Faster streaming speech recognition on @Arm edge devices: see how ExecuTorch, SME2, FP16 and QC8W balance accuracy and latency—delivering up to 2.12× faster processing on Pixel 11. Read our latest blog: ✍ Josiah Davis, Kshitij Sisodia

Faster than fueling up! A senior executive from an overseas state-owned power grid enterprise experienced #BYD’s# flash charging technology on-site, getting a firsthand look at China’s hard-tech innovation in new energy and coming away deeply impressed by China’s remarkable flash-charging speed! #ChinaTech#

Faster settlement improves tokenized products, but distribution is the near-term growth lever. Tokenization lets institutions reach stablecoin holders who wouldn't otherwise invest in products that only exist off-chain, says @benafisch.

Faster checkpoints, less GPU idle time. Our collaboration with Google Cloud brings GCSFS and Rapid buckets to PyTorch Lightning, cutting checkpoint write times by up to 95%. Full breakdown and how to try it →

Faster GLM-5.3-Flash is now live: up to 200 tokens/s. Model code: glm-5.3-flashx. Priced at 2.5× GLM-5.3-Flash on both the Coding Plan and API. Open to all API users. Coding Plan users can apply here:

Faster slot times are coming to Ethereum. Every time you send ETH, swap an asset on a DEX, or deposit funds to an L2, there's a gap between pressing the button and the transaction being processed onchain. Today that wait time is 12 seconds. The upgrade shaves 2 seconds off the top. Sounds small, doesn't it? Two seconds by itself isn't really a revolution. You might not even notice it at first. But those seconds don't operate in isolation. They sit right at the top of a complicated stack. Ethereum has been locked at 12 seconds for years. Every other part of the stack (gas limits, L2 proofs, confirmations) kept improving while the basic clock stayed frozen. Once the software no longer assumes "slots are ALWAYS 12 seconds", later upgrades can shave off more of it without adding more technical debt. That's engineering jargon for teams chasing easy software shortcuts instead of a clean, sustainable solution. The small cut adds up across the network: > You get 6 chances to land a transaction instead of 5. > Onchain prices refresh 20% more often, reducing transaction slippage. > Full block finality drops to 10.7 minutes from 12.8. > Bridges and L2s settle transactions slightly faster than before. The upgrade has no go-live date yet and is tentatively scheduled for the upcoming Hegotá fork, with mainnet expected in Q2 2027.