InfraNVIDIA showed Vera Rubin: AI hardware that keeps learning nonstopNVIDIA announced Vera Rubin - a platform for 'continuous post-training' of agentic models, meaning systems that don't stop learning from their own actions after the initial training. The company claims it does the same job with a quarter of the chips compared to the previous generation. The numbers come from NVIDIA. The direction is real though, and everyone paying the AI bill pays for it.InfraNVIDIA brings its robot chips down to the mass marketOn 15 July NVIDIA launched two new Jetson Thor modules - T3000 and T2000 - on the Blackwell architecture, for robotics and edge AI. The cheaper one lands exactly where earlier robot boards were too expensive for a device made in volume.InfraCloudflare rolled out a code that shows when DNSSEC has been bypassedOn 3 July the Albanian domain .AL stopped opening for resolvers with DNSSEC checking on. The operator AKEP had changed the key. But the record in the root zone still pointed to the old one. Cloudflare's public resolver 1.1.1.1 restored access with a workaround, but for the first time marked the answer with a code that openly says: this check was skipped.InfraGoogle released a sandbox for agent-written code, with the network closed by defaultCloud Run sandboxes entered public preview on July 10. You run code written by an AI agent, and it doesn't touch the rest of your system. What matters is in the default: zero outbound network access, zero visibility into the service's keys.InfraCloudflare stopped guessing where your server is and started askingAs of 10 July you can specify yourself which AWS, Google Cloud, Azure or Oracle region your server sits in, instead of the network guessing by response time and getting it wrong. A shorter path for traffic, no extra cost, even on the free plan.InfraAMD reports up to 6.3x faster image generation on its MI350XOn 10 July AMD published benchmarks on its blog for generative visual models on the accelerator MI350X. The numbers are impressive, but the baseline is an unoptimized reference implementation - not a previous-generation chip, not NVIDIA. The tests are AMD's own, with no independent verification.InfraNVIDIA produced half a million made-up financial headlines to train other modelsNVIDIA showed how to manufacture financial news that never happened - to feed models where real data is expensive and locked behind licenses. The result looks like news. Behind it there isn't a single real event.InfraCloudflare showed a consensus that holds where the classic one breaks330 data centers around the world, agreeing on the same thing at the same time - Cloudflare Research showed how: the Meerkat service, built on the QuePaxa algorithm, with about 10 times higher throughput than Raft under a bad network, per the company.
48 stories