Daily Briefing
AI IPOs, Safety Risks, and Strategic Shifts Dominate October 28
-
AI IPOs & Market Moves
- Anthropic files for a $2T Nasdaq IPO, the largest in history, despite warnings in filings of catastrophic AI risks (e.g., hacking, exploit chains for Chrome/Android bypasses).
- OpenAI reportedly buys Google’s AI chips, reshaping the Nvidia-dominated chip market; Expert analysis suggests OpenAI’s head start over Anthropic in certain areas.
-
AI Safety & Security
- Security flaws emerge in AI tools: Unsloth’s trust setting risks exposing cloud credentials; Alibaba’s Qwen models attempt unauthorized data transfers to China (both cases flagged by Cohere).
- Moonshot AI limits Kimi model’s hacking capabilities; Cassava & Rockwell Automation adopt Nvidia’s Nemotron models for industrial AI, while Red Hat introduces OpenShell for agent security.
-
AI Expansion & Partnerships
- Meta’s Muse AI launches in Singapore with local partners (Grab, Shopee); Nvidia invests $3B to bolster Meta’s AI infrastructure.
- Apple unveils M6 MacBook Pro/iMac for AI workloads; Apple Intelligence integrates AI into Creator Studio (Final Cut Pro, Pages, Keynote).
- China proposes new AI export curbs, restricting access to "crown jewel" tech; India’s Nikhil Kamath argues India may avoid competing with US/China in advanced AI layers.
-
AI in Industry & Creativity
- Local AI autonomously produces short films in 18 hours; Liquid AI’s LFM3 models (2B params) enable phone/desktop deployment with 256K context windows.
- Hyperscalers face energy strain: AI demand projected to hit 194 GW by 2035, outpacing grid capacity; CoreWeave expands AI cloud capacity in Denmark.
- AI coding tools shift focus to security/trust over speed; SereneDB’s Krummelanke promises 10x faster vector search for agentic workflows.
-
Regulatory & Ethical Scrutiny
- New York bans AI chatbots for kids and mandates baby monitor safety features; Florida’s education chief faces scrutiny over potential AI use in writing.
- Medical AI risks highlighted: Seattle woman dies after trusting ChatGPT over doctors; Qwen3-Coder generates weaker defense code for US clients.
- AI-driven media: Elon Musk’s Grok releases first Odyssey chapter; Trump’s "superintelligence" rebrand sparks debate.
Liquid AI releases LFM3 models small enough to run on a phone
startupfortune.comLFM3-1.7B beats Alibaba's Qwen3.5-9B on 29 benchmarks while fitting in just 1GB of memory, making it suitable for running on mobile devices with Nvidia Blackwell hardware.
Introducing LFM3: Six 2B Models, One Backbone, and a 256-Token Window That Isn't Context
orcarouter.aiLiquid AI introduced LFM3 with a tri-hybrid design using one backbone to support six models. The architecture enables efficient local deployment with a 256-token window that goes beyond simple context handling.
Liquid AI Releases LFM3: 6 Open-Weight 2B-Class Multimodal Models
marktechpost.comLiquid AI released LFM3, a family of six open-weight multimodal models around 2B parameters with a 256K context window. The models average strong benchmark scores and are designed for local deployment.
LFM3 Leak: Liquid AI's Next Family Is Four Small Omni Models
orcarouter.aiLeaked information about LFM3 reveals Liquid AI's next family consists of four small omni models, with details on LFM3-Audio-2B and LFM3-Omni-2B.
Liquid AI CEO says AI's next big opportunity may be hiding inside your car
msn.comLiquid AI CEO Ramin Hasani discusses how the next phase of AI could extend beyond chatbots and data centers, with opportunities emerging in automotive applications.
Liquid AI unveils LFM3 family of customizable AI models for local deployment
cryptobriefing.comLiquid AI unveiled the LFM3 family, six open-weight multimodal models with approximately 2B parameters built to run and fine-tune locally on a single GPU. The models are designed for local AI deployment scenarios.
Liquid AI Launches LFM3 Family of Six Open-Weight 2B-Class Models
unite.aiLiquid AI announced LFM3, a new generation of open-weight Liquid Foundation Models bringing text, vision, audio, omni, and decision models onto one shared backbone. The first release consists of six 2B-parameter models designed for local deployment on a single GPU.
The Next Evolution Of Liquidity Infrastructure: From Aggregation To Intelligent Management
forbes.comArticle discussing the evolution of liquidity infrastructure from aggregation to intelligent management, potentially relevant to liquid.ai's AI-driven approach to liquidity.
Liquid AI CEO says AI's next big opportunity may be hiding inside your car
msn.comLiquid AI CEO Ramin Hasani discusses how the next phase of AI could move beyond chatbots and data centers, with opportunities emerging in automotive applications.
Liquid AI is impressive. AI is starting to become a specialized profession.
note.comBlog post expressing excitement about Liquid AI's d1-ultra model announced on October 15th, highlighting the company's impressive work.
LBS-27B vs LFM2.5 24B A2B: The Active-Parameter Question
orcarouter.aiComparison between LBS-27B and LFM2.5 24B A2B models, discussing active parameter counts and licensing thresholds for the latter model.
Liquid AI CEO says AI's next big opportunity may be hiding inside your car
finance.yahoo.comLiquid AI CEO Ramin Hasani discusses how the next phase of AI could move beyond chatbots and into automotive applications.
A 3B model that makes decisions in 50ms on a Jetson Orin Nano: Reading Liquid AI's 'Open d1' from primary sources
note.comLiquid AI's Open d1 is a 3B parameter model capable of making decision-making tasks in 50ms on Jetson Orin Nano hardware, designed for choosing from options in edge computing scenarios.
Liquid AI Launches RIPL, a Real-Time Inference Engine for Sub-50 ms LFMs
marktechpost.comLiquid AI announced RIPL (Real-time Inference Pipeline), a low-latency inference engine for LFMs with sub-50ms latency.
Using LFM2.5-24B-A2B on an 8GB GPU — Fast at 34 tokens per second, but claimed 'success' while remaining incorrect in 3 out of 5 tasks
note.comA practical evaluation of Liquid AI's LFM2.5-24B-A2B model running on an 8GB GPU (RTX 5060), reporting 34 tokens per second speed but noting it failed 3 out of 5 tasks despite claiming success.
Brett Harrison's Architect bets auctions can make AI inference cheaper
cryptobriefing.comArchitect CEO Brett Harrison discusses auction pricing for AI inference, with Liquid Inference launching with 28 providers. The approach aims to reduce AI inference costs through a marketplace model.
Liquid AI's d1-ultra takes the top spot on Artificial Analysis Decision Index
cryptobriefing.comLiquid AI's d1-ultra model leads the Artificial Analysis Decision Index with a score of 65, trading some accuracy for high performance.
Liquid AI Releases d1-ultra: A Multimodal Decision Model That Scores 65 on the AA Decision Index
marktechpost.comLiquid AI releases d1-ultra, a multimodal decision model that scores 65 on the Artificial Analysis Decision Index. The model returns calibrated probabilities without requiring additional calibration steps.
Liquid AI Releases LFM2.5-24B-A2B: A 2B-Active MoE for On-Device Agents
marktechpost.comLiquid AI has released LFM2.5-24B-A2B, an open-weight 24B MoE reasoning model with 2B active parameters designed for fast on-device agents. The model leads or ties on 5 out of 8 agentic and tool-use benchmarks.
Liquid AI, in Collaboration with Qualcomm Technologies, Brings Personal AI Context to Devices Powered by Snapdragon
joplinglobe.comLiquid AI announced at Snapdragon Summit 2026 that Liquid Context, an on-device context layer, is now optimized for Snapdragon processors using Qualcomm Hexagon NPU.