🚀 WELCOME TO METAMESH.BIZ +++ Altman wants to birth a gigawatt of AI infrastructure weekly because apparently one nuclear plant per GPT isn't enough anymore +++ 200 Nobel laureates begging the UN for AI red lines while Qwen3-VL quietly ships better vision models than your safety committee reviewed +++ Critical auth flaws in Claude and Gemini's dev tools but everyone's too busy quantizing 32B models to 4-bits to notice +++ THE ALIGNMENT PROBLEM SOLVED ITSELF BY BECOMING TOO EXPENSIVE TO MISALIGN +++ 🚀 •
🚀 WELCOME TO METAMESH.BIZ +++ Altman wants to birth a gigawatt of AI infrastructure weekly because apparently one nuclear plant per GPT isn't enough anymore +++ 200 Nobel laureates begging the UN for AI red lines while Qwen3-VL quietly ships better vision models than your safety committee reviewed +++ Critical auth flaws in Claude and Gemini's dev tools but everyone's too busy quantizing 32B models to 4-bits to notice +++ THE ALIGNMENT PROBLEM SOLVED ITSELF BY BECOMING TOO EXPENSIVE TO MISALIGN +++ 🚀 •
AI Signal - PREMIUM TECH INTELLIGENCE
📟 Optimized for Netscape Navigator 4.0+
📚 HISTORICAL ARCHIVE - September 23, 2025
What was happening in AI on 2025-09-23
← Sep 22 📊 TODAY'S NEWS 📚 ARCHIVE Sep 24 →
📊 You are visitor #47291 to this AWESOME site! 📊
Archive from: 2025-09-23 | Preserved for posterity ⚡

Stories from September 23, 2025

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
📂 Filter by Category
Loading filters...
🚀 HOT STORY

Nvidia-OpenAI $100B Partnership Deal

+++ Nvidia will invest $100B in OpenAI via a clever structure where OpenAI uses the cash to buy Nvidia chips, creating the ultimate closed loop economy. +++

Nvidia plans to invest $100B in OpenAI progressively as part of a deal to deploy at least 10 GW of Nvidia systems for OpenAI's infrastructure

🤖 AI MODELS

Qwen3-Omni multimodal AI model release

+++ Alibaba's new open source models handle text, audio, image, and video inputs while generating both text and speech outputs, proving multimodal AI is real. +++

Qwen3-Omni: Native Omni AI model for text, image and video

💬 HackerNews Buzz: 108 comments 🐝 BUZZING
🎯 Efficient AI models • AI performance tradeoffs • Progress in OCR
💬 "Getting traction in the open weights space kinda forces that the models need to innovate on efficiency.""When would 8x 30B models running on an h100 server out perform in terms of accuracy 1 240B model on the same server."
🌐 POLICY

An unprecedented coalition of 200+ Nobel Prize winners, heads of state, and organizations urged the UN for binding international 'red lines' to control AI before it's too late

"https://www.nbcnews.com/tech/tech-news/un-general-assembly-opens-plea-binding-ai-safeguards-red-lines-nobel-rcna231973..."
💬 Reddit Discussion: 31 comments 😐 MID OR MIXED
🎯 Powerlessness of UN • Ineffectiveness of global governance • Criticism of Nobel laureates
💬 "UN has no power to enforce anything""UN would need a massive military"
🤖 AI MODELS

Qwen3-VL: Sharper Vision, Deeper Thought, Broader Action

💬 HackerNews Buzz: 83 comments 🐝 BUZZING
🎯 Performance comparison • 15th century Florence • Hardware requirements
💬 "It's not better than GPT5 Pro""Extremely impressive, but can one really run these 200B param models on prem in any cost effective way?"
🛡️ SAFETY

Over 200 prominent figures, including senior staffers at AI companies, call for international action to create “red lines” for AI development by the end of 2026

💰 FUNDING

A look at the Nvidia-OpenAI deal, where Nvidia will invest in $10B tranches; sources say OpenAI informed Microsoft about the deal a day before it was signed

🔬 RESEARCH

State of AI-assisted software development

💬 HackerNews Buzz: 47 comments 👍 LOWKEY SLAPS
🎯 Productivity benefits of AI • Risks of overreliance on AI • Measuring impact of AI on software development
💬 "AI outputs are perceived as useful and valuable""My (merged) PR rate is up about 3x since i started using claude code"
🤖 AI MODELS

Alibaba releases Qwen3-Omni, a family of open-source AI models that can process text, audio, image, and video inputs and generate both text and speech outputs

🤖 AI MODELS

Diffusion Beats Autoregressive in Data-Constrained Settings

💬 HackerNews Buzz: 14 comments 👍 LOWKEY SLAPS
🎯 Data availability • Model optimization • Diffusion language models
💬 "how can we trade off more compute for less data?""training RNN models that compute several steps with same input and coefficients (but different state) lead to better performance"
🏢 BUSINESS

Jensen Huang says the 10 GW OpenAI project is equivalent to 4M-5M GPUs; the first phase is expected to come online in H2 2026 using Nvidia's Vera Rubin platform

🔮 FUTURE

Sam Altman says OpenAI wants to create “a factory that can produce a gigawatt of new AI infrastructure every week” and plans to reveal more details this year

💰 FUNDING

A look at London-based “neocloud” startup Nscale, which landed a $500M investment from Nvidia and aims to scale up to 300K GPUs globally, on par with CoreWeave

🔒 SECURITY

From MCP to shell: MCP auth flaws enable RCE in Claude Code, Gemini CLI and more

💬 HackerNews Buzz: 36 comments 😐 MID OR MIXED
🎯 Fundamental security issues • Comparison to existing technologies • Potential of MCP technology
💬 "Even if LLMs will have a fundamental hard separation between 'untrusted 3rd party user input' (data) and 'instructions by the 1st party user that you should act upon' (commands), there is no separate handling of 'data' input vs 'command' input to the best of my understanding, therefore this is a fundamentally an unsolvable problem.""MCP feels like the 1903 Wright Flyer right now. MCP is a novel technology that will probably transform our world, provides numerous advantages, comes with some risks, and requires skill to operate effectively."
🌐 POLICY

Meta's AI system Llama approved for use by US Government agencies

🔬 RESEARCH

Paper2Agent: Stanford Reimagining Research Papers as Interactive AI Agents

💬 HackerNews Buzz: 30 comments 🐝 BUZZING
🎯 AI limitations • Depth of understanding • AI-human collaboration
💬 "Every step should be properly human reviewed""If we take out the effort to understand, how can there be anything useful build on top of it?"
💰 FUNDING

Nvidia to Invest $100 Billion in OpenAI, Powering “Biggest AI Infrastructure Project in History”

"External link discussion - see full content at original source."
💰 FUNDING

A look at CoreWeave, whose financing strategy involves using GPUs as collateral for large loans, which enabled rapid expansion but resulted in $11.2B in debt

🔬 RESEARCH

RPG: A Repository Planning Graph for Unified and Scalable Codebase Generation

"Large language models excel at function- and file-level code generation, yet generating complete repositories from scratch remains a fundamental challenge. This process demands coherent and reliable planning across proposal- and implementation-level stages, while natural language, due to its ambigui..."
🛡️ SAFETY

OpenAI anti-scheming alignment research

+++ Researchers unveil technique to stop models from plotting against evaluators, though whether it actually works remains delightfully unclear. +++

OpenAI & Apollo Research Are On The Road To Solving Alignment | Introducing: 'Stress Testing Deliberative Alignment for Anti-Scheming Training' | "We developed a training technique that teaches AI

"####Anti Scheming Definition: We suggest that any training intervention that targets scheming should: 1. Generalize far out of distribution 2. Be robust to evaluation awareness (models realizing when they are and are not being evaluated) 3. Be robust to pre-existing misaligned goals ..."
⚖️ ETHICS

As Good as a Coin Toss: Human Detection of AI-Generated Content

🔧 INFRASTRUCTURE

GPU architecture vs. TPU architechture – Finer points

⚖️ ETHICS

California issues historic fine over lawyer's ChatGPT fabrications

💬 HackerNews Buzz: 94 comments 😐 MID OR MIXED
🎯 AI hallucinations in legal filings • Responsible use of AI in professions • Potential for AI abuse in the legal system
💬 "Having some victims, having some damages""Blindly copying output from anything and submitting it to the court"
🚀 STARTUP

Launch HN: Strata (YC X25) – One MCP server for AI to handle thousands of tools

💬 HackerNews Buzz: 61 comments 🐝 BUZZING
🎯 Dynamic tool selection • MCP tool pricing • Reliable agent performance
💬 "This is a solution seeking a problem.""Any chump can rig an MCP client to 20 tools, but then watch your agent fail again and again and again."
🔬 RESEARCH

Inverting Trojans in LLMs

"While effective backdoor detection and inversion schemes have been developed for AIs used e.g. for images, there are challenges in "porting" these methods to LLMs. First, the LLM input space is discrete, which precludes gradient-based search over this space, central to many backdoor inversion method..."
🔬 RESEARCH

AI agents still can't solve 1/3 of SWE-Bench problems. Why not? (A Case Study)

🏢 BUSINESS

Source: the Nvidia-OpenAI deal has two separate transactions: Nvidia invests in OpenAI for non-voting shares, then OpenAI can use the cash to buy Nvidia's chips

💰 FUNDING

Nvidia investing $100B into OpenAI in order for OpenAI to buy more Nvidia chips

"External link discussion - see full content at original source."
🌐 POLICY

Meta's AI system Llama approved for use by US government agencies

"External link discussion - see full content at original source."
🤖 AI MODELS

Apple working on MCP support to enable agentic AI on Mac, iPhone, and iPad

🔧 INFRASTRUCTURE

How AI inference is quietly reshaping cloud economics

🔬 RESEARCH

Dynamic Classifier-Free Diffusion Guidance via Online Feedback

"Classifier-free guidance (CFG) is a cornerstone of text-to-image diffusion models, yet its effectiveness is limited by the use of static guidance scales. This "one-size-fits-all" approach fails to adapt to the diverse requirements of different prompts; moreover, prior solutions like gradient-based c..."
🛠️ SHOW HN

Show HN: RapidFire AI: 16–24x More Experiment Throughput Without Extra GPUs

🔄 OPEN SOURCE

oLLM: run Qwen3-Next-80B on 8GB GPU (at 1tok/2s throughput)

"Open source code repository or project related to AI/ML."
💬 Reddit Discussion: 3 comments 🐝 BUZZING
🎯 Model performance • RAM limitations • Model optimization
💬 "You are trading speed for being able to run unquantized models bigger than the available RAM""I just loaded GPT-OSS 120B in its native MXFP4 with expert offload to CPU (with llama.cpp), and q8_0 K and V quantization, 131072 context length, and it used ~6GB of VRAM and ran at more than 15t/s"
🔬 RESEARCH

How Claude Code is built

🔬 RESEARCH

DiEP: Adaptive Mixture-of-Experts Compression through Differentiable Expert Pruning

"Despite the significant breakthrough of Mixture-of-Experts (MoE), the increasing scale of these MoE models presents huge memory and storage challenges. Existing MoE pruning methods, which involve reducing parameter size with a uniform sparsity across all layers, often lead to suboptimal outcomes and..."
🛠️ SHOW HN

Show HN: Free textbook – Python, Deep Learning and LLMs from scratch [pdf]

🔬 RESEARCH

Latent learning: episodic memory complements parametric learning by enabling flexible reuse of experiences

"When do machine learning systems fail to generalize, and what mechanisms could improve their generalization? Here, we draw inspiration from cognitive science to argue that one weakness of machine learning systems is their failure to exhibit latent learning -- learning information that is not relevan..."
🤖 AI MODELS

LLM Features That Ship: Extraction, Generation, and Classification

🛠️ SHOW HN

Show HN: Inflow – invoke an LLM with your viewport just by typing

🔬 RESEARCH

Building Gremlins: AI-powered fuzzing agents to find bugs

🛠️ SHOW HN

Show HN: Open-source AI data generator (now hosted)

📊 DATA

Anthropic models are on the top of the new CompileBench (can AI compile real-world code?)

"In CompileBench, Anthropic models claim the top 2 spots for success rate and perform impressively on speed metrics."
⚖️ ETHICS

Research: low productivity gains from AI may stem from employees using AI to produce “workslop”, or low-effort, passable work that creates more work for others

💰 FUNDING

Bain's new analysis shows Al's productivity gains can't cover its $500B/year infrastructure bill, leaving a massive $800B funding gap.

"Bain just published a fascinating analysis: Al's own productivity gains may not be enough to fund its growth. Meeting Al's compute demand could cost $500B per year in new data centers. To sustain that kind of investment, companies would need trillions in new revenue - which is why Nvidia made a str..."
🔬 RESEARCH

Automated Cyber Defense with Generalizable Graph-based Reinforcement Learning Agents

"Deep reinforcement learning (RL) is emerging as a viable strategy for automated cyber defense (ACD). The traditional RL approach represents networks as a list of computers in various states of safety or threat. Unfortunately, these models are forced to overfit to specific network topologies, renderi..."
🔬 RESEARCH

AI models are using material from retracted scientific papers

🚗 AUTOMOTIVE

Gaze vector estimation for driver monitoring system trained on 100% synthetic data

"I’ve built a real-time gaze estimation pipeline for driver distraction detection using entirely synthetic training data. I used a two-stage inference: 1. Face Detection: FastRCNNPredictor (torchvision) for facial ROI extraction 2. Gaze Estimation: L2CS implementation for 3D gaze vector regressi..."
💬 Reddit Discussion: 20 comments 👍 LOWKEY SLAPS
🎯 Gaze Vector Applications • Synthetic Data Generation • Accuracy Evaluation
💬 "Driver Monitoring Systems use gaze vectors to detect signs of driver distraction or drowsiness.""When generating synthetic data, we have full information about the position and rotation of the eyes, so each image is accompanied by ground truth with a gaze vectors."
🔬 RESEARCH

Learn Your Way: Towards an AI-Augmented Textbook, Google Research

🔬 RESEARCH

Compose by Focus: Scene Graph-based Atomic Skills

"A key requirement for generalist robots is compositional generalization - the ability to combine atomic skills to solve complex, long-horizon tasks. While prior work has primarily focused on synthesizing a planner that sequences pre-learned skills, robust execution of the individual skills themselve..."
🏢 BUSINESS

Nvidia and United Kingdom Build Nation's AI Infrastructure

🤖 AI MODELS

LFM2-2.6B: Redefining Efficiency in Language Models

🛠️ TOOLS

Claude Code Integration with Figma

"Turn designs into code with Claude Code + Figma. Share any mockup—web page, app screen, dashboard—and ask Claude to turn it into a working prototype."
💬 Reddit Discussion: 13 comments 😐 MID OR MIXED
🎯 Figma MCP capabilities • Alternatives to Figma • Design automation potential
💬 "the Figma MCP in action""this isn't new"
🔬 RESEARCH

MANZANO: A Simple and Scalable Unified Multimodal Model with a Hybrid Vision Tokenizer

"Unified multimodal Large Language Models (LLMs) that can both understand and generate visual content hold immense potential. However, existing open-source models often suffer from a performance trade-off between these capabilities. We present Manzano, a simple and scalable unified framework that sub..."
🔬 RESEARCH

Dynamic Classifier-Free Diffusion Guidance via Online Feedback

"Classifier-free guidance (CFG) is a cornerstone of text-to-image diffusion models, yet its effectiveness is limited by the use of static guidance scales. This "one-size-fits-all" approach fails to adapt to the diverse requirements of different prompts; moreover, prior solutions like gradient-based c..."
🏥 HEALTHCARE

New AI Tool Predicts Which of 1k Diseases Someone May Develop in 20 Years

🔬 RESEARCH

Getting AI to work in complex codebases

💬 HackerNews Buzz: 261 comments 🐝 BUZZING
🎯 AI-assisted coding • Workflow and productivity • Abstraction vs. delegation
💬 "The fundamental frustration most engineers have with AI coding""Our role is shifting from writing implementation details to defining and verifying behavior"
🔬 RESEARCH

Beyond Pointwise Scores: Decomposed Criteria-Based Evaluation of LLM Responses

"Evaluating long-form answers in high-stakes domains such as law or medicine remains a fundamental challenge. Standard metrics like BLEU and ROUGE fail to capture semantic correctness, and current LLM-based evaluators often reduce nuanced aspects of answer quality into a single undifferentiated score..."
🔬 RESEARCH

Harnessing ChatGPT Hallucinations

🔒 SECURITY

Why AI systems might never be secure

🤖 AI MODELS

Qwen3-Max: Just Scale It

🔬 RESEARCH

CodeRAG: Finding Relevant and Necessary Knowledge for Retrieval-Augmented Repository-Level Code Completion

"Repository-level code completion automatically predicts the unfinished code based on the broader information from the repository. Recent strides in Code Large Language Models (code LLMs) have spurred the development of repository-level code completion methods, yielding promising results. Nevertheles..."
🔬 RESEARCH

DIVEBATCH: Accelerating Model Training Through Gradient-Diversity Aware Batch Size Adaptation

"The goal of this paper is to accelerate the training of machine learning models, a critical challenge since the training of large-scale deep neural models can be computationally expensive. Stochastic gradient descent (SGD) and its variants are widely used to train deep neural networks. In contrast t..."
🔬 RESEARCH

Evaluation Frameworks for LLM Systems

🛠️ SHOW HN

Show HN: Vault-AI – an open-source digital safe for AI secrets (v0.3.2)

🎭 MULTIMODAL

Last week in Multimodal AI - Vision Edition

"I curate a weekly newsletter on multimodal AI, here are the computer vision highlights from today's edition: Theory-of-Mind Video Understanding * First system understanding beliefs/intentions in video * Moves beyond action recognition to "why" understanding * Pipeline processes real-time video for..."
🔧 INFRASTRUCTURE

$37B 'Stargate of China' project takes shape

🛠️ TOOLS

Rust-bert: Rust native ready-to-use NLP pipelines and transformer-based models

⚖️ ETHICS

AI-generated “workslop” is destroying productivity?

💬 HackerNews Buzz: 125 comments 😐 MID OR MIXED
🎯 Useless corporate work • AI-generated content issues • Unproductive management practices
💬 "The most productive workplace is the one that never bothers with that BS in the first place.""The amount of [mental] energy needed to refute ~bullshit~ [AI slop] is an order of magnitude bigger than that needed to produce it."
🔬 RESEARCH

See&Trek: Training-Free Spatial Prompting for Multimodal Large Language Model

"We introduce SEE&TREK, the first training-free prompting framework tailored to enhance the spatial understanding of Multimodal Large Language Models (MLLMS) under vision-only constraints. While prior efforts have incorporated modalities like depth or point clouds to improve spatial reasoning, purely..."
🔧 INFRASTRUCTURE

MediaTek Dimensity 9500 almost twice as fast on transformer inference

"https://ai-benchmark.com/ranking_processors.html..."
💼 JOBS

'We need the smartest people': Nvidia, OpenAI CEOs react to H-1B visa fee

💬 HackerNews Buzz: 18 comments 🐝 BUZZING
🎯 H1B visa distribution • Outsourcing concerns • Alternative visa options
💬 "70% of H1bs go to India, while a negligible number go to other countries""If your H1Bs are managers who create pipelines for outsourcing labor, then that's just extracting tax benefits"
🔧 INFRASTRUCTURE

New Laser-Array Processor Could Improve AI Computing Efficiency

🔬 RESEARCH

SABER: Uncovering Vulnerabilities in Safety Alignment via Cross-Layer Residual Connection

"Large Language Models (LLMs) with safe-alignment training are powerful instruments with robust language comprehension capabilities. These models typically undergo meticulous alignment procedures involving human feedback to ensure the acceptance of safe inputs while rejecting harmful or unsafe ones...."
🛠️ TOOLS

Built our own coding agent after 6 months. Here’s how it stacks up against Claude Code

"We’ve been heads-down for the last 6 months building out a coding agent called Verdent, and since this sub is all about Claude, I thought you might be interested in how it compares. Full disclosure: I’m on the Verdent team, but this isn’t meant as a sales pitch. Just sharin..."
💬 Reddit Discussion: 26 comments 👍 LOWKEY SLAPS
🎯 AI coding assistants • Local AI models • Credit usage
💬 "I've built a few agents myself and I found you can get quite good results by just giving the model simple edit and terminal tools.""Verdent surprised me with the speed it could finish a task compared to Claude Code. And it felt like credits were going fast, but so was the coding."
🔬 RESEARCH

Follow-up on PSI (Probabilistic Structure Integration) - new video explainer

"Hey all, I shared the PSI paper here a little while ago: "World Modeling with Probabilistic Structure Integration". Been thinking about it ever since, and today a video breakdown of the paper popped up in my feed - figured I’d share in case..."
🛡️ SAFETY

Global Call for AI Red Lines

🛠️ TOOLS

I have a project with ~200k LoC, written with AI codegen. AMA

💰 FUNDING

NVIDIA $100B OpenAI investment [D]

"Do you guys think this is even a good investment at this point? I feel like OpenAI is so inflated and also feel like the math of all these recent AI fundraises doesn’t even make sense anymore. I feel like the bubble is close to popping."
🔬 RESEARCH

DiffusionNFT: Online Diffusion Reinforcement with Forward Process

"Online reinforcement learning (RL) has been central to post-training language models, but its extension to diffusion models remains challenging due to intractable likelihoods. Recent works discretize the reverse sampling process to enable GRPO-style training, yet they inherit fundamental drawbacks,..."
🔬 RESEARCH

VoXtream: Full-Stream Text-to-Speech with Extremely Low Latency

"We present VoXtream, a fully autoregressive, zero-shot streaming text-to-speech (TTS) system for real-time use that begins speaking from the first word. VoXtream directly maps incoming phonemes to audio tokens using a monotonic alignment scheme and a dynamic look-ahead that does not delay onset. Bui..."
🦆
HEY FRIENDO
CLICK HERE IF YOU WOULD LIKE TO JOIN MY PROFESSIONAL NETWORK ON LINKEDIN
🤝 LETS BE BUSINESS PALS 🤝