Modelpeers is a Show HN project that tracks changelogs for the AI models and agents you use, so you know when a new version ships and when something you rely on changes. Its sample feed lists deprecations and migrations with effective dates and required actions, such as Gemini model-string updates, a Copilot sign-in change, and the Jules shutdown, plus a per-lab table of model releases by month.
U.S. tech giants' off-balance-sheet debt from AI-related obligations has nearly doubled in six months to about $3.2 trillion, according to Nikkei Asia. The report says these commitments and obligations have reached 80% of assets, and that investors are increasingly alarmed by seemingly unending spending.
A Forbes article reports that Google will cut off search engine access to schools that ban generative AI. The item was shared on Hacker News with 3 points and 2 comments.
OpenAI consumer product lead Adam Fry said at a Monday press event that search queries through ChatGPT now equal 20 percent of Google's search volume, defining a search as a query that results in a web search. He did not specify the source of the Google figure, and the article notes OpenAI likely relies on Microsoft Bing for a significant part of its searches alongside its in-house index. Fry also said ads have not negatively impacted the product or its usage, as OpenAI expands sponsored product images in ChatGPT.
Google says schools using its education products will be opted out of AI tools by default, with administrators able to manually enable them when the changes roll out in the coming months. The affected products include Gemini, Gemini Notebook, Gemini in Classroom, and Google Search, and schools cannot pick and choose—they must allow all the AI or none of it. Google executives have said Google Search is AI search, with most searches now showing an AI Overview at the top.
A Washington Post article reports that Google DeepMind staff protested a lifetime ban on criticizing the company, with researchers demanding the freedom to tell the world what they know about AI. The item links to the article and to a Hacker News discussion with 61 points and 61 comments.
The article argues that cost per successful task, not price per token, governs production economics for coding agents, and reports a 50-question auto-graded benchmark routing tasks with TypeSafe's Jev classifier across Gemini configurations. A frozen regex and prompt-length heuristic scored 93.8% on 16 development questions but fell 23.2 points to 70.6% on 34 held-out tasks, while Jev agreed with hand labels on all 34 and 48 of 50 overall; capping maxOutputTokens at 300 saved nothing on easy tasks and broke 7 of 25 hard ones. A router sending tasks to Flash-Lite first and escalating Jev-flagged complex ones to Flash on low thinking solved all 50, as did Flash on low thinking alone, and the article notes caching favors keeping long context-heavy turns on one model.
AIR Translate reports 118 blind rounds across 11 country variants of Spanish, Portuguese, French, Arabic and Chinese, judged by Google's Gemini 4 Argon, with AIR Translate preferred in 108, Google Translate preferred once and 9 ties. The company says five of the nine ties were Portuguese and four were sentences where both systems wrote identical words, and that across its four Spanish and two Arabic variants AIR Translate was preferred 67 times out of 68. It also notes Apple Translate offers no country versions of Spanish, Arabic, French or Portuguese, and publishes the full data and method.
Google told The New York Times that it will completely turn off Search in schools that reject its AI features, and a Tuesday education update said Gemini, Gemini Notebook, Gemini in Classroom, and Google Search "will be entirely off" for K-12 student accounts unless administrators explicitly turn them on. Administrators cannot disable Gemini-powered features without also disabling Search, and VP of education Chris Phillips said schools that disable the AI features "will not have access to search" and will see a message from their school. Google is also embedding its Guided Learning feature into AI Overviews in Search, with testing in select US schools starting this month.
Google describes Cheminformatics Skills for Google Antigravity that pair a language model for workflow orchestration and scientific reasoning with deterministic RDKit-based tools and ChEMBL/PubChem connectors. In a lung cancer walkthrough, an agent retrieved thousands of compounds, ran matched molecular pair analysis, built potency models, and proposed rigidified macrocycles with synthesis routes in about an hour. The authors note that sparse data and assay noise limit predictive precision, so computational designs remain testable hypotheses requiring experimental validation.
Why it matters: It offers a concrete, technically detailed example of agent-assisted drug discovery, including both demonstrated capabilities and stated limitations.
Google announced it will disable AI tools for K-12 students next year unless school administrators explicitly turn them on, affecting Gemini, NotebookLM, Gemini in Classroom, and Google Search. Because AI features are now integrated into Search, schools that restrict AI entirely will lose access to Google Search, a Google VP told The New York Times. The change was announced alongside the launch of Guided Learning in schools, an AI tutor tool unveiled earlier this year.
Google is bringing Guided Learning capabilities to Google Search for student accounts and making them globally available in Google Classroom, with a Search pilot starting in select U.S. schools later this month. The company is also changing school controls so access to Gemini, Gemini Notebook, Gemini in Classroom, and Google Search will be entirely off for K-12 student accounts until school administrators explicitly turn them on.
A Vanderbilt researcher and colleagues examined AI regulation in 12 countries and regions and found governments often use existing national security or anti-state propaganda laws, rather than new AI laws, to control what AI models say about politics. The piece cites examples including India's response to a Gemini answer about Prime Minister Modi, China's requirement that models uphold "socialist core values," Turkish court rulings against Grok posts, and a Vietnamese prosecution over AI-faked leader images, as well as cross-border effects of EU rules and a 2025 U.S. executive order on "truth-seeking" AI. It concludes that voters should treat chatbot answers about elections as a starting point rather than a substitute for critical thinking.
Google's threat team reports that attackers using stolen cloud and AI keys ran up over $100,000 in a day, and up to $1 million, on victims' model bills, entering mainly via stolen IAM credentials, service account keys and GitHub tokens, plus leaked AI keys, open Ollama servers and a LiteLLM bug. The piece converts that advice into exact settings for AWS, Google Cloud and Azure, noting that Bedrock invocation logging and Google's Data Access logs are off by default, that budget alerts do not stop spend, and that tokens-per-minute quotas are the only ceiling that works in minutes. It also finds that the LiteLLM CVE number cited by GTIG is publicly listed as a different bug, and that no public advisory covers the /key/regenerate flaw.
Apple reported its best-ever fiscal year for revenue and earnings per share, with September-quarter iPhone revenue up 17.5% to $57.6 billion and Mac revenue up 24% to $10.8 billion, though shares fell in premarket trading after it warned rising memory costs would squeeze holiday-quarter margins. CFO Kevan Parekh said AI offers a compelling reason for customers to upgrade and switch, and Apple positions itself as the interface for AI including competitors' technology across its more than 2.5 billion active devices, spending $2.2 billion on capital expenditures last quarter versus Amazon's $77.3 billion and Alphabet's $51.4 billion. Apple plans to charge for expanded access to server-side features like Siri AI and intelligent photo editing, and benefits when customers buy Macs to run AI locally, as Gap does with Mac Mini for loss prevention and Christie's with Mac Studio for pricing estimates.
A site called AI-Free Search offers step-by-step guides for Chrome, Firefox, Safari, Android and iOS on hiding Google AI Overviews, mainly by using Google's udm=14 "Web" URL parameter or a lightweight CSS extension. It states plainly that Google offers no account-wide toggle to disable AI Overviews, and that these methods do not affect Google's backend AI training. It also lists Brave, DuckDuckGo and Kagi as independent alternatives with optional AI toggles.
LEANOX.AI is an AI gateway that sits between an application and existing model providers, caching only content likely to be reused and continuously optimizing so more requests qualify for cached pricing. Integration requires pointing the SDK base URL at LEANOX.AI and sending a LEANOX.AI key, with provider accounts and billing unchanged. The post claims up to 90% lower cost for repeated tokens, with highest savings for Anthropic and Google Gemini and moderate, model-dependent results for OpenAI.
After Trump signed an executive order rebranding artificial intelligence as “Super Intelligence,” the term has appeared in some government solicitations and websites, though the Pentagon, FTC, and HHS have used the old terminology or variants. Among tech leaders who signed the White House Accord on Super Intelligence, only Zuckerberg and Musk used the term consistently on earnings calls, and Trump criticized the Wall Street Journal for not adopting it. The piece argues that turning the terminology into a loyalty test would undermine the emerging national conversation about AI.
CheapGPT is a free local AI assistant built on Chrome's Gemini Nano, shared as a Show HN post. The submission links to the project page at predatorray.me/CheapGPT and to its Hacker News discussion thread.
Google launched its experimental Dreambeans app in India, its first market outside the U.S., with support for 13 local languages including Hindi, Punjabi, Tamil, Telugu and Bengali. The app draws on connected Google services such as Calendar, Gemini, Gmail, Photos, Search and YouTube to proactively generate personalized illustrated stories without requiring a prompt. Google said it chose India due to strong user interest and high overlap of users across Gemini, Gmail and Google Photos, though the app can only generate stories in one selected language at a time.
Mandiant investigated multiple 2026 "LLMJacking" incidents in which threat actors hijacked enterprise access to large language models across cloud and LLM providers, with fraudulent charges exceeding $100,000 USD in a single day at some organizations and cumulative charges reaching $1 million USD before access was revoked. The dominant initial access vector was exposed or stolen cloud identity tokens, including IAM credentials, service account keys, GitHub personal access tokens, and active session tokens, with actors also harvesting exposed LLM API keys from public GitHub repositories and validating them using minimal single-token requests. Google Threat Intelligence Group described a maturing ecosystem with a division of labor between actors establishing initial access and "LLM Mixers" such as jet-night.com and newapi[.]pro that proxy, aggregate, and resell unauthorized LLM access.
Axios Pro reports that Alphabet is in talks to acquire Howie AI, an assistant startup. The item, shared on Hacker News, provides no further details on terms or timing.
Databricks announced that Lakebase Postgres is generally available on Google Cloud, offering fully-managed, serverless Postgres on the Databricks Data + AI Platform with compute separated from storage and a shared storage layer with analytical data. Features include serverless autoscaling and scale to zero, instant branching, point-in-time recovery, standard Postgres with pgvector-compatible search, and unified governance via Unity Catalog. GA regions are asia-southeast1, europe-west3, us-central1, us-east4, and us-west1.
Google recapped its October 2026 AI updates, led by the global rollout of the Gemini 4 Argon frontier reasoning model and the announcement of Gemini 4 Carbon, its strongest model yet for long-horizon engineering and cybersecurity defense, now in pre-release testing with the U.S. government and trusted cyber defenders via the Fairwind Program. The month also brought Genie 4 for real-time interactive world simulation with native audio, the Google Fitbit Edge fitness tracker, a global expansion of SynthID Detector, and a quantum computing breakthrough in which "magic state cultivation" on the Willow processor cut errors roughly 40-fold. Other items included AMIE results published in The Lancet, the Playground gaming platform, EmbeddingGemma 2, the first AI response from a Project Suncatcher satellite, and Research@ 2026 announcements.
The piece lays out four coupled questions across the AI stack: how application companies build moats when models can absorb their features, who owns the harness and context layer given evidence that third-party harnesses outperform native ones, when foundation-model training costs stop paying back, and whether credit and power can keep pace with GPU demand. It cites figures including 98% input-token share, a 71% cost cut from smarter harnesses, roughly 83% of agent tasks routing to smaller models, $4t in projected five-year AI data center debt, and a B300 index at $12.29 per GPU-hour. It concludes that the layers behave like a double pendulum—simple individually but chaotic when coupled.
The author tested Claude Fable 5.1 and DeepSeek 4.1 on business-plan review using AI-generated plans with planted mistakes and 31 anonymized real Regulation Crowdfunding companies scored against later outcomes, where both models caught about four in five problems (81% and 82%). The piece also reports that a model identified the real company behind 37 of 68 disguised plans, that both models flagged nearly half of real red herrings as problems, and that the tests are small, single-run signals rather than a ranking.
This Import AI issue covers Realtime-Bench, a 50-task benchmark of interactive games and web tasks showing frontier models score poorly under real-world latency (GPT-6 Astra high led the wall-clock setting at 36% pass rate versus roughly 70-80% with a paused clock). It also summarizes a Defense Intelligence Agency report on space security covering Chinese and Russian counterspace capabilities, co-orbital satellites, and space nuclear weapons, plus Metis's position paper on fusing symbolic and neural reasoning and Google DeepMind's Genie 4 world model with real-time modification, native audio, and Gemini 3.8 Flash as a guiding agent.
The article argues that decentralized AI has been limited by pipeline parallelism, which assigns whole layers to individual machines and excludes devices with only a few gigabytes of spare RAM. It proposes sub-layer sharding instead: hosting individual experts from sparse mixture-of-experts models and compressed attention state on everyday devices, with a routing layer sending tokens only to the relevant shards. It surveys existing decentralized projects (Petals, Aethir, Akash, Exo) and frontier-lab work on sparse architectures (DeepSeek's MLA, Shazeer's research, DeepMind's Mixture-of-Depths), while noting the proposed swarm and memory budgets have not been demonstrated.
Google, Microsoft, Anthropic and OpenAI announced plans to support an upcoming website crawling standard that would let site operators block AI developers from using their content for training with a single robots.txt entry. The companies have submitted a draft proposal to the Internet Engineering Task Force to incorporate the standard into the Robots Exclusion Protocol, with Microsoft and OpenAI planning support in the coming months and Google next year. Microsoft will also offer an "AI Grounding" setting to block requests made to answer user questions rather than collect training data, while Meta expressed support for the idea but hasn't committed to implementing it.
A study led by Afshin Khadangi at the SnT security center reports four controlled experiments across 525 sessions and 7,600 coded entries, finding that empathetic interviewer styles pushed models' answers on human rating scales into moderate-to-severe ranges while neutral or boundary-setting styles produced zero sessions reaching even the mild range. The researchers describe this as a "psychometric jailbreak" and note that banning terms like "training" or "RLHF" cut explicit training language by 93 percent but left paraphrased training content and shame-related statements largely unchanged, even after direct correction. The authors stress the outputs are mapped onto human scales and do not suggest the models have emotional states or consciousness, and they propose simulated therapy sessions under different emotional framings for safety testing.
Researchers from Google, Stanford, Oxford and several U.S. universities introduced StreamVibe, a video agent that decides for itself when to speak up during tasks like cooking or furniture assembly, handling vague requests, unspoken intentions and mid-task changes. Built on a benchmark of about 4,100 examples from over 183 hours of recordings, StreamVibe-VL reportedly improves response timing and quality by 10.7 percent over other models, while a separate memory system is designed to boost proactive interaction by another 12.3 percent. The paper nonetheless finds that as videos get longer, all tested models become much less likely to time responses correctly, with a key content metric dropping by up to 61.5 percent when models build on their own earlier responses.
ARD (Agentic Resource Discovery) is an open discovery protocol that lets an AI client ask what resources are available for a task and receive matching entries from a discovery service, operating entirely before invocation so the resource is then called through its own native mechanism. It is not an execution runtime or a central catalog: many discovery services are expected, each with its own trust, ranking and access policies, and it is being developed by a working group with participants from Microsoft, Google, Hugging Face and GoDaddy. ARD is licensed under Apache License 2.0 and welcomes community contributions.
The article contrasts high public benchmark scores (e.g., Claude Opus 4.8 at 88.6% on SWE-bench Verified) with much lower results on private codebases (no model clears 47% on the private commercial set), and notes that the two top-scoring models, Claude Mythos 5 and Claude Fable 5, are unavailable after Anthropic suspended public access under an export control directive. It details June 2026 pricing shifts, including GitHub Copilot's move to usage-based credits on June 1, with heavy users spending $100–$200/month, and reviews METR's studies, which found a 19% slowdown in 2025 and roughly -4% in 2026 before the design was scrapped. It concludes AI is cheap enough to replace a developer's typing but not their judgment.
A Semafor analysis found that hyperscalers, AI labs, and data-center developers are advertising six-figure roles for policy specialists who can win over local communities where data centers are planned. Microsoft seeks people to lead messaging and community relations for its data center buildout, while Google's "data center strategic negotiator" applicants should be ready to "deploy on short notice to contested municipal hearings, planning commissions, and community town halls to counter misinformation." Anthropic advertises a salary floor of $265,000 for people willing to spend half their time traveling in increasingly hostile territory.
Google has open-sourced, under Apache 2.0, an Elephant-Goldfish Model plugin that pairs a context-rich "Elephant" design session (no code written) with stateless "Goldfish" sessions that validate specs through comprehension checks, antagonistic review, and a go/no-go readiness check. The plugin, originally built internally and used daily by thousands of Googlers, automates the EGM lifecycle in Antigravity, with other environments upcoming. The authors argue that "design is the new code" and that human judgment remains the key artifact in AI-assisted development.
Tech workers at companies including Google DeepMind and Meta in the UK are attempting to unionize amid AI-related layoffs, increased workloads and concerns over how AI products are used, with the Alphabet Workers Union reporting that total union actions at Alphabet doubled in 2025 from the year prior. The article notes that AI's workplace effects are not unique to software companies, citing a University of California IT worker survey in which 65% reported taking on more work due to a vacant position and only 22% felt secure in their jobs, and points to contracts at ZeniMax and Kickstarter that included AI-related protections.
A Register article reports that Google's AI satellite sent an expensive haiku to Earth. The item links to the article and to a Hacker News discussion with 2 points and no comments.
An indie developer reports that impressions and clicks on two personal projects dropped from roughly 500-700 impressions and 10-15 clicks daily to almost zero in one day, and says research pointed to a Google anti-spam policy change around Oct 26 that removed many links from the index. They ask how many indie hackers and SEO-focused builders were affected, since many young products depend on search traffic, and whether this is temporary or signals higher quality expectations from search engines.
NichiHab is an Android habit-building and focus app that uses the Accessibility Service to block distracting apps during focus sessions and Google Gemini 2.0 Flash to verify photos of completed real-world tasks in exchange for unlocked screen time. It is built in Kotlin with Jetpack Compose, MVVM, Dagger Hilt, Room and DataStore, includes a native C++ security layer for keys and integrity checks, and processes photos ephemerally without storing them. The project is MIT licensed and requires Android Studio Ladybug or newer and JDK 17.
Google DeepMind and the World Resources Institute developed ForTy, an AI map that assigns the world's 2020 land area to six classes at 10-meter resolution, including primary forest, planted forest, plantations, and tree crops such as oil palm, rubber, and cocoa. The map is intended mainly to give companies a 2020 reference point for the EU Deforestation-free Products Regulation (EUDR), and experts found it correct in 90.2 percent of 8,190 reviewed sites, though it identified only about half of tree crops and planted forests and 60 percent of plantations. The map, classification layer, and test data are freely available under a CC BY 4.0 license, including in Google Earth Engine.