Granite Switch 4.1 combines decision and speech adapters into one hosted model endpoint so it can handle chat, decision requests, and speech-to-text tasks. Granite-Speech-5.0-Turbo-CTC is claimed to be the fastest ASR model on the OpenASR leaderboard. The release reuses the base model’s KV cache across adapters and aims to deliver more capabilities without additional GPU memory beyond what the base model alone would use.
Vercel added Daemons for continuously running backend workers and made its Python support and Drives persistent storage for AI agents generally available. Daemons is in public beta and supports running a specified number of instances per region. These updates move more asynchronous workloads onto Vercel, expanding what developers can run on its infrastructure without maintaining separate worker and storage systems.
openKylin used Open Source Summit Europe 2026 to outline how Linux should change in the AI era by introducing KART, a unified AI runtime, plus a system-level agent execution environment. It said RVV kernel optimizations provide a 2–5× speedup in FP32 inference with ONNX Runtime. openKylin 3.0 is presented as shifting desktop interaction to intent-based control, updating search and organization around semantics and tasks, and extending support toward the RISC-V and open hardware ecosystem.
Modal moved from infra unicorn to infra decacorn after reporting a $16B Series D alongside $1B ARR and additional Sandboxes usage growth. The announcement cites close to 3 billion sandboxes started, scaling to a million concurrent. As a result, Modal’s valuation and usage rose about 3x within 5 months and it reached a claimed 15x increase over 1 year.
Harness by Autonomous launched an open-source command center for coding agents that runs on the desktop app, terminal, and iOS/Android, with voice control and end-to-end encryption. It includes 49 harnesses for CAD, circuits, robots, games, and music. Free on Mac and Linux, the release adds a centralized way to coordinate multiple agents on different machines.
Every AI story that matters,
in your inbox by 8am.
TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the
day in two minutes. Follow companies and topics for alerts, or get the
briefing in Slack. Free, no spam, unsubscribe anytime.
Reading TLDRocket needs no cookies, and the readership counts we rely on come from
our own cookieless analytics. Google Analytics is the exception: it sets cookies and
reports to Google, so it stays switched off until you allow it. You can change your
mind any time from “Cookie settings” in the footer.
Strictly necessary
Session security and form protection (tldrocket-session,
XSRF-TOKEN, 2 hours). The site cannot work without them,
so they need no consent.
Always on
Google Analytics 4 (_ga,
_ga_<id>, up to 2 years). Measures which
stories and sections readers use. Google acts as a third-party processor and may
store the data outside the EU. No advertising, no profiling, no data sold.
