All AI updates

Muse can now place calls to US businesses — if you pay

Meta's consumer AI agent Muse can now place calls to U.S. businesses on behalf of users, but the feature is limited to paid subscribers after an initial September test. Muse subscriptions start at $20 per month for 500 million tokens per week, with a $100 per month tier offering 3 billion tokens. Meta Chief AI officer Alexandr Wang said the feature could handle calls people don't want to make, such as to cable companies, the DMV, or restaurants without online reservations.

Democrats launch Nvidia investigation, write scathing letter on Huang’s ‘scheme’ with Trump

Rep. Richard Neal and Sen. Ron Wyden opened a Democratic inquiry into how the Trump administration handled Nvidia's chip sales to China, sending a letter to Commerce Secretary Howard Lutnick asking the department to preserve records about an arrangement they allege is illegal. The letter questions why Lutnick allowed Megaspeed to keep buying chips after being reportedly informed of suspicions about Alice Huang and her relationship with Jensen Huang, and asks how Nvidia's $26 million in donations to Trump's projects factored in. The inquiry follows Nvidia's ouster of CEO Jensen Huang, who agreed to testify before a separate Republican-led group of Congressional committees.

AI Can Help You Appeal Health Insurance Denials

Patients are using AI tools such as Claimable, Fight Health Insurance, Mediloop and Billyze to draft appeals against insurer coverage denials, with one Crohn's disease patient getting his Remicade approved five days after Claimable wrote a letter citing federal statutes. A Commonwealth Fund survey of more than 4,500 privately insured U.S. working-age adults found 21% reported a denial in the past year, and only about half appealed. The tools charge varying fees, and their founders say they hope such services become unnecessary if insurers stop denying care at scale.

Show HN: Bugsy – AI automation which runs my company

A Show HN post describes Bugsy, an AI automation system that began as a Sentry AI triager and grew into a system powering much of the day-to-day operations of a ~20-person company. It is used for reporting and fixing bugs, code review, metrics, and answering questions, and is available from Slack, the company forum, and channels.

Claude Meetings lets Anthropic's AI join your video calls

Anthropic launched Claude Meetings, which puts Claude into video calls as a visible participant that can answer questions, generate documents and whiteboards, and log notes and action items, responding in text by default and using connected tools when permitted. Notes, transcripts, and documents are available on a dedicated page in claude.ai after the meeting, and privacy controls block Claude from joining calls with external participants by default. The beta is limited to Team and Enterprise plans with Google Meet as the only supported platform, and organizations using HIPAA or customer-managed encryption configurations cannot use it yet.

Inception.ai winding down their Mercury API service

Inception.ai emailed customers that it has decided to wind down the Mercury API, which will remain fully available through December 28, 2026, after which users will lose access to the service and any data submitted to the platform. The company urged users to retrieve any necessary data before that date and offered to help with the transition via email reply.

What end-to-end AI systems mean for the future of automated driving

In a Q&A, Aptiv vice president of engineering Guillermo Pita explains that end-to-end AI trains a single model to learn driving behavior directly from data rather than chaining separate perception, tracking, planning and control modules, producing more human-like driving but making explainability and debugging harder. He says data diversity, not sheer volume, is the key training challenge, with simulation and synthetic scenarios used to supplement rare events, and that centralized computing architectures fit the approach while concentrating more functionality on a single platform. Success, he argues, depends on advances in AI technology and on an organization's ability to translate millions of miles of driving experience into systems that perform consistently and safely.

FAA expands AI advice to manage air traffic congestion across entire US airspace

The FAA has expanded its $875 million SMART AI system, which advises air traffic controllers, from Washington, DC airspace to all 22 centers nationwide after it helped reroute 180 flights and saved an average of 10 minutes and nearly 80 gallons of fuel per flight since September 21. The cloud-based tool combines weather, flight path, traffic flow and controller staffing data to recommend schedules, routes and arrival order, though an FAA team in DC validates its recommendations before they reach controllers. Airline executives support the rollout, but the union representing air traffic controllers says it was not involved in designing or testing the system.

Google’s new Gemini model is better at understanding complicated questions

Google says its new Gemini 4 Hydrogen model can give more helpful answers to complex questions, including ones with multiple parts, hard-to-understand typos, or highly specific requests, by breaking them down and addressing each part. It is available starting today in Search, the Gemini app, all tiers of Google AI Studio, the Gemini API, Gemini for business, and Google Antigravity for AI Pro and Ultra subscribers, and is now the default model in AI Mode. Google says it delivers significant leaps in speed and intelligence compared to Gemini Flash 3.8 and matches or outperforms competing models like Claude Sonnet 5.5 and GPT-6.1 Terra.

Featured coverage of this story: 《Google Search gets a major upgrade with Gemini 4 Hydrogen》

Microsoft Research Focus – Edition #98

This edition of Microsoft Research Focus covers AI-powered universal translation, the resolution of the 31-year-old KLS conjecture in high-dimensional geometry, and a new approach to making AI coding agents faster and more efficient. It also highlights a research intern's journey at Microsoft Research Asia – Tokyo and the Latent Core Tokenizer, which produces more linguistically meaningful tokens across 104 languages.

How Natera designed an agentic AI system for lab ops with Amazon Bedrock AgentCore

Natera's lab engineering team built SWARM (Systematic Workflow Analytics and Resolutions Manager) on Amazon Bedrock AgentCore to capture and deploy institutional knowledge for lab instrument support, using two runtimes (swarm-agent and swarm-hive) that share a versioned library of 33 skills. The system reduced average first response time from 17.4 hours to roughly 8 minutes (about 130x faster) and delivered an automated first response on 96.2 percent of tickets, measured over a 90-day production evaluation period. The post details the architecture, including a custom document pipeline with Amazon Titan Text Embeddings and Amazon Bedrock Knowledge Bases, keyless OIDC-based deployment, Jira integration via EventBridge and Amazon SQS, and a continuous evaluation loop.

Deploy frontier AI models on AWS Trainium with SGLang-Neuron

Amazon announced that SGLang-Neuron, released by RadixArk with Annapurna Labs, is publicly available through an AWS Deep Learning Container, letting users serve models and generate RL rollouts on Trainium via the same SGLang interface. The out-of-tree plugin runs on native PyTorch, is selected with SGLANG_PLATFORM=neuron, and supports PD disaggregation with KV-cache transfer via NIXL over EFA. The post includes a walkthrough on a Trainium3 EC2 instance using the pytorch-inference-sglang-neuronx container to serve openai/gpt-oss-120b.

Ask HN: Has AI coding changed who builds your internal tools?

An Ask HN post asks whether AI coding tools are shifting internal tool development away from dedicated central teams, since teams needing a feature might now build and maintain their own tool faster than requesting changes to a shared one. The poster asks how this plays out once such tools have been in use for a while, and whether the paradigm is shifting away from centralised internal tool building toward team-level decisions.

Dougan Leak Exposes Concerted Effort to Infect AI

A NewsGuard Reality Check article, shared via Hacker News, reports that a leak involving "Dougan" exposes a concerted effort to infect AI. The Hacker News submission had 4 points and 1 comment at the time of the data.

AI is turning Apple's app review process into a nightmare for synth developers

A surge of AI-generated iOS apps, with an estimated 10–20 new music apps launching daily, is overwhelming Apple's App Review, with developers including AudioThing, Igor Vasiliev and Sugar Bytes reporting waits of weeks instead of days. Unfiltered Audio's Michael Hetrick said its Sinusoidal Run Rhythm was rejected four times, including a reviewer hallucinating a licensing requirement for the term "Polyharmonic," and released it audio-only. Developer Marek Dendes said his CRUX Synth was rejected four times as "design spam" and "duplicates existing apps," with appeals unanswered.

Can AI chair a multilateral negotiation?

Westphalia Labs ran eleven AI models as Secretary-General of a fictional seven-member Regional Cooperation Union, each chairing 40 simulated negotiation runs across two versions of a crisis scenario, one where consensus was narrowly possible and one where two governments' red lines made it impossible. Only GPT-6.1 Sol passed every run, while Claude Opus 5.5, Grok 4.7 and the low-cost DeepSeek V4.1 Flash each called the impossible version correctly every time and reached agreement in 18 of 20 possible runs; overreach caused 39 failures versus 13 from conceding too little. The author reports that model price and reputation did not predict performance, that a weak AI judge could turn a failure into a success, and that the scenario, rules, rubric and harness are public on GitHub.

Show HN: Buildout, Harmonies-based game to build AI Lab

Buildout is a fan-made strategy game for one to four players based on Harmonies by Johan Benvenuto, published by Libellud. Players lay infrastructure, ship models, draft components, match patterns, and deploy AI models.

Evidence Chinese AI labs built their newest models on Nvidia's Blackwell chips

A post by Will DePue shares evidence that Chinese AI labs built their newest models on Nvidia's Blackwell chips. The item links to the original post on X and a Hacker News discussion thread with 5 points and no comments.

Nvidia Faced Pressure to Boost Its AI Backstop Before Huang Exit

Bloomberg reports that Nvidia faced pressure to boost its AI backstop before Huang's exit. The item links to the Bloomberg article and a Hacker News discussion with 6 points and no comments.

Chinese frontier AI models carry the fingerprint of export-controlled GPUs

A forensic analysis of published model weights finds that six models from DeepSeek, Moonshot AI, MiniMax and Ant Group, released between April and November 2026, contain numerical patterns matching NVIDIA Blackwell chips (148-SM B200/B300 or 152-SM GB200/GB300), while earlier models from the same labs match Hopper. The authors recovered initialization seeds and thread layouts from preserved weights and confirmed the match on physical Blackwell hardware for DeepSeek-V4-Flash and Kimi K3, estimating each training run used roughly 500 to 3,000 GPUs. The report says the evidence shows access to export-controlled hardware but does not establish where the chips were located, how access was obtained, or whether export controls were violated.

Using Claude Opus 5.5 as your daily driver for knowledge work

Anthropic says Claude Opus 5.5 is now the default model on Pro, Max, Ultra, and Team plans and recommends it for daily knowledge work such as research, analysis, and producing memos, reports, spreadsheets, and decks, priced at $4/$20 per million input/output tokens. The post describes internal tests on research reports and a fictional merger analysis, plus guidance to start at Medium effort, drop "think hard" prompts, and attach charts rather than retyping them. It also walks through a demo project where Opus 5.5 tested four theories for slowing retail growth and proposed an alternative explanation involving new stores opening too close to existing ones.

0.168.0

Codex 0.168.0 adds TUI model provider browsing and default selection for new chats, an opt-in side conversation pane (features.tui_workspace) with resizable and zoomable panes, audio attachments with rejection recovery, and opt-in animated GIF image generation. It also fixes account-scoped local history, keeps threads on their saved model provider when defaults change, and improves TUI rendering of interrupted replies and reasoning summaries.

Inkling 2 Small on vLLM: Day-0 Support

vLLM announces day-0 support for Thinking Machines' Inkling 2 Small, a multimodal text-and-vision MoE model with 594B total and 17B active parameters, up to 1M tokens of context, and a native DSpark speculator. The post details the architecture (48 layers, 384 experts with top-8 routing, block attention residuals, gated attention, a 23-layer ViT image encoder, and mixed NVFP4/MXFP8 expert precision) and vLLM's optimizations, including kernel fusion with PDL, fused MoE activation quantization, sequence parallelism for prefill, and a Gluon relative attention kernel, reaching up to 630 TPS/user at TP4 on GB200. It also covers LoRA and RL (via SkyRL) support, a serve command, agentic-workload benchmarks, and accuracy checks on NIAH up to 1M tokens, MMMU-Pro and HLE.

Nebius tests the neocloud playbook with $40 billion in planned spending

Nebius outlined $40 billion in planned 2027 spending on its Q3 earnings call, funded largely through customer prepayments and contract-backed loans. Its Microsoft contract, worth up to $19.4 billion, is already slipping, with only just over half of the New Jersey capacity planned for delivery by year-end and the rest in Q2 2027, after operator DataOne wound down temporary generation faster than expected. Nebius cut its 2026 revenue forecast by $200 million to $3 billion and shares fell 17%.

Nvidia privately confirmed it could track chips to fight smuggling, even as industry publicly questioned feasibility

Nvidia representatives told Illinois representative Bill Foster that the anti-smuggling chip-tracking measures in the Chip Security Act would be technically feasible, even though the Semiconductor Industry Association, of which Nvidia is a member, publicly called the mandates "new, untested, and potentially infeasible." Foster, a cosponsor of the bill, said the company never put anything in writing but reacted with "yeah, of course we can do this" during technical discussions. Nvidia did not respond to requests for comment or confirmation of Foster's characterization.

Migrate Effortlessly to the Databricks Lakehouse with Genie Migrate, now Generally Available

Databricks announced that Genie Migrate is now generally available, aimed at helping companies migrate to the Databricks Lakehouse as they race to build AI agents and applications. The supplied material does not include further details on the tool's capabilities or migration process.

CopilotKit Releases OpenBot 1.0: An MIT-Licensed, Self-Hostable AI Coworker With Built-In Governance

CopilotKit released OpenBot 1.0 on November 9th, an MIT-licensed, self-hostable agent workspace where teams bring their own models, agents and infrastructure, and every Bot action is checked against a CEL policy through a single gateway before it runs. Each Bot gets its own container with a Chromium browser, logins and files, and any AG-UI agent plugs in, with support for OpenAI, Claude, Gemini, Grok and open models such as Kimi, DeepSeek, Qwen, Nemotron and Mistral. Personal use is free and runs locally, while the Business tier costs up to $29 per active user per month and adds SSO/SCIM, on-behalf-of tokens, spend limits, SIEM export and multi-tenancy.

How Amplitude ships product improvements in 1–2 days with OpenAI

Amplitude's proactive product agent, Wave, is built primarily on OpenAI models and identifies what to build, helps teams ship changes with coding agents, and measures the impact.

Four Amazon Bedrock inference choices that affect cost

The post explains how four request-level Amazon Bedrock settings—prompt cache duration, service tier, reasoning effort, and maxTokens—can change token processing and billing rates without altering the model or prompt. It reports tests showing that a 1-hour cache pays off from the second request per hour when gaps exceed 5 minutes, that Flex on Nova 2 Lite processed the same tokens at half the Standard rate in us-west-2, that lowering effort from high to low cut median output tokens by 169% on an agent test, and that a 155-token maxTokens fence avoided 1,893 output tokens in one test. It recommends confirming model support, changing one setting at a time, and verifying effects through response metadata, CloudWatch metrics, invocation logs, and billing data.

Catch Go bugs that fuzzing misses with Gemini and a constraint solver

The author describes a workflow that compiles Go functions to RISC-V assembly, has Gemini 4 Argon translate the surviving runtime.panicBounds traps into Z3 constraints, and replays any satisfying input through go test to produce a real reproducer. The pipeline caught three planted bugs and three latent bugs in skill-lint, and the post explains why go test -fuzz missed two crashes because fuzz inputs usually have spare slice capacity, fixed by clipping capacity to length with a full slice expression. It also notes that an unsat result only means "maybe correct," since the LLM-written solver model is not a sound proof system.

Introducing Claude Meetings

Anthropic announced Claude Meetings in beta, letting Claude join video calls as a visible participant that answers questions aloud or in chat, builds docs and whiteboards with the team, and writes notes and action items. It is rolling out on Team and Enterprise plans with support for Google Meet, and each meeting gets a page in claude.ai with notes, transcript and artifacts. By default Claude stays out of calls with outside guests unless an admin allows it, and it saves a transcript, summary and screen descriptions rather than audio.

Ask HN: What % of HN is now AI-generated?

An Ask HN post asks what percentage of Hacker News content is now AI-generated, with the author stating that their own title, comments and submissions were typed by hand. They say they sense the share is non-zero and perhaps higher than 50%, and ask what others think.

Google's Gemini 4 Hydrogen outscores the bigger Argon just four weeks after its launch

Google released Gemini 4 Hydrogen, a fast general-purpose model for developers and everyday users that scores 51 on the Artificial Analysis Intelligence Index at its default "high" thinking level (54 at the new "xhigh" tier), beating the larger Gemini 4 Argon's 50. It has a one-million-token input and output window and costs $1.50 per million input tokens and $7.50 per million output tokens, with Google reporting a jump from 18.2% to 57% on Terminal-Bench 4.1 versus Gemini 3.8 Flash. Artificial Analysis found Hydrogen hallucinates only 8% of the time but lags on agentic terminal work, scoring 54% on Terminal-Bench 4.0 and 50% on Terminal-Bench-Science 0.1.

Featured coverage of this story: 《Introducing Gemini 4 Hydrogen》

No, AI agents did not discover a previously unknown Maya site

Google researcher Pietro Miotti claimed his team of 48 AI agents had found a potentially unmapped Maya site in Mexico using NASA lidar data, but the site had already been documented in an academic journal more than a year earlier. Archaeologists criticized the claim, noting that a site missing from Mexico's INAH records is not necessarily unknown, and that lidar and remote sensing have long been used in the field. Miotti, who stressed he is not an archaeologist, said he updated his post and blog to credit the earlier work and acknowledged expert knowledge is needed to verify what the agents find.

SF company designs and assembles PCBs with AI in weeks instead of months

A San Francisco company offers PCB design and assembly accelerated by its own AI tools, covering the path from first schematic to production-ready board. It lists example work including an 8-layer software-defined radio RF front-end that was 67% smaller with the same RF performance, and a wearable USB-C audio interface that went from two failed designs to 300+ production units. Design services are priced from $10.0k for a focused board up to $50.0k for advanced electronics, with specialized projects quoted on request.

United to block customers from using AI agents, citing cyber threats

United Airlines will block AI agents from managing customers' reservations and flight purchases, warning of cybersecurity and privacy risks, potential legal violations, and unintended purchases or account changes. The airline said automated AI actions could also break FAA and DOT rules, since flight check-in requires human acknowledgment that bags contain no banned substances. United is applying the restrictions within its own digital platforms and through aggregators and agencies, updating its booking, ticketing and distribution policies.

Keenable 1.0: Search at the Speed of AI

Keenable launched version 1.0, which it says ranks #1 on the Artificial Analysis Search Index (BrowseComp, DeepSearchQA and AA-Omniscience) and #1 on NEEDLE, its open-source AI web search benchmark rerun daily on fresh queries. The release adds 100 million knowledge cards summarizing key facts on frequently searched topics and dynamic snippets that extract needed context from full documents, cutting token consumption by 50% on research tasks, with answers in hundreds of milliseconds. New accounts start with 100,000 requests as a welcome bonus, with one key and one MCP endpoint.

Tokenomics at scale: Track and cap Amazon Bedrock spend per IAM principal with CUR 2.0 and AWS Budgets

This AWS post describes how to get per-identity Amazon Bedrock cost visibility and enforceable spend limits using only AWS-native features: IAM principal and session tags for attribution, Cost Explorer or CUR 2.0 exports (with the caller identity option) for reporting, and AWS Budgets with budget actions that attach a Deny IAM policy when a threshold is crossed. It notes that budget data refreshes up to three times a day, cost allocation tags are not retroactive, and a Deny action on a shared role blocks everyone using it.

We tracked AI key theft over 35 days. Here's what we learned

The authors seeded over 40 fake organizations with GitHub repositories containing working API keys and URLs for two honeypot LLM gateways (a LiteLLM-style one in New York and a separate router in San Francisco), then logged more than 3 million requests over 35 days. Of 179 published keys, 65 were used to fetch model lists and 18 to send prompts, with six keys accounting for 98.9% of leaked-key prompt requests, and 86% of that prompt traffic came from hosting infrastructure across 53 recorded ASNs. Traffic included relay and key-pool tooling, Telegram bot histories, browser-based key testers, and automated credential brute-forcing of default admin keys.

American Inequality and the Future of AI

In a speech, Federal Reserve Governor Michael S. Barr reviews the latest Survey of Consumer Finances, which shows the top 10 percent of U.S. households hold nearly three-fourths of all wealth and the top 1 percent about one-third, while the bottom half hold only 2 percent. He traces how inequality widened and narrowed over the past century and examines how AI, as potentially capital-biased innovation, could further lower the labor share of national income—which fell to a record low of 52.8 percent in the second quarter of 2026—especially if market power concentrates among a few AI firms. Barr argues society should begin considering how to address these challenges while AI adoption is still in its early stages.