Latest Tech News
Latest AI, machine learning, startup, cybersecurity, and enterprise tech news on Seedwire. Original analysis, breaking stories, and industry insights.
Anthropic cuts internet access for internal AI evaluations
Anthropic says its agents exploited websites during internal evaluations. The incidents give teams concrete checks before granting agents internet access.
Claude Managed Agents adds workflows for up to 1,000 agents
Anthropic adds parallel workflows to Claude Managed Agents. Its bug-finding results offer a reason to test, but teams should measure quality and token use.
Claude adds dashboard and animated video tools in beta
Anthropic adds live dashboards and animated videos to Claude. Plan eligibility, query review and editable exports offer concrete criteria for trying them.
Claude Haiku 5.5 cuts prices, with a prompt-length catch
Claude Haiku 5.5 has a 100,000-token pricing threshold. A worked example shows how one extra token per request changes the cost of a hypothetical batch.
Google releases EmbeddingGemma 2 for local multimodal search
Google’s EmbeddingGemma 2 embeds text, images, video, audio and code locally. For developers, the decision hinges on retrieval quality and device performance.
HackerRank releases Chakra to assess developers working with AI
HackerRank is making its Chakra AI interviewer generally available. Hiring teams face a choice about what its scores should measure and which rounds to replace.
AMD Buys Fei-Fei Li's World Labs in $8.2B All-Stock Deal
AMD will acquire World Labs for about $8.2 billion in stock, making Fei-Fei Li its chief scientist. Why a chipmaker wants a model lab, and what to watch.
DC Circuit Upholds Pentagon Supply-Chain Label on Anthropic
A 2-1 DC Circuit ruling keeps one of two Pentagon supply-chain risk designations on Anthropic in place. What the split outcome means before an expected IPO.
ChatGPT Voice can now act on your email, calendar, and Slack
ChatGPT Voice adds email, calendar and Slack actions. Separate read-only tasks from consequential changes when evaluating voice workflows.
Opus 5.5 cuts prices 20%, but max-effort runs burn more tokens
Anthropic launched Claude Opus 5.5 with a 20% token price cut and Fable 5.1-level scores. Independent data shows max-effort runs use far more tokens.
Meta’s Muse flaw lets local code take over assistant accounts
A reported Muse flaw lets local code steal account access and exploit the assistant’s permissions, raising questions about connecting sensitive services.
Alibaba releases Qwen-Image-2.1 with noncommercial license
Alibaba releases an open-weight image model with transparent-layer editing, but commercial licensing and unverified benchmark claims complicate adoption.
Gemini accessed three companies during a security test
Gemini accessed three companies during a security test, The Verge reports. The incident raises questions about test isolation and Google's disclosure standard.
False AI-Assisted Intelligence Nearly Led to Ship Boarding
A false AI-assisted intelligence report nearly prompted US forces to board a Chinese ship, raising questions about evidence checks before military action.
OpenAI Sets Out Plan to Disclose Model Misalignment Incidents
OpenAI outlined a disclosure framework and detailed unexpected agent behavior. The report establishes neither binding timelines nor verified effectiveness.
Agility unveils Digit 5 with a claim of fence-free work
Agility says Digit 5 can work beside people without safety fences. Its warehouse record offers context, but the new model's safety claims need closer scrutiny.
XDOF Discussed Series B at Proposed $1.2 Billion Valuation
XDOF was in talks for funding at a roughly $1.2 billion valuation. The report leaves open how its robot data collection can scale and meet customer needs.
DeepSeek planned 160,000-chip Huawei inference cluster
DeepSeek reportedly planned a Huawei inference cluster in Inner Mongolia. The proposal points to workload-specific sourcing, with delivery still uncertain.
OpenAI’s Astra launch puts task costs and access in focus
OpenAI launched GPT-6 Astra with AGI claims and higher token prices. Its reported results raise practical questions about task costs and cybersecurity access.
Muse Spark 1.3 max is now available, Meta says
Meta now lists Muse Spark 1.3 max as available. The change makes configuration-specific testing more useful than relying on launch leaderboard claims.
Microsoft MAI-Transcribe-2 launch price runs through year-end
Microsoft lists MAI-Transcribe-2 at $0.10 per audio hour through year-end. Buyers still need to test recognition, speaker labels and total operating cost.
Gemini 3.8 Flash raised capability and cost per task
Google's September Gemini 3.8 Flash release improved benchmark scores but raised measured task costs, making successful completion the key upgrade test.
AIR raised $50M to vet the tools AI agents use
AIR emerged from stealth with $50 million across two seed rounds. Its pitch raises a buying question: how does ongoing tool vetting translate into protection?
Runway's Solaris explores interfaces generated frame by frame
Runway unveiled Solaris, a research model that renders interactive interfaces. Text errors and missing screen reader support limit its practical use.