NVIDIA is developing a new AI chip for China. But there's a catch.
Plus, more links to make you a little bit smarter today.
We’ve made a course about using AI to develop your own RAG-based chatbot. Impress your clients/future employers and develop a potential source of income by following along with our lessons!
From the vlog…
From the main channel…
Finally Finishing the Heuristics Arc… with a free bundle.
Short article this week, but for a good cause: I’ve made a free flashcard set based on my last 3 programming articles!
NVIDIA is developing a new AI chip for China. But there's a catch.
This new chip is actually *half* as powerful as their current top-of-the-line Blackwell GPU. The reason for this is new regulations that prevent US-based companies to sell AI equipment to "enemies of the state", i.e. China.
This rule isn't new. Similar regulations were put into place on aerospace companies during the height of the Cold War. Of course, it barely mattered -- Soviets were usually able to smuggle in the new technology anyway, and anything they couldn't steal they just researched themselves.
Will history repeat itself?
Read the article: https://techcrunch.com/2025/08/19/nvidia-said-to-be-developing-new-more-powerful-ai-chip-for-sale-in-china/
Self-driving has found its way to Japan.
Tesla, still in a heated rivalry with other self-driving projects such as Waymo, have expanded beyond the US and China and have begun "playtesting" driver-assisted self-driving for their cars in the Japanese market.
It will be interesting to see how these cars will adapt to different traffic laws and street systems. Technically speaking, "niches" could form where certain self-driving software that does particularly good in a given traffic system could end up dominating in that market just based on that. Waymo owns US, Tesla owns Japan, etc. etc...
Read the article: https://asia.nikkei.com/business/automobiles/tesla-tests-ai-powered-autonomous-cars-on-public-roads-in-japan
First, AI companies took on images. Then video. Now, games are next.
In a recent interview on Bloomberg Tech, Cristóbal Valenzuela talked about how Runway -- best known for its foray into AI video -- was targeting gaming next. Runway's current Game World concept is rather simple, mostly using AI to enhance text adventures which have existed for decades. But earlier in the month, we had Google DeepMind's Genie 3 demo, which resembles something much closer to what gamers -- and companies -- want to see.
It'll be a race to see who can get there first.
Watch the video: https://www.bloomberg.com/news/videos/2025-08-21/ai-startup-runway-s-play-for-video-gaming-video
The Pentagon is looking to space to revolutionize how it transports cargo, and two major players just landed key contracts to make it happen.
Blue Origin and Anduril Industries have secured new study contracts with the U.S. Air Force under its Rocket Cargo program, specifically the Rocket Experimentation for Global Agile Logistics (REGAL) initiative. These contracts, totaling over $2.3 million, aim to explore how their rocket technology can move military cargo and potentially even humans around the globe in a "delivery as a service" model, with the goal of delivering to remote locations in less than an hour.
While these initial contracts are for studies, they are strong signals for future larger funding and the long-term vision of global, rapid logistics. Anduril's specific contract to design a 5-10 ton payload reentry container highlights the notoriously difficult challenge of safe atmospheric reentry, a critical component for enabling this 'delivery as a service' vision. This pioneering work could fundamentally transform military supply chains, making rapid deliveries to remote theaters possible and demonstrating a significant stride in space-based logistics.
Read the article: https://techcrunch.com/2025/08/21/anduril-blue-origin-to-study-how-to-transport-cargo-from-orbit-to-earth-for-the-pentagon/
AWS CEO Matt Garman calls the idea of using AI to replace junior staff "the dumbest thing I've ever heard."
In a recent discussion with AI investor Matthew Berman, Garman, CEO of Amazon Web Services (AWS), criticized business leaders who believe AI tools can replace junior employees. He argues that these junior staff are often the least expensive employees and the most engaged with AI tools. Replacing them now would lead to a significant talent gap in the future, as there would be no one learning and growing into senior roles. Garman advocates for continuing to hire college graduates and teaching them how to build software and decompose problems, suggesting AI — like AWS’s Kiro tool — can assist in this education. He also questions the value of measuring AI by the percentage of code it contributes, stating that fewer, higher-quality lines of code are often better than more, potentially bad, code. Data suggests over 80 percent of AWS’s developers use AI for various tasks, including writing unit tests, documentation, and code.
This perspective offers a crucial counter-narrative to the common fear that AI will simply eliminate entry-level positions. It reframes AI as a tool for enhancing human capabilities and fostering the next generation of talent, rather than solely a cost-cutting measure. Garman's emphasis on developing critical reasoning, creativity, and a continuous "learning mindset" underscores the adaptability required in the rapidly evolving AI age, suggesting a path for career longevity beyond narrow skill sets. This aligns with the broader theme of "reskilling" for the AI era.
Link to the article: https://www.theregister.com/2025/08/21/aws_ceo_entry_level_jobs_opinion/