead before we see robots as colleagues rather than just cool gadgets.

Robbert
Robbert

Continue Reading

Tools 3 min read

Agoda's AI assistant can now plan and book your entire trip for you in Malaysia

Agoda has switched on its AI Assistant for users in Malaysia, offering end-to-end trip planning that goes beyond simple hotel search. The tool can take a loose request, like a weekend budget trip to a specific city, and generate an itinerary, suggest accommodation and flights, and complete the booking within the same conversational flow. It is aimed at travellers who want to skip juggling multiple tabs and comparison sites, letting the AI assistant handle the back-and-forth of narrowing down dates, budget and preferences. The rollout follows similar AI trip-planning pushes from other travel platforms as the industry races to make booking feel more like chatting with a travel agent than filling out search forms.

Unknown
Tools 3 min read

Meta's AI smart glasses are now on sale in Malaysia, here's which model to pick

Meta's AI-powered smart glasses, built with Ray-Ban and Oakley, are now officially available for purchase in Malaysia. The glasses let wearers take hands-free photos and video, get real-time AI voice assistance, translate speech, and stream audio, all without pulling out a phone. Different frame styles and lens options target different uses, from everyday wear to sports, with pricing varying by model and lens type. For buyers, the key decisions are frame style, battery life expectations, and whether built-in AI features like live translation or visual search are supported for Malaysian language settings. Check official Meta or authorised retailer listings for exact Malaysian pricing and warranty terms before buying, since grey-market imports may lack local support.

Unknown
Tools 3 min read

Microsoft starts rolling out Copilot Code so you can build apps just by describing them

Microsoft has begun rolling out Copilot Code, a tool that lets users describe an app in plain language and have it generated and assembled automatically, without writing code line by line. It builds on Microsoft's existing Copilot family but is aimed squarely at turning natural-language prompts into working software, from internal business tools to simple apps. The rollout is gradual, so availability may vary by region and subscription tier. For non-developers, this lowers the bar to automating small workflows or prototyping ideas without hiring a developer; for professional coders, it's another layer of AI-assisted scaffolding on top of existing Copilot features. Try it by checking your Microsoft 365 or Copilot subscription for the new build option, and start with a small, well-defined task to see how it handles your instructions.

Unknown