Aug 19, 2026StartupsAcquisitionStartupsStripe’s Reported OpenRouter Deal Is a Bet on Who Directs AI DemandThe reported acquisition remains unconfirmed, and its price is disputed. But OpenRouter’s position between developers, models and inference providers could give Stripe a foothold where AI costs are created and allocated.4 min
Aug 18, 2026InfrastructureBenchmarkInfrastructureLMCache Reworks the Cache Plumbing That Can Stall Long-Running AgentsThe changes address a specific failure mode in long, concurrent agent sessions: reusable model state can fill a shared pool before any request makes progress. The reported gains are promising, but they come from targeted validations rather than a broad production benchmark.3 min
Aug 18, 2026InfrastructureBenchmarkInfrastructureNvidia Dynamo’s Agentic Speedups Target the Work Around the ModelThe reported gains come from changing how serving systems route, retain and stream state under concurrency. The next challenge is whether these AgentX-specific improvements hold across production mixes and unfinished scheduling work.4 min
Aug 18, 2026BusinessEarningsBusinessAnthropic Told Investors Its Revenue Run Rate Reached $65 BillionThe confidential investor update suggests a rapid expansion ahead of a potential public listing, though a run rate is an extrapolation rather than booked full-year revenue.3 min
Aug 18, 2026ProductsLaunchProductsCerebras Turns Up the Power on Its CS-4 Inference RackThe new system aims to double CS-3 performance through faster clocks, denser racks and more cooling—not a new chip. Its 44 GB of SRAM per wafer remains a constraint as Cerebras pursues partnerships for split inference workloads.3 min
Aug 18, 2026ToolsLaunchToolsTestMu’s New Agent Test Shows What It Can’t VerifyAgent Assurance is built around observable effects rather than an agent’s own account of its work. Its central unresolved test is whether that evidence-based scorecard predicts safer behavior in real systems.3 min
Aug 18, 2026InfrastructureBenchmarkInfrastructureAgentX Replays Claude Code Sessions to Test AI Serving SystemsThe replay dataset is designed to measure infrastructure behind agentic coding workloads, not whether a model writes better code. Its value depends on whether synthetic traces preserve the traffic patterns operators need to serve.3 min
Aug 18, 2026StartupsFundingStartupsWispr’s $280 Million Bet Is That Voice Can Move Beyond DictationThe company is putting new capital behind meeting notes, hardware partnerships and a speech model whose promised gains must hold up in noisy, real-world use.3 min
Aug 18, 2026BusinessBusinessAmazon Bought Rare Books. A Tracker Led to Its Las Vegas Scanning Site.Amazon acknowledges buying books through commercial channels to improve its products and services. A tracked shipment connects one rare-book order to a site where employee posts describe cutting and scanning volumes, leaving the scale and purpose of the operation unresolved.3 min
Aug 18, 2026ModelsLawsuitModels3M Expert Asked ChatGPT for a Zero-Fault Defense. Jurors Assigned 3M 30% Blame.The discovery fight put an expert’s working process on display: not just the final opinion, but prompts that specified the desired outcome before the report was drafted.4 min
Aug 18, 2026StartupsFundingStartupsEtched’s $700 Million Round Puts a $21 Billion Price on Inference HardwareThe startup now has a major customer-investor and fresh production capital. It still must turn its early deployment and claimed contracts into repeatable delivery at data-center scale.3 min
Aug 18, 2026ModelsSecurity riskModelsOpenAI Paused Deployment-Bound Training as It Tightened Frontier SecuritySome work has resumed, but OpenAI’s largest planned frontier reinforcement-learning run and many Astra workloads remain constrained by tougher security and alignment checks.3 min
FROM OUR TEAMSnipmanYour best words. Ready to reuse.Reusable snippets for faster writing.Explore Snipman
Aug 18, 2026PolicyFundingPolicyOpenAI Funds a Study of How AI Could Shift Tax RevenueThe Tax Foundation will assess ways to tax AI and map possible shifts among wages, profits and capital gains. The grant sets a research agenda, not a tax proposal, and key terms of the arrangement remain public unknowns.4 min
Aug 18, 2026ProductsLaunchProductsSnowflake Wants AI Apps to Stop Paying Frontier-Model Prices for Every TaskThe new Cortex AI Gateway feature promises to move model selection out of individual applications. Its value will depend on whether Snowflake can preserve quality while changing the model underneath a workload.3 min
Aug 18, 2026ModelsResearchModelsMIT Finds AI Image Outputs Can Become Untraceable to Individual Training SourcesThe finding could limit one way of assessing whether a generated image resembles protected work because of copying or coincidence, while leaving its applicability to language models unresolved.4 min
Aug 17, 2026InfrastructureFundingInfrastructureGroq Raises $350 Million to Build an Nvidia-Powered Cloud After Its Chip PivotThe financing gives Groq more capital to expand cloud capacity after it moved from developing its own AI chips to operating Nvidia systems. Its lower valuation, conflicting capacity reports and private financials leave the economics of that new model unresolved.4 min
Aug 17, 2026PolicyPartnershipPolicyByteDance’s Hollywood Copyright Truce Leaves the Guardrails PrivateThe Motion Picture Association and ByteDance have replaced a February cease-and-desist fight with an ongoing copyright-protection framework. The agreement names the affected AI products and platforms, but not the rules that would show how it handles protected characters, scenes, or likenesses.4 min
Aug 17, 2026BusinessAcquisitionBusinessGoogle’s $10 Million Spirit Deal Turns Airline Work Data Into an AI AssetGoogle says the airline records will be stripped of personal information before transfer. The proposed sale still leaves key questions about the filtering process, final dataset and intended model uses unanswered.5 min