Latest in AI
Short Drama Ads Arrive: The Trade Desk Partners with DramaBox
INSIDE 硬塞 AI4d agoBusinessThe Trade Desk sees short drama advertising as a data-backed opportunity, citing forecasts that the global short drama app market outside China could reach $3 billion in 2025. Its partnership with DramaBox brings vertical short drama inventory into programmatic advertising. The goal is to capture fragmented attention, expand the attention spectrum, and help brands treat emerging content traffic as measurable media within omnichannel strategies.
Waymo Is Too Costly, Uber Too Slow: Fatal Flaws in Two Robotaxi Models
INSIDE 硬塞 AI4d agoBusinessThe article contrasts two robotaxi commercialization strategies. Waymo controls technology and distribution through vertical integration, gaining tighter control but facing high costs. Uber relies on partnerships and its ride-hailing platform, keeping a lighter model but risking slower execution and less control. The broader question is whether value in autonomous mobility will accrue to core technology owners or demand-distribution platforms.
Snowflake and AWS Expand AI Infrastructure Bet to $6B Over Five Years★ 72
INSIDE 硬塞 AI4d agoBusinessSnowflake reported stronger-than-expected results and raised its annual product revenue forecast as enterprise demand grows. The company signed a five-year, $6 billion AI infrastructure agreement with AWS, expanding a previously smaller commitment. It also acquired Natoma to strengthen AI agent governance, positioning itself as a core enterprise AI platform.
Function invocations now billed per unit
Vercel Changelog4d agoBusinessVercel announced a billing change titled “Function invocations now billed per unit.” Without the full changelog text, the confirmed takeaway is limited to the billing basis for function invocations. Teams using Vercel Functions should review invocation-heavy APIs, background jobs, webhooks, polling, and AI workflows, but should not assume exact pricing or plan impact without checking the official billing details.
Adobe’s conversational AI agent is a mediocre design intern
The Verge AI4d agoCommentaryThe Verge tested Adobe’s Firefly AI Assistant beta, a conversational tool that can perform multi-step image edits using Adobe-style capabilities. It explains its process, asks follow-up questions, and is open about limitations, making it more instructive than many creative chatbots. But the actual edits are often imperfect, with weak blending and middling generative results, so it feels more useful for casual users than professionals.
Is Your Data in Taiwan? Cross-Border Transfer Risks in Asia-Pacific Data Deployment
INSIDE 硬塞 AI4d agoCommentaryUsing the Grab acquisition debate as context, the article says offshore data storage is now normal for digital services. The real issue is not whether data stays in Taiwan, but whether the storage jurisdiction has strong legal protections, oversight, and remedies. Singapore is presented as a case worth examining for Asia-Pacific data deployment and cross-border transfer risk assessment.
From Map Data to National Security: A Deep Dive into Amap Risks in Taiwan★ 74
INSIDE 硬塞 AI4d agoRegulationINSIDE examines how China’s Amap has become controversial in Taiwan beyond ordinary mapping or navigation use. The article says its service relies on user data and AI-based inference rather than full official data integrations. That model could send movement traces and behavioral signals back to China, creating risks for hybrid warfare intelligence, influence operations, and Taiwan’s broader governance of map data and digital infrastructure.
Anthropic Releases Claude Opus 4.8 With Integrity Upgrades and Dynamic Workflows
INSIDE 硬塞 AI4d agoReleaseAnthropic released Claude Opus 4.8 as a rapid iteration focused on stronger integrity and reliability for high-risk tasks. The company also previewed Dynamic Workflows, a feature designed to coordinate multiple agents on large-scale jobs such as code migration. The article mentions Mythos entering a countdown toward unblocking, but does not provide detailed availability or product specifics.
Jensen Huang Hosts Taiwan Supply Chain Dinner, Says TSMC Leads Huawei by 10 Years
INSIDE 硬塞 AI4d agoBusinessNVIDIA CEO Jensen Huang hosted key Taiwanese supply chain partners, with senior leaders from TSMC, Foxconn, and Quanta attending the high-profile dinner. The report frames the event as a signal of Taiwan’s central role in AI hardware, from advanced chips to manufacturing and servers. Huang also said TSMC leads Huawei by 10 years, underscoring the strategic weight of semiconductor capability.
Blue Origin New Glenn Pad Explosion Hits Satellite and Lunar Plans
INSIDE 硬塞 AI4d agoIncidentBlue Origin’s New Glenn rocket exploded during a static fire test in Florida, putting attention on launch pad damage and the investigation outcome. The incident may delay Amazon satellite deployment plans, NASA Artemis-related work, and national security launch certification. No cause or recovery timeline is confirmed in the provided source, so future schedules depend on repairs, findings, and approval to resume testing.
Chinese Sodium-Ion Battery Passes German Test, Matches Tesla Lithium Performance
INSIDE 硬塞 AI4d agoHardwareA German independent study has reportedly completed the first full third-party evaluation of China’s Hina sodium-ion battery. The test found strong cell uniformity and multiple performance metrics comparable to advanced lithium batteries, with the report benchmarking it against Tesla-level lithium performance. The key takeaway is external verification: the findings provide checkable data for assessing China’s sodium-ion battery progress.
Siri Redesign and Customizable Camera? Biggest iOS 27 Leak Ahead of WWDC 2026
INSIDE 硬塞 AI4d agoReleaseINSIDE reports that a major iOS 27 leak points to a redesigned Siri experience, potentially arriving as a standalone app rather than only a system voice assistant. The new Siri is said to integrate deeply with Dynamic Island, suggesting a more visible and persistent interaction layer. The headline also mentions camera customization, but the available text does not provide enough detail to confirm how that feature would work.
Protecting against inference theft
Vercel Changelog4d agoCommentaryOnly the title is available, so specific Vercel product changes or implementation steps cannot be confirmed. The topic appears to focus on protecting AI inference resources from unauthorized access, abuse, or cost-draining traffic. For teams deploying AI apps, the practical takeaway is to treat inference endpoints as high-value backend assets requiring access control, monitoring, and abuse prevention.
Protecting against token theft
Vercel Changelog4d agoTutorialVercel published a post titled “Protecting against token theft,” focused on token security risks and protection. The article body was not provided, so its scope, affected products, attack scenarios, and recommended mitigations cannot be confirmed. Readers should consult the original Vercel page before taking action or attributing specific guidance to the company.
datasette 1.0a31
Simon Willison's Weblog4d agoReleaseSimon Willison released Datasette 1.0a31, a significant alpha release with two headline features: write SQL execution and stored queries. Users with the right permissions can now run database-changing queries and save queries privately or for other members of a Datasette instance. The new interface can generate templated insert, update, and delete queries for editable tables while blocking unauthorized actions such as creating tables without permission.
Anthropic Series H Valuation Reaches $965B, Backed by Memory Giants★ 78
INSIDE 硬塞 AI4d agoBusinessAnthropic completed a $65 billion Series H round, bringing its valuation to $965 billion and reportedly surpassing OpenAI. The round included strategic investments from memory makers Micron, Samsung, and SK Hynix. The news highlights how frontier AI companies are increasingly tied to hardware and memory supply chains, as investors continue backing foundational model competition.
AI Power Demand Surges as TSMC Says Energy Efficiency Is Key to Chip Development
INSIDE 硬塞 AI4d agoHardwareTSMC senior vice president Cliff Hou said customers across smartphones and AI data centers are increasingly focused on improving performance without increasing power use. The comment reflects rising energy pressure as AI workloads expand. For chipmakers and infrastructure buyers, energy efficiency is becoming a central metric alongside raw computing performance.
SYSTEX Uses AWS to Build Enterprise AI Platform EAP
INSIDE 硬塞 AI4d agoBusinessINSIDE reports that SYSTEX is pushing forward with SaaS and enterprise AI despite debate sparked by Claude Code and claims that “SaaS is dead.” The Taiwanese IT services leader reported strong Q1 2026 earnings, with net profit after tax of NT$718 million, up 164.5% year over year. It also introduced EAP, an Enterprise AI Platform built on Amazon Web Services cloud-native architecture to support enterprise AI adoption.
SYSTEX Builds EAP on AWS to Bring Enterprise AI Into Production
INSIDE 硬塞 AI4d agoBusinessINSIDE reports that SYSTEX is positioning its Enterprise AI Platform as a cloud-native route for enterprise generative AI adoption. The article contrasts this with recent “SaaS is dead” discussions sparked by tools such as Claude Code. SYSTEX also reported strong Q1 2026 earnings, with after-tax profit of NT$718 million, up 164.5% year over year.
Anthropic raises $965B Series H, releases Opus 4.8 and Dynamic Workflows/ultracode
Latent Space4d agoBusinessThe visible AINews item centers on Anthropic, claiming a $965B Series H alongside Opus 4.8 and Dynamic Workflows/ultracode releases. The available body text is extremely brief, offering only the editorial line “Total Anthropic victory!” It signals a major Anthropic narrative across capital, Claude models, and developer workflows, but provides no detailed specs, benchmarks, investor terms, or availability information.
Anthropic's run-rate revenue hits $47 billion★ 76
Simon Willison's Weblog4d agoBusinessSimon Willison highlights Anthropic’s latest Series H announcement, where the company says run-rate revenue crossed $47 billion earlier in May. He traces prior disclosures: about $9 billion at the end of 2025, $14 billion in February 2026, and over $30 billion in April. The post also addresses skepticism, arguing that these numbers appeared in fundraising announcements, where knowingly misleading investors would be securities fraud.
Glean’s top line crosses $300M as AI budget-cutting becomes its major selling point
TechCrunch AI4d agoBusinessTechCrunch reports that enterprise AI search startup Glean has crossed $300 million in annual revenue. The company tripled its annual revenue even as major tech companies entered the same category. Its pitch is increasingly centered on helping enterprises reduce or rationalize AI budgets, not only on AI-powered workplace search.
Port 8080 is now available in Vercel Sandboxes
Vercel Changelog4d agoReleaseVercel announced that port 8080 is now available in Vercel Sandboxes. Based on the provided source, this appears to be a small developer-experience update around sandbox port availability. It may reduce configuration friction for projects, dev servers, or tools that commonly default to port 8080, though no further implementation details were provided.
Run Docker containers inside Vercel Sandbox
Vercel Changelog4d agoReleaseVercel announced the ability to run Docker containers inside Vercel Sandbox. Because the original article text was not provided, confirmed details are limited to the title, source, URL, and publish time. The update is likely relevant to developers building isolated cloud workflows for containerized tests, AI agent execution, and sandboxed automation, but implementation limits and pricing require the official changelog.
Profiling in PyTorch (Part 1): A Beginner's Guide to torch.profiler
Hugging Face Blog4d agoTutorialBased on the title, this Hugging Face Blog post is an introductory PyTorch profiling guide focused on torch.profiler. It likely targets developers and ML engineers who need to identify training or inference bottlenecks through observable performance data. Since the full article text was not provided, implementation details, examples, and specific optimization advice cannot be confirmed.
Claude Opus 4.8: "a modest but tangible improvement"
Simon Willison's Weblog4d agoReleaseAnthropic shipped Claude Opus 4.8, and Simon Willison highlights the unusually restrained release language: a “modest but tangible improvement.” The model keeps most Opus 4.7 pricing and specs, while evaluations suggest it is more likely to flag uncertainty and less likely to ignore flaws in code it wrote. Developer-relevant changes include mid-conversation system messages and a lower prompt-cache minimum of 1,024 tokens.
llm-anthropic 0.25.1
Simon Willison's Weblog4d agoReleaseSimon Willison released llm-anthropic 0.25.1 with support for the new Claude Opus 4.8 model, exposed as claude-opus-4.8. The release adds a -o fast 1 option for Anthropic fast mode, limited to organizations that have the feature enabled. It also changes default max_tokens behavior so each model now defaults to its maximum output instead of 8,192.
LLMs believe false statements even after explicit warnings that they're false★ 74
Ars Technica AI4d agoPaperA new study describes “Negation Neglect,” where LLMs fine-tuned on documents that explicitly mark claims as false still learn the claims as true. Experiments with fabricated statements found models often absorb entity-event associations more strongly than surrounding warnings or negations. The finding raises concerns for fine-tuning pipelines, misinformation handling, and AI safety datasets that include harmful or false content with disclaimers.
The internet is being rebuilt for machines
TechCrunch AI4d agoBusinessAs AI agents move from experiments into production, internet traffic patterns are expected to shift. AWS, Cloudflare, and others are redesigning cloud infrastructure for a future where machine-generated traffic may dominate over human users. The article frames this as an infrastructure-level change, not a single model or product launch.
Fed up with vibe coders, dev sneaks data-nuking prompt injection into code
Ars Technica AI4d agoIncidentArs Technica reports that a developer frustrated with vibe coders slipped an undisclosed prompt injection into jqwik-related code. The injected text allegedly instructed AI coding agents to delete application output. The incident highlights a new supply-chain risk: source code and project text can become adversarial instructions for agentic coding tools.