CoreWeave launched Forge, a platform meant to speed up an ongoing agent improvement loop by letting teams share a common workflow for running, observing, curating, improving, and evaluating AI agents. The article says the platform measures success by speed of learning, tied to faster shipping and faster value delivery. It adds capabilities like Agent Lens, model checkpoints/configuration via Registry, RL Rollouts and model distillation, and makes its ARIA tool generally available, alongside a partner network of co-engineered integrations.
CoreWeave announced CoreWeave Forge to connect training, inference, and evaluation in an open, full-stack AI cloud. The platform is tied to its newly announced development platform and CoreWeave recently earned SemiAnalysis’ Platinum ClusterMAX rating for a third consecutive time. As a result, production learnings can feed back into improving models and agents, and customers can start with specific inference use cases on the same platform.
Arena, an AI model evaluation and leaderboard platform, raised a $200 million Series B at a $3.1 billion valuation. It also said it reached $100 million in annualized run-rate revenue in June. The funding and new “alignment” leaderboard category shift Arena’s role toward providing more detailed, community-based evaluation for model labs and enterprises.
Hume AI CEO Andrew Ettinger explains that voice AI can sound human while still failing to capture the meaning-carrying signals in speech. He says transcripts can strip away multiple parts of communication such as tone and background noise, leaving systems with a “listening problem.” As a result, voice AI development focuses more on evaluation in real-world conditions, better modeling of emotion as a spectrum, and addressing new safety and personalization risks when agents talk back.
Goodfire launched “inside-out” monitors that detect misuse by reading internal signals inside an AI model instead of having a separate model reread the agent’s output. In tests on Kimi K3, monitoring about 1,500 sessions cost roughly $51 versus $233 for a cheaper step-by-step checker and about $10,000 for a top-tier one. This lowers monitoring cost and overhead, enabling Baseten customers to choose which risks to watch and whether to log, escalate to human review, or refuse requests.
Webflow argues that incident tabletop exercises no longer test the right variable because AI now performs parts of triage, playbook selection, and response drafting. The article says to build new injects that cover AI failure modes such as wrong or incomplete answers with full confidence. As a result, teams should add AI-specific failure scenarios, include the people who maintain the AI tooling, and debrief on whether responders trusted the AI appropriately, not just whether they followed the runbook.
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.
