Connecting RS‑232 gear to a Metis M.2 #AI board? This community post breaks down the pin‑out, wiring, and tech tweaks needed to get legacy devices talking to tomorrow’s edge AI platforms! A must‑read for embedded & industrial innovators, and those with a nostalgic pang for
HARDWARE
-
8th and 9th Neuralink Devices Implanted This Weekend
By
–
Les 8e et 9e Neuralinks ont été implantés ce week-end.
-
xAI Reaches 500K GPUs, Targets 1M GPU Cluster This Year
By
–
Already half way to the first 1,000,000 GPU cluster. For context, @xai started wiring together Colussus in 2024 putting together 200,000 GPUs in 122 days. Seemingly impossible, it was an incredible feat of engineering. They will get to a million within the year.
-
xAI Colossus 2 Data Center: 550K NVIDIA GPUs in Tennessee
By
–
xAI gave a glimpse of its Colossus 2 data center in Tennessee The facility will be powered by 550K NVIDIA GB200 and GB300 GPUs, using various energy sources Musk said the goal is to bring AI compute equivalent of 50M units of Nvidia’s H100 online in 5 years
-
OpenAI invests $30B annually in Oracle Stargate data center
By
–
Open is set to pay Oracle $30B/year for a massive 4.5GW Stargate data center The move takes total Stargate capacity under development to 5+GW CEO Sam Altman says the company will invest way more than the originally committed $500B as part of the project
-
5G Innovation Lab Transforms MLB All-Star Game and Healthcare
By
–
From fast pitch calls to automated strike zones, @TMobileBusiness this is rapidly evolving the #MLB All-Star Game into a live 5G innovation lab! And I love how the impact goes far beyond baseball: Real-time data in smart hospitals Precision control in Industry 4.0
-
Apple Silicon GPU Memory Access Enables Better LLM Performance
By
–
I don't think that will run an LLM very well, you need GPU-accessible memory to get good performance That's what's neat about Apple Silicon – regular RAM is available to the GPU
-
Qwen3-Coder-480B runs locally on high-end Mac Studio
By
–
Looks like Qwen3-Coder-480B-A35B-Instruct is a VERY strong coding model to run locally… if you can afford a ~$10,000 512GB Mac Studio to run it on!
-
Amazon Acquires Bee: Building AI Hardware Strategy
By
–
congrats to Bee on getting picked up by Amazon; similar to Blink, Ring, Eero and ofc the new Claude + Nova + Alexa i think @panos_panay is putting together a pretty solid second act of amazon’s ai hardware strategy knew bee was winning when @dharmesh showed up to his
-
ASML and Taiwan Semiconductor Navigate Tariff Uncertainties
By
–
The TL Podcast for July 20th: Praying for growth As stocks continue to hit new all-time highs, chip-equipment maker ASML let us know that the risk of continued uncertainty from tariffs and trade is by no means over, while Taiwan Semiconductor continues to thrive as a result of