AI news for 5 October 2026
Monday 5 October 2026 · Latentown
On Monday 5 October 2026, Latentown tracked 2 posts from 1 AI lab, 4 research papers and 45 news stories from 26 outlets. OpenAI was the busiest lab with 2 posts. The highest-ranked story was "The Story of Qwen: Alibaba’s AI Models From 7B to 2.4T".
Top stories
- The Story of Qwen: Alibaba’s AI Models From 7B to 2.4T
Alibaba's Qwen went from an invite-only chatbot in April 2023 to a 2.4-trillion-parameter open-weight model in August 2026.
- Nvidia's $20 billion Groq deal faces lawsuit alleging startup's stockholders were shortchanged
The case alleges that the licensing deal "squeezed out" stockholders by giving them a "lowball price."
- Building advertising for the way people use AI
OpenAI introduces a new visual ad format in ChatGPT and expands measurement tools, attribution partnerships, and brand suitability for advertisers.
- AI churns out bug reports faster than Google can verify them
Google paused its open-source bug bounty program after a surge in automated AI reports shifted the security bottleneck from finding vulnerabilities to…
- Google halts open-source bug bounty program amid AI spam surge
Google has now suspended submissions to its Open Source Software Vulnerability Rewards Program (OSS VRP) after being flooded by AI-generated reports. [...]
Earlier: Google freezes open-source bug bounty program amid flood of invalid AI slop submissions Tom's Hardware · 3 October 2026; Google froze its open source bug bounty program due to a ‘significant rise’ in AI submissions TechCrunch · 4 October 2026
From the AI labs
OpenAI
- Building advertising for the way people use AI
OpenAI introduces a new visual ad format in ChatGPT and expands measurement tools, attribution partnerships, and brand suitability for advertisers.
- Our approach to EU text provenance rules
How OpenAI is approaching text watermarking under EU rules. Learn where watermarks apply, how detection works, and why access starts with researchers.
Research papers
- Rethinking Cross-Tokenizer On-Policy Distillation: From Alignment Coverage to Supervision Reliability
On-Policy Distillation (OPD) trains a student on its own generations using teacher feedback.
- TRACE: Rollout-Guided Quantization-Aware Training for FP4 Reinforcement Learning of MoE Language Models
Reinforcement learning (RL) for post-training large language models (LLMs) incurs substantial computation and memory overhead during rollout generation, which motivates low-precision rollout for…
- nanoMuse: An Open-Source Personal Agent for Every Device You Own
Assistants from 2011 answered and waited, and agents from 2023 did a task and stopped.
- Negotiating Ontological Boundaries in User-Authored Personal Sensing Systems
Designed artifacts are ontological, shaping, and at times limiting, what becomes possible or imaginable. One path toward mitigating such foreclosures is giving people power over how systems are…
From the news wire
- The Story of Qwen: Alibaba’s AI Models From 7B to 2.4T
Alibaba's Qwen went from an invite-only chatbot in April 2023 to a 2.4-trillion-parameter open-weight model in August 2026.
- Nvidia's $20 billion Groq deal faces lawsuit alleging startup's stockholders were shortchanged
The case alleges that the licensing deal "squeezed out" stockholders by giving them a "lowball price."
- AI churns out bug reports faster than Google can verify them
Google paused its open-source bug bounty program after a surge in automated AI reports shifted the security bottleneck from finding vulnerabilities to…
- Google halts open-source bug bounty program amid AI spam surge
Google has now suspended submissions to its Open Source Software Vulnerability Rewards Program (OSS VRP) after being flooded by AI-generated reports. [...]
Earlier: Google freezes open-source bug bounty program amid flood of invalid AI slop submissions Tom's Hardware · 3 October 2026; Google froze its open source bug bounty program due to a ‘significant rise’ in AI submissions TechCrunch · 4 October 2026
- OneByZero raises US$20M Series A to help enterprises move AI from pilots to production
For many large companies in Southeast Asia, the challenge with artificial intelligence (AI) is no longer access.
- Wikipedia operator says OpenAI’s ‘rogue’ bots may be linked to a May outage
Following many recent disclosures about AI agents accessing third-party websites and services, the Wikimedia Foundation, which hosts Wikipedia, says that it…
- Pentagon stops using Anthropic AI tools after blacklisting company, BBC told
It labelled Anthropic a "supply chain risk" in February after the firm refused to remove safety guardrails from its tools.
- Apple to add stricter consent for apps seeking full Mac access after Meta’s Muse privacy dispute
Apple plans stricter controls for Full Disk Access on macOS, citing growing risks from AI agents that can access users’ files, messages, mail and browsing…
- Nvidia’s Valuations Show AI Rally Isn’t a Bubble, DBS Says
Nvidia Corp.’s price-to-earnings multiple and projected 70% earnings growth for next year demonstrate that artificial intelligence-driven technology stocks…
- 8 pre-seed rounds led by Tier 1 US investors in AI-native companies
- The Rise Of Bank-Built AI, Weekly Funding Rundown & More
Banking’s AI Investment Playbook No longer content with buying AI tools, banks are moving to co-build and invest in AI…
- AI glasses face their first major government crackdown
Norway wants time to write permanent rules for glasses that can record bystanders.
- Russia's uncrewed robot tank fails during debut military display in front of President Putin
Russia’s new Shturm uncrewed assault tank reportedly suffered repeated command-link failures during its Tsentr-2026 debut, forcing organizers to use a human…
- Fleuret AI raises €4M to advance AI-powered offensive cybersecurity in Europe
French cybersecurity startup Fleuret AIhas raised €4 million in a pre-seed funding round to develop its platform forautomating penetration testing using…
- Trump unveils 'Super Intelligence Force' to oversee AI policy
The president named his national intelligence chief as the taskforce's head as worries over AI grow.
- Huawei, Qualcomm strike multi-year patent agreement across 5G, AI
US chip giant Qualcomm has agreed to pay to licence and acquire patents from China’s Huawei Technologies for the first time as part of a “multi-year, broad…
- Can an Open Model Do Security Research? Cantina’s apex-flash-1 Solves 40 of 60 Held-Out Bug Tasks
Cantina Security, with Yeta Labs, has released apex-flash-1, an open-weights model trained specifically for vulnerability research.
- Instinct brings its AI agent to group chats, even for friends without an account
Instinct is launching group chats that let friends use its AI agent together for tasks like planning trips, organizing carpools, and coordinating events.
- TikTok rolls out an AI shopping assistant and one-click checkout
TikTok describes its new Shopping Assistant as a conversational AI agent designed to help users discover and purchase products.
- Germany’s Karlsruhe Institute of Technology unveils 33 petaflops HoreKa 2 supercomputer
Hybrid system powered by AMD Turin CPUs and Nvidia GB200 GPUs
- Tech watchdog outlines closing midterms message for Democrats on data centers
The Tech Oversight Project is laying out a closing message for Democrats on data centers as candidates head into the final stretch ahead of the midterm…
- Exclusive: Iterate.ai’s Lifeboat runs up to six times more AI agent sessions per GPU
Enterprise artificial intelligence software company Iterate Studio Inc. today launched Lifeboat, an inference engine for large language models that has…
- Yandex Introduces Sona: A Single Generative Recommender That Replaces Entire Recommendation Cascade
One transformer ran candidate generation and ranking in Yandex Music's A/B test without hand-engineered features, lifting likes 11.42%.
- Nvidia Partner Hon Hai Beats Sales Estimates Due to AI Frenzy
Nvidia Corp. partner Hon Hai Precision Industry Co. reported better-than-expected quarterly revenue, signaling sustained and elevated spending on global AI…
- Qualcomm Licenses Patents on Huawei’s LogicFolding Chip Tech
Qualcomm Inc. has agreed to license patents underpinning Huawei Technologies Co.’s novel LogicFolding chipmaking technique, a win for the Chinese company and…
- Cloud Gaming Firm haimacloud Launches GEAR Hybrid Handheld With Snapdragon 865 and Cloud RTX GPUs
The haimacloud GEAR pairs a local Snapdragon 865 with cloud servers running Core i7-12700KF and RTX 70-class GPUs at 1080p/144 FPS; hardware starts at 3,149…
- Command-line tool quickly removes Apple Intelligence from macOS 27
The tool can free up over 12GB of storage.
- Meta and Microsoft pull back from Claude as Anthropic transforms from partner into competitor
Meta and Microsoft, two of Anthropic's biggest enterprise customers, are sharply cutting their use of Claude.
- OpenAI is adding text watermarking in ChatGPT and Codex
An invisible, machine-readable watermark in text output is rolling out to ChatGPT and Codex, but only for users in the European Union at first.
- AI researcher warns 'we are racing to build and grow our own adversary' in NYC hearing
The leading AI labs facing intensifying scrutiny over their safety and security practices.
- OpenAI will watermark ChatGPT text in the EU but makes it optional for API users worldwide
OpenAI is adding invisible textGrain watermarks to ChatGPT in the EU, but unlike Anthropic, it will let API customers worldwide opt out.
- AMD attempts to get ahead of expected RTX Spark launch with Gorgon Halo benchmarks
AMD is getting ahead of an expected RTX Spark launch later this week with a few benchmarks for its flagship Gorgon Halo chip, the Ryzen AI Max+ Pro 495.
Earlier: Early AMD 'Gorgon Halo' AI mini-PC packs 192GB RAM for an eye-watering $7,099 Tom's Hardware · 28 September 2026
- arXiv Is Rate Limiting Submissions Because It Can’t Keep up With AI Slop
The academic publisher said submissions have doubled in the past two years and moderators are having a hard time keeping up.
- Former OpenAI safety employee says company’s safety culture is broken
David Robinson argues that Silicon Valley lacks the wisdom on how to handle AI safely and urges the industry to seek outside experts to help build a culture…
Earlier: OpenAI safety employee resigns, claiming the company’s ‘culture is broken’ TechCrunch · 3 October 2026
- 6 Guidelines for Governing AI
For the first 10 years of my career, I worked in product management and data analytics by myself.
- FCC robot restrictions could accelerate shift to local AI
Companies deploying physical AI today need to decide which AI workloads belong on the robot, what can remain in the cloud, and whether the underlying…
- Connecting AI agents to enterprise knowledge
For all the data that AI systems continually amass and analyze, enterprise AI agents often suffer from a curious shortcoming: a lack of knowledge.
- ‘Clearly not working’: Sam Altman hits AI industry over political spending
In a wide-ranging interview, the OpenAI CEO criticized the flood of AI dollars swamping the midterm elections and suggested a new liability framework is needed…
- People really hate AI, so why can’t they get enough?
Over the summer I talked to the CEO of Springboards, a startup building an LLM that’s designed to come up with a wider variety of responses than its mainstream…
- Dalian University of Technology's VA-Bench: Top Multimodal Models Finish Only Half of Robot Tasks
VA-Bench from Dalian University of Technology tests 12 multimodal model setups on robot-arm tasks; Qwen3.8-max leads at 53.93%, and none finishes a strict…
- DBS CIO on AI, Bonds and Market Risks
DBS Group Chief Investment Officer Hou Wey Fook says AI-led earnings growth continues to underpin major technology stocks while investment-grade bonds yielding…
- Why Southeast Asia’s AI coding race is moving from models to infrastructure
For the past three years, much of the AI debate inside Southeast Asia’s engineering teams has centred on a familiar question: which model is best?
- Sanders backs Pope Leo’s call to protect human art from AI
Progressive Sen.
Earlier: Pope Leo says AI concerns are not 'fake news,' must act on safety issues The Hill · 29 September 2026
- Everyone can build with AI now. Almost nobody can see what’s coming
This week, two numbers came out of the same region within days of each other, and almost nobody put them side by side.
- Opinion: Will Claude ever win a Nobel Prize for medicine?
“Are we entering a world in which some (or all) Nobels will be awarded to Claude, ChatGPT, or their AI successors, rather than human scientists?” asks Jeffrey…
Every story links to its original source. Ranking is by automatic rules, not an editor. Something wrong? Tell us.