LLMs

Twitter

Facebook

Copy Link

Key Facts

Year2018

Key PlayersOpenAI, Google, Anthropic

Top News

Companies flood Reddit with fake posts to manipulate ChatGPT and Google AI search results

Reddit moderators are battling a new form of spam designed to manipulate AI chatbots. Companies selling peptides and hormone replacement therapy have been flooding the biohackers subreddit with strategically crafted posts intended to shape how AI systems like ChatGPT and Google AI search answer questions. The tactic, called Answer Engine Optimization, exploits Reddit's role as a primary source for large language models.

Technology3 sources

Companies flood Reddit with fake posts to manipulate ChatGPT and Google AI search results

Technology3 sources

AI outperforms law professors 75% of the time in Stanford Law School study on legal reasoning

A Stanford Law School study reveals that AI-generated answers to contract law questions were preferred over responses from law professors in 75% of blind comparisons. The research, involving 16 professors from 14 U.S. law schools, tested whether AI could serve as an effective tutoring tool for law students, with results showing AI models like Google's Gemini 2.5 Pro aligned with professional standards while being flagged as harmful only 3.5% of the time compared to 12% for human answers.

Science3 sources

AI outperforms law professors 75% of the time in Stanford Law School study on legal reasoning

Science3 sources

Phison aiDAPTIV and Intel tackle local AI's memory problem with 26B-parameter models on 16 GB RAM

Phison unveiled aiDAPTIV at Computex 2026, a memory extension technology that runs 26-billion-parameter models on PCs with just 16 GB RAM. Developed with Intel, the system uses high-performance NAND flash storage as an AI cache to bypass memory and storage limitations. But the specialized Pascari SSDs cost $2,516 for 1TB, raising questions about adoption and echoing Intel's past struggles with proprietary technologies like Optane.

Technology2 sources

Phison aiDAPTIV and Intel tackle local AI's memory problem with 26B-parameter models on 16 GB RAM

Technology2 sources

MiniMax M3 beats GPT-5.5 and Gemini 3.1 Pro on benchmarks for 5-10% of the cost

Chinese AI startup MiniMax launched its M3 large language model, delivering frontier-tier performance that eclipses GPT-5.5 and Gemini 3.1 Pro on key benchmarks while costing just 5-10% as much. The model features a 1-million-token context window, native multimodality, and will be released as open-source with full weights within 10 days, fundamentally challenging the cost-performance trade-off in enterprise AI.

Technology2 sources

MiniMax M3 beats GPT-5.5 and Gemini 3.1 Pro on benchmarks for 5-10% of the cost

Technology2 sources

AI solves 80-year math problem, mathematicians push back with new declaration on AI guardrails

OpenAI's AI model disproved the famous Erdős unit distance conjecture that stumped mathematicians for eight decades. But the breakthrough sparked concern across the field. In response, 16 experts released the Leiden Declaration on Artificial Intelligence and Mathematics, calling for transparency, proper attribution, and guardrails to protect research integrity as AI reshapes their discipline.

Science10 sources

AI solves 80-year math problem, mathematicians push back with new declaration on AI guardrails

Science10 sources

Developers ditch cloud AI for local LLM setups running on low-power hardware

Tech enthusiasts are replacing expensive cloud-based AI coding tools like Cursor with local LLM configurations powered by Ollama and Qwen models. These setups run on surprisingly modest hardware, including 15W CPUs and consumer GPUs, delivering privacy, cost savings, and offline access. From coding assistants to smart home control, local AI is proving capable enough to challenge cloud alternatives.

Technology8 sources

Developers ditch cloud AI for local LLM setups running on low-power hardware

Technology8 sources

OpenAI retires GPT-4.5 and o3, closing the chapter on ChatGPT models that sparked the AI boom

OpenAI quietly announced it's retiring GPT-4.5 on June 27 and OpenAI o3 on August 26, marking the end of the GPT-4 family that transformed AI from curiosity into mainstream tool. Despite newer models offering better benchmarks, paid ChatGPT subscribers are expressing nostalgia for the personalities of these older models, highlighting an unexpected emotional connection users have formed with specific AI versions.

Technology2 sources

OpenAI retires GPT-4.5 and o3, closing the chapter on ChatGPT models that sparked the AI boom

Technology2 sources

Russia-Linked GREYVIBE Uses ChatGPT and Gemini to Launch AI-Powered Cyberattacks on Ukraine

A Russia-linked threat group called GREYVIBE has been using ChatGPT, Google Gemini, and Ideogram AI throughout cyberattacks targeting Ukrainian military and government entities since August 2025. WithSecure researchers found the group integrated AI tools across malware development, infrastructure setup, and lure creation—but operational security mistakes exposed their backend systems.

Technology3 sources

Russia-Linked GREYVIBE Uses ChatGPT and Gemini to Launch AI-Powered Cyberattacks on Ukraine

Technology3 sources

Drexel University study reveals bond paradox in AI mental health tools used by millions

A Drexel University study analyzing over 4 million Reddit posts found that while nearly half of U.S. adults use AI chatbots for mental health support, most view them as supplements to human therapy. The research uncovered a critical bond paradox: AI tools help with practical tasks, but deep emotional bonds without clear goals correlate with dependence, worsening symptoms, and difficulty disengaging.

Health2 sources

Drexel University study reveals bond paradox in AI mental health tools used by millions

Health2 sources

CoreWeave launches platform letting AI agents learn and improve autonomously in production

CoreWeave unveiled a platform that enables enterprises to deploy AI agents capable of autonomous improvement using real-world data. The solution combines serverless reinforcement learning with production inference to eliminate slow testing cycles. By allowing agents to adapt in the field rather than through iterative evaluations, CoreWeave reduces costs by over 40% and accelerates training by 1.4 times.

Technology2 sources

CoreWeave launches platform letting AI agents learn and improve autonomously in production

Technology2 sources

Anthropic launches Claude Opus 4.8 as Mythos-class AI model nears public release

Anthropic released Claude Opus 4.8 with significant improvements in honesty and coding capabilities, just over a month after version 4.7. The AI model is 4x less likely to make unsupported claims and shows notable gains in agentic coding. Meanwhile, the company plans to release its powerful Mythos-class AI model to all customers within weeks after restricting access through Project Glasswing.

Technology28 sources

Anthropic launches Claude Opus 4.8 as Mythos-class AI model nears public release

Technology28 sources

Anthropic opens Milan office, names Generali, Pirelli and Enel as Italian customers

Anthropic has formally opened its sixth European office in Milan, announcing a roster of Italian enterprise customers including Generali, Pirelli, and Enel. The AI startup plans to triple its international workforce as it accelerates European expansion, with the London office alone set to reach 800 employees.

Business2 sources

Anthropic opens Milan office, names Generali, Pirelli and Enel as Italian customers

Business2 sources

Fujitsu unveils AI agent technology that evolves by learning from business operations

Fujitsu has developed self-evolving multi-AI agent technology that continuously learns from daily operations, human feedback, and policy changes. The Adaptive AI Agent Framework achieved a 28-point accuracy improvement across multiple domains including healthcare and finance. This technology enables AI agents to autonomously improve without constant expert intervention, addressing personnel shortages and knowledge succession challenges.

Technology2 sources

Fujitsu unveils AI agent technology that evolves by learning from business operations

Technology2 sources

Argonne National Laboratory launches AI inference service from spare supercomputing capacity

The Department of Energy's Argonne National Laboratory has unveiled a new AI inference service built from spare supercomputing capacity to accelerate scientific discovery. The service provides researchers across DOE labs with secure, cloud-like access to large language models without relying on public platforms like ChatGPT, supporting critical work including the Genesis Mission.

Technology2 sources

Argonne National Laboratory launches AI inference service from spare supercomputing capacity

Technology2 sources

Roku launches biggest home screen update in a decade with AI-powered app organization

Roku unveiled its most significant home screen redesign in over ten years, rolling out automatically to all US devices starting today. The update introduces AI-powered features including Quick Access for app organization, Top Picks for You, and Your Daily Scoop for trending content. The redesign maintains the familiar three-column layout while shifting priorities toward personalized content discovery.

Technology3 sources

Roku launches biggest home screen update in a decade with AI-powered app organization

Technology3 sources

Prev 123 4…22 Next