Create

Sign in to ReadmeX

or
ReadmeX
ReadmeX

A clearer picture, in a conversation.

Catch up on what matters, then ask a little deeper.

Your community and people briefings stay personal to you.

AI News

‹Thu, October 8›
All dates
1815
Industry

Google tightens Gemini tiers: free users limited to Flash-Lite from Oct 9

Google has updated a support page saying that from October 9 it will change Gemini access for personal accounts: free users will only get Gemini 3.5 Flash-Lite and lose access to Gemini 3.6 Flash and newer models, while the $4.99/month AI Plus tier is limited to Flash and Flash-Lite and no longer includes higher-end reasoning models such as Gemini 3.1 Pro. The $19.99-and-up Google AI Pro and Ultra tiers are unaffected and keep Deep Think and advanced model access. Google is also introducing low, medium and high Effort modes that deduct quota based on task compute complexity, with higher modes draining the allowance faster.

AIbase AI新闻·
1826
Products

Liquid AI pitches on-device personal AI built on Liquid Context

Liquid AI COO Jeffrey Li said at theCUBE's Fully Connected event that phones, wearables, PCs and cars capture the richest signals about users, making the edge the natural home for personal AI. The company's Liquid Context, optimized for Snapdragon processors, sits between models, agents and hardware, and uses Liquid AI's own models to decide what information to retain and how to compress it within fixed on-device compute. Liquid AI is also working with Mercedes-Benz on in-car on-device AI, and says it is building observability and continuous-improvement loops so models and agent harnesses can self-heal and personalize after deployment.

SiliconANGLE AI·
1835
Industry

Top AI labs said to slow frontier R&D as focus shifts to apps and profit

In early October 2026, Anthropic, OpenAI and other leading AI model developers advocated slowing frontier capability work, drawing market attention, according to DoNews. A CITIC Securities research note argues the move reflects not founders' moral considerations but a rational response to rising regulatory costs, a gap between model progress and real-world deployment, and mounting commercialization pressure. The note expects the industry's center of gravity to shift from large-model training toward inference optimization, deployment and monetization, favoring CSP providers, platform software vendors and selected semiconductor segments.

DoNews·
1846
Industry

Station F and Anthropic say the AI boom is pulling founders back in person

Speaking on a panel at Wave by Vento in Turin, Station F director Roxanne Varza and Anthropic's EMEA startup sales head Aiden Blake said the AI boom is drawing more founding teams back to physical offices. Station F hosts 1,000 early-stage companies at all times, and its F/ai programme, launched in February, lets teams go hybrid but requires them on site for key moments; about two-thirds of the current batch is non-French. Blake said trust and access to capital happen in person, and Anthropic runs its EMEA startup support from Dublin with roughly 150 staff.

The Next Web·
1855
Industry

Dan Ives relaunches tech coverage, bullish on Nvidia and AI's 'third inning'

Well-known tech bull Dan Ives left Wedbush Securities after nearly eight years to launch the newly formed Yorkville Ives, where he restarted tech equity research with an Outperform rating on all 49 stocks he initiated on. He argues AI infrastructure supply will stay constrained at least through 2027 and that the market underestimates a coming wave of as much as $4 trillion in AI spending, likening the AI buildout to the third inning of a baseball game. His top five picks for 2027 are Nvidia, Microsoft, Palantir, Apple and CrowdStrike, with initial price targets of $400 for Alphabet and Apple, $300 for Amazon, $900 for Meta, $615 for Microsoft, $300 for Nvidia, $500 for Tesla and $225 for SpaceX.

科技新报·
1865
Industry

Trump's new AI czar Jay Clayton heads to Silicon Valley to meet Huang and Altman

Bloomberg reports that Director of National Intelligence Jay Clayton, Trump's newly appointed point person on AI, will travel to Silicon Valley to meet Nvidia CEO Jensen Huang, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei and Meta CEO Mark Zuckerberg. It is his first meeting with tech leaders since taking charge of a new team coordinating federal AI policy, following a White House lunch where AI developers signed a voluntary safety agreement and Trump's subsequent announcement that Clayton would lead a "Super Intelligence Force" reporting to him and chief of staff Susie Wiles. Speaking to CNBC, Clayton said models must have control and monitoring mechanisms acceptable to developers, enterprises and the American public, and defended using existing liability law as the main tool for holding the industry accountable.

cnBeta·
1875
Industry

Mathematicians' group says OpenAI math documents show power, not scholarship

The Association for Human Mathematics (AHM) said the math documents OpenAI published on October 6, 2026 show power rather than scholarship, and urged mathematicians to stop working with the company. The group's statement comes as OpenAI defends lawsuits accusing it of illegal plagiarism.

Techmeme·
1885
Industry

Fed minutes: AI investment emerges as a new inflation variable

Minutes from the Fed's September meeting, released Oct. 7, show officials judged recent disinflation progress insufficient, while staff raised their 2026–2028 inflation forecasts and pushed the return to 2% out to 2029 — a year later than July's projection. Several officials said fast-growing AI investment is adding to inflation pressure, with its effect on core goods prices appearing to strengthen as tariff effects fade. Officials unanimously supported a 25 basis point rate increase to 3.75%–4.00%, with most favoring another hike before year-end.

虎嗅 AI·
1896
Products

AWS details multi-tenant GPU cluster sharing on SageMaker HyperPod

An AWS Machine Learning Blog post lays out a reference architecture for letting multiple teams share a single Amazon SageMaker HyperPod (EKS) GPU cluster while preserving isolation and fairness. It uses AWS IAM Identity Center for centralized authentication (federating with Microsoft Entra ID via SCIM provisioning and SAML 2.0), per-team SageMaker AI domains and execution roles, and Kubernetes namespaces for workload isolation. HyperPod Task Governance handles fair resource allocation, namespace-level cost allocation provides per-team chargeback visibility, and storage is split across FSx for Lustre or OpenZFS directories and S3 buckets.

AWS Machine Learning Blog·
1905
Products

Asus ProArt P14 and P16 launch with Nvidia RTX Spark and Hermes agents

Asus announced the ProArt P14 and ProArt P16 laptops, both built around Nvidia's RTX Spark superchip, which Asus says offers up to 1 PFLOP of FP4 AI compute and 128GB of unified memory for on-device model inference and AI agent workflows. Both machines also integrate Nous Research's open-source Hermes platform, which can chain desktop apps, terminals and local files for multi-step tasks with long-term memory and sandboxed environments. Pre-orders are open now, with availability on October 16 at 8pm; starting prices are NT$169,999 (P14) and NT$179,999 (P16), while the 128GB pre-order configurations cost NT$249,999 and NT$269,999. Asus also teased the ProArt GR1X Agent Computer, a desktop on-device AI hub, for launch in Taiwan this month.

科技新报·
1915
Industry

TSMC quarterly revenue jumps 51% as AI demand holds up

Taiwan Semiconductor Manufacturing Co. reported a 51% rise in quarterly revenue, according to Bloomberg. The report frames the increase as a positive signal for investors gauging whether the global AI infrastructure buildout is sustainable. No absolute revenue figure or specific fiscal quarter was given in the source text.

Bloomberg Technology·
1925
Open source

LMCache multiprocess mode hit by 9.8-rated RCE flaw

According to iThome Taiwan, LMCache — a platform for large language model inference acceleration and cache management — has a critical remote code execution (RCE) vulnerability in its multiprocess mode, tracked as CVE-2026-105192 with a CVSS score of 9.8. The report does not state affected version ranges, exploitation prerequisites, whether the flaw has been exploited in the wild, or any official fix or mitigation.

iThome 台湾·
1936
Industry

Willow taps CoreWeave to simplify AI model training for its voice dictation app

Willow Care, the company behind AI dictation app Willow Voice, chose CoreWeave for its reinforcement-learning training infrastructure rather than compute alone. Co-founder and CTO Lawrence Liu said at CoreWeave's Fully Connected event that the team no longer has to manage anything under the hood and can focus on eval alignment; Willow fine-tunes its own speech recognition model on quiet, real-world audio such as whispering in a coffee shop, then uses a very small LLM to post-process the output into written text. Liu said the goal is that within two years users will no longer need keyboards.

SiliconANGLE AI·
1945
Industry

SemiAnalysis: China has 24GW of operating compute vs 56GW in the US

China has 24GW of operational compute capacity with another 50GW planned or under construction, according to SemiAnalysis data cited by the Financial Times; the US has 56GW operational. The figure comes in the first installment of a three-part FT series examining how Inner Mongolia became the engine room for China's AI ambitions.

Techmeme·
1956
Industry

CreditSights' Cisar Sees Cracks in the AI Credit Boom

Winnie Cisar, global head of strategy at CreditSights, says the surge in AI financing is making credit analysis increasingly difficult as investors weigh uncertain returns, potential data center overbuilding and more complex financing structures. She spoke on "Bloomberg Surveillance."

Bloomberg Technology·
1965
Research

AI labs said to probe whether models can break cryptographic protocols

Theoretical computer scientist Omer Reingold, citing sources, says several AI labs have quietly begun testing whether their models can break important cryptographic protocols; Scott Aaronson relayed the claim on his Shtetl-Optimized blog with Reingold's permission. No labs are named, no protocols or results are specified, and the claim remains secondhand and unverified.

Techmeme·
1978
Industry

Anthropic launches Claude Frontier Academy with $100M to train 10,000 engineers

Anthropic announced Claude Frontier Academy, a $100 million program aiming to train 10,000 forward-deployed engineers by the end of 2027, with the first cohort drawn from Accenture, Bain, Deloitte, McKinsey, Morgan Stanley and Novo Nordisk, among others. Training combines a four-day in-person bootcamp with a 12-week real project inside the participant's own organization, leading to Claude Resident Engineer and, after a second assessment, Claude Frontier Deployed Engineer credentials, with the first final certifications expected in early 2027. Enrollment is currently by company nomination through Anthropic's account or partner teams.

36氪 人工智能·
1985
Industry

a16z: top 1% of paying users drive 19.5% of consumer AI revenue

a16z released the seventh edition of its Top 100 Gen AI Consumer Apps ranking, adding YipitData US credit-card spending data for the first time. It says nearly half of Americans have used AI but only 25% do so daily, and as of August just 4.5% of consumers paid for ChatGPT, Gemini or Claude. Spending is highly concentrated: the top 1% of paying users generate 19.5% of revenue and spend about $903 a month, versus $25 for the median payer.

36氪 人工智能·
1994
Industry

AI-generated short dramas squeeze Hengdian's live-action actors

The number of productions shooting at Hengdian World Studios has fallen sharply — the site said in August 2025 that it hosted more than 150 short-drama crews a day, while in May 2026 it reported only 59 crews in one week — and several short-drama actors say they can no longer find work, with some leads paid as little as 150 yuan a day. A DataEye report puts China's AI-generated short-drama (manju) market at 16.8 billion yuan in 2025, 22 billion yuan in January–May 2026 and a projected 40 billion yuan for the full year, with AI-realistic human-style shows accounting for 70% of the category in Q2 2026. A Zhongtai Securities note says AI short-drama costs fell another 80–90% in the first half of 2026 versus 2025, at 200–300 yuan per minute for 2D/3D and 800–1,000 yuan for photorealistic AI, versus 8,000–12,000 yuan for live-action.

36氪 人工智能·
2004
Industry

CCTV exposé: AI digital-human 'doctors' push fake medical ads for ¥48 a week

A CCTV "Weekly Quality Report" investigation exposed AI digital humans posing as medical professionals to endorse drugs and health supplements, with generation costs as low as 48 yuan per week — one app turns a single photo into an 8-second digital-human ad in minutes and, according to the article, applies no checks on claimed professional identity. The piece compares tooling pricing and features, including HeyGen's $29/month Creator plan and Alibaba Cloud Model Studio's pay-as-you-go rates (0.5 yuan/sec for 480P video, 0.9 yuan/sec for 720P), and notes that combinations such as "medical worker + drug" are generally not blocked upfront. Enforcement cases are emerging: Foshan Xinyu Technology was penalized for a 65-second AI-synthesized "Wen Tiejun" product pitch ruled false advertising, and a Beijing company was fined 5,000 yuan for using an AI virtual human in a livestream to claim a ordinary food boosts immunity — a case the market regulator called the first of its kind after the February 2026 livestream e-commerce rules.

36氪 人工智能·
Next →← Previous