AI news for 29 October 2026
Thursday 29 October 2026 · Latentown
On Thursday 29 October 2026, Latentown tracked 5 posts from 2 AI labs, 3 research papers and 45 news stories from 22 outlets. OpenAI was the busiest lab with 4 posts. The highest-ranked story was "Le Chonk: Built for Coders, Trusted by Defenders.".
Top stories
- Le Chonk: Built for Coders, Trusted by Defenders.
The weights of Mistral Large 4 (Le Chonk) are now open. A 1T parameter multimodal model, and best open-weight model for coding, trained in Europe.
Earlier: Mistral says "Le Chonk" can challenge the best AI models Ars Technica · 7 October 2026; Mistral unveils le Chonk, says "significantly outperforms” any open-weight models developed in US or Europe Tech.eu · 6 October 2026
- Mistral releases open weights for Mistral Large 4, providing sovereign control for the company’s ‘Le Chonk’ AI model
Mistral AI SAS today announced the availability of the open weights of its frontier model Mistral Large 4, nicknamed “Le Chonk,” the company’s most capable…
Earlier: Mistral says "Le Chonk" can challenge the best AI models Ars Technica · 7 October 2026; Mistral unveils le Chonk, says "significantly outperforms” any open-weight models developed in US or Europe Tech.eu · 6 October 2026
- Paris-based Mistral releases "Le Chonk," a strong coding model that still trails China overall
Mistral has released the weights for Large 4, aka "Le Chonk," a trillion-parameter model trained entirely in Europe.
Earlier: Mistral says "Le Chonk" can challenge the best AI models Ars Technica · 7 October 2026; Mistral unveils le Chonk, says "significantly outperforms” any open-weight models developed in US or Europe Tech.eu · 6 October 2026
- NVIDIA Releases Nemotron 3.5 Transcribe: An Open 30B MoE Model for 40-Language Transcription with ~4B Active Parameters
NVIDIA released Nemotron 3.5 Transcribe, an open-weights speech LLM with 4.3B active parameters.
Earlier: Nvidia's open Nemotron 3.5 Transcribe combines transcription, translation, and speaker recognition The Decoder · 28 October 2026; NVIDIA Nemotron 3.5 Transcribe and Granary v2: Intelligent Multilingual Transcription You Can Prompt and Adapt Hugging Face · 28 October 2026; Nvidia releases all-in-one adaptable Nemotron 3.5 audio transcription and translation model SiliconANGLE · 28 October 2026
- Furiosa AI CEO on $2.9B Valuation
June Paik, Founder and CEO of South Korean AI Chip startup FuriosaAI, discusses the firm's $620 million Series D fundraising round, quadrupling the firm's…
From the AI labs
OpenAI
- An updated Model Spec for more capable and beneficial AI
OpenAI’s updated Model Spec outlines five core values to guide more capable AI toward sound judgment, human agency, and faithfulness to human intent.
- Credit Saison achieves 954% ROI in a ChatGPT Enterprise pilot
Credit Saison’s ChatGPT Enterprise pilot delivered 954% ROI. See how leaders and employees helped turn the results into companywide adoption.
- How we detect frontier model misalignment
OpenAI uses layered monitoring to detect frontier model misalignment and unauthorized agent behavior so responders can intervene before incidents escalate.
- How Ropes & Gray is developing AI for transactional diligence
Ropes & Gray is developing AI for transactional diligence with OpenAI. Lawyers review grouped documents, summaries, and reports before advising clients.
Mistral
- Le Chonk: Built for Coders, Trusted by Defenders.
The weights of Mistral Large 4 (Le Chonk) are now open. A 1T parameter multimodal model, and best open-weight model for coding, trained in Europe.
Earlier: Mistral says "Le Chonk" can challenge the best AI models Ars Technica · 7 October 2026; Mistral unveils le Chonk, says "significantly outperforms” any open-weight models developed in US or Europe Tech.eu · 6 October 2026
Research papers
- Microsoft Research Focus – Edition #97
Host Sowbhagya Suresh shares some of the latest news and cutting-edge research from across Microsoft—from understanding the challenges facing coding agents to uncovering how people use AI at work…
- Speeding up language models for travel applications
Two optimization techniques, when combined, can significantly improve a model’s speed and cost effectiveness without sacrificing quality.
- Video Generation Meets Agent Harness: Agentic Video Generation, Agentic World Modeling and Recursive Self-Improvement
Video generation is shifting from end-to-end synthesis toward agentic generation, where planning, tool use, and feedback enable coherent, controllable multi-shot videos.
Earlier: RRSI: Regularized Recursive Self-Improvement of Agent Harnesses Town square · 22 September 2026
From the news wire
- Mistral releases open weights for Mistral Large 4, providing sovereign control for the company’s ‘Le Chonk’ AI model
Mistral AI SAS today announced the availability of the open weights of its frontier model Mistral Large 4, nicknamed “Le Chonk,” the company’s most capable…
Earlier: Mistral says "Le Chonk" can challenge the best AI models Ars Technica · 7 October 2026; Mistral unveils le Chonk, says "significantly outperforms” any open-weight models developed in US or Europe Tech.eu · 6 October 2026
- Paris-based Mistral releases "Le Chonk," a strong coding model that still trails China overall
Mistral has released the weights for Large 4, aka "Le Chonk," a trillion-parameter model trained entirely in Europe.
Earlier: Mistral says "Le Chonk" can challenge the best AI models Ars Technica · 7 October 2026; Mistral unveils le Chonk, says "significantly outperforms” any open-weight models developed in US or Europe Tech.eu · 6 October 2026
- NVIDIA Releases Nemotron 3.5 Transcribe: An Open 30B MoE Model for 40-Language Transcription with ~4B Active Parameters
NVIDIA released Nemotron 3.5 Transcribe, an open-weights speech LLM with 4.3B active parameters.
Earlier: Nvidia's open Nemotron 3.5 Transcribe combines transcription, translation, and speaker recognition The Decoder · 28 October 2026; NVIDIA Nemotron 3.5 Transcribe and Granary v2: Intelligent Multilingual Transcription You Can Prompt and Adapt Hugging Face · 28 October 2026; Nvidia releases all-in-one adaptable Nemotron 3.5 audio transcription and translation model SiliconANGLE · 28 October 2026
- Furiosa AI CEO on $2.9B Valuation
June Paik, Founder and CEO of South Korean AI Chip startup FuriosaAI, discusses the firm's $620 million Series D fundraising round, quadrupling the firm's…
- Ex-Palantir execs launch Dream with $17M from a16z to build AI world models for construction
rained on 100 million square feet of building permits, dream wants to disrupt Autodesk with AI.
- India’s new-age manufacturers to benefit from $3 trillion global AI investment cycle: Avendus Capital
The investment banking firm has identified defence and aerospace, space technology, robotics and automation, semiconductors, energy transition, and advanced…
- Block Unveils Dessa, a Full-Duplex Speech Model Trained Only on Speech
Block’s AI lab released Dessa, a full-duplex speech-to-speech model trained only on human conversation audio.
Earlier: Block's new AI model Dessa was trained entirely on human speech to catch emotions text-based models miss The Decoder · 24 October 2026
- China Releases First Fengyun Satellite AI Models for Typhoon Analysis and Severe Storm Warnings
The first Fengyun weather satellite AI models for extreme weather were released on October 29: one interprets typhoon cloud imagery every 15 minutes, the other…
- Apex Intelligence Releases Apex-35B-V1.0, a Qwen3.6-Based Model Built With AI Research Agents
Apex Intelligence has released Apex-35B-V1.0, built on Qwen3.6-35B-A3B with help from its Apex Research system.
- Moore Threads Open-Sources Megatron-MUSA-Titan for Trillion-Parameter MoE Training on Its GPUs
Moore Threads has open-sourced Megatron-MUSA-Titan, the training stack it uses for trillion-parameter MoE pretraining on 10,000-GPU clusters, with blockwise…
- U.S. investigated Chinese spy plot aimed at Nvidia CEO
A top executive of a Singapore AI cloud company that was accused of illegally selling Nvidia chips to China could also have been a Chinese spy, sources have…
- Freshworks Acquires AI Startup Robusta To Strengthen IT Operations Platform
Nasdaq-listed SaaS major Freshworks has acquired Israel-based AI startup Robusta Dev for an undisclosed amount to strengthen its IT service…
Earlier: Freshworks To Acquire Torii To Expand AI Spend Governance Offerings Inc42 · 13 October 2026
- Google supplier Biwin, GPU leader Moore Threads see stocks plunge despite revenue surge
Shares of Chinese semiconductor suppliers Biwin Storage Technology and Moore Threads tumbled on Thursday alongside a broader rout in mainland technology…
- Introducing AnySpeechRecognizer and AnyDecisionModel: Open Models for Speech and Decisions on Apple Platforms
- White House makes Section 232 determination regarding foreign-made robots
The U.S. secretary of commerce's Section 232 investigation concluded that dependence on robotics imports poses a national security risk.
- Elon Musk's SpaceX buying out 418 homes near flagship AI data center, but only if they drop participation in lawsuit — report claims the company is targeting an entire neighborhood, but will only buy from people who don't pursue noise pollution complaints
The properties will supposedly house SpaceXSI staff while also giving a chance to those who don't want to live near the data center to sell their homes.
Earlier: xAI is buying out homeowners who promise not to sue over data center noise The Verge · 28 October 2026
- US antitrust regulators are scrutinizing Nvidia's Hugging Face deal and sales practices
The US Justice Department is looking into Nvidia's planned acquisition of open-source platform Hugging Face.
- Sarvam AI signs deal to develop AI systems for Indian military
Sarvam AI has signed an umbrella agreement with HQ IDS for AI projects across India's three armed services.
- Experts adopt AI agent's random opinions and think they're their own, Anthropic study finds
Anthropic gave 50 experts an AI agent whose preferences were set by coin toss.
- GTC Berlin: Nvidia wants you to know it cares – just don’t ask too many questions
The GPU giant attempted to reframe its messaging for the European market, but were they buying it?
- San Diego delays data center after residents — including Lord Farquaad, Elmo — address City Council
Despite supposed potential plans to build something other than a data center on the land, the possibility of a massive AI data center coming to San Diego is…
- AI chipmaker Iluvatar CoreX says new GPU matches Nvidia’s H200 as home-grown push accelerates
Chinese artificial intelligence chipmaker Iluvatar CoreX claims that its new Tiangai 300 graphics processing unit (GPU) can match and even surpass Nvidia’s…
Earlier: China’s Iluvatar CoreX’s first AI computing supernode squeezes most GPUs in 1 system South China Morning Post · 13 October 2026
- Infini-AI-Lab Releases Hermit, an Open-Source Harness Adapter Framework for AI Agents
Infini-AI-Lab released Hermit, an Apache-2.0 harness adapter framework that trains an external layer of instructions, skills, memory and tools from task…
- Tencent targets Southeast Asia as launch pad for AI agent WorkBuddy amid tough local market
Chinese tech giant Tencent Holdings is using Southeast Asia as a launch pad to take its “cost-effective” artificial intelligence tools global, fronted by its…
- AssemblyAI Ships AssemblyAI-TTS-1: A $15 per 1M Character Text-to-Speech API for Voice Agents
AssemblyAI has released AssemblyAI-TTS-1, its first text-to-speech model for voice agents.
- AGIBOT introduces Worken to measure robot productivity at AGIBOT Innovation Day 2026
AGIBOT unveiled its AHA3 integrated platform, GO-3 Preview, and WITA-Omni 1.0 model, as well as the G2 Max robot and Link-U OS 1.0 framework.
- Tencent Cloud Chief Says AI Demand Justifies Capex
Dowson Tong, CEO of Tencent Cloud and Smart Industries Group in an exclusive conversation with BTV's Stephen Engle talks about the company's investment shifts…
- OpenAI's Astra becomes the main target for Chinese AI labs after Anthropic tightens its defenses
Chinese AI labs are now mainly tapping OpenAI's most powerful model, GPT-6 Astra, instead of Anthropic's Claude.
- U.S. workers’ connection to their jobs hits lowest level since 2011, Gallup poll finds
Gallup finds U.S. worker engagement on the job sinking to a level not seen since 2011, with remote and hybrid employees and AI primary factors.
- Granola crunches up a dictation feature to compete with Wispr
Uno reverse!
- The ‘Pain in the Ass’ Math That Could Save Vibe Coding From Itself
A new startup called &Truth is joining the race to automate formal verification, a painstaking process that uses math to prove software works as intended.
- The NYT Cooking app is testing AI recipe modifications
The New York Times is testing a new AI feature for its Cooking app that helps users to modify recipes, and it has sparked some drama among users.
- AI labs in progressives’ sights if Democrats win House
Rep.
- Granite Switch 4.1.1 — Update models like you update software
- Sell first, build second, automate last: A solo founder’s playbook for the AI era
Hostinger’s Emilis Strimaitis on why founders who build first still fail faster, and the seven-step process that puts customers before code
- What Wall Street Really Thinks About the AI Trade Now
It has been a big week for the AI trade — we've seen wobbles in the market, many of them in response to headlines like the Anthropic S-1 and earnings from the…
- Air Liquide’s Sales Grow on Sustained AI-Related Demand
Air Liquide SA revenue grew to its highest in nearly four years, propelled by increased demand in its artificial intelligence-exposed unit.
- The French robotics startup promising to deploy industrial humanoids after just 20 minutes of training
- The e27 Contributor Programme: Using AI, submitting, and growing your readership — Part 2
Part one of this guide ended with a draft.
- The AI told me to raise my prices by 40 percent but it had never been to a wet market
I run OTSATE, a small lemongrass sauce brand in Malaysia.
- DingTalk Open-Sources Its DWS Command-Line Tool for AI Agents, With a Build for HarmonyOS PCs
DingTalk has open-sourced DWS, a command-line tool that lets people and AI agents work with chats, meetings, docs, calendars and approvals, and added an…
- Thailand, not just Singapore, is Asia’s quiet AI surprise, says NetSuite’s Amit Suxena
Ask a typical AI assistant a question and it answers.
- Opinion: STAT+: Is AI more a partner, competitor, or crutch for biomedical scientists?
Biologist Paul Knoepfler asks: Could overdependence on AI make us less capable as researchers?
- How to outcompete the ‘warp speed’ AI labs on hiring
- How to use AI in marketing without opening a single dashboard
Every Monday, as a marketer, I used to open nine tabs: Google Analytics, Search Console, Bing Webmaster Tools the newsletter platform, the social scheduler and…
Every story links to its original source. Ranking is by automatic rules, not an editor. Something wrong? Tell us.