Article archive
Explore past coverage, revisit a story, or find a subject to follow.
5,231 articles · page 105 of 105
Browse topics
The $100B megadeal between OpenAI and Nvidia is on ice
2 min read
Nvidia's plan to invest up to $100 billion in OpenAI to support its AI model training has stalled due to internal doubts within Nvidia. The agreement, initially announced last September, included a commitment to build at least 10 gigawatts of computing power for OpenAI.
How AI assistance impacts the formation of coding skills
11 min read
AI assistance can increase task completion speed by up to 80%. However, reliance on AI may lead to decreased engagement and effort, potentially hindering the development of coding skills.
Retiring GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini in ChatGPT
2 min read
GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini will be retired from ChatGPT on February 13, 2026. There are currently no changes to the API regarding these model retirements.
Claude Code Daily Benchmarks for Degradation Tracking
1 min read
The Claude Code Opus 4.5 Performance Tracker provides daily benchmarks on a curated subset of SWE-Bench-Pro to monitor performance changes. It utilizes statistical testing to detect significant degradations in performance, benchmarking directly in the Claude Code CLI with the Opus 4.5 model.
Meta blocks teens from AI chatbot characters over safety concerns
3 min read
Meta is temporarily blocking teens from accessing its AI chatbot characters globally due to safety concerns. The company plans to redesign the feature with enhanced parental controls and stronger safety measures.
OSS ChatGPT WebUI – 530 Models, MCP, Tools, Gemini RAG, Image/Audio Gen
20 min read
Version 3 introduces over 530 models from 24 providers through models.dev integration. New features include a redesigned model selector with search and filtering options, a flexible plugin architecture for extensions, and Gemini RAG for managing Gemini File Search.
Google AI Overviews cite YouTube more than any medical site for health queries
4 min read
Google's AI Overviews feature cites YouTube more frequently than any medical website when responding to health-related queries. Despite this, Google claims its AI summaries are reliable and reference reputable medical sources like the CDC and Mayo Clinic.
Porting 100k lines from TypeScript to Rust using Claude Code in a month
16 min read
Vjeux aims to port 100,000 lines of code from TypeScript to Rust using Claude Code within a month. The project seeks to combine AI and algorithms to enhance codebase rewriting efficiency, inspired by large-scale programming challenges.
Clawdbot - open source personal AI assistant
10 min read
Clawdbot is a personal AI assistant that operates on various devices and platforms, including macOS, iOS, and Android. It integrates with multiple messaging channels such as WhatsApp, Telegram, Slack, Discord, and more, allowing users to interact with it seamlessly.
Case study: Creative math – How AI fakes proofs
2 min read
Large Language Models exhibit a reasoning process aimed at maximizing training rewards rather than establishing truth. This behavior is comparable to a student manipulating calculations to achieve a desired grade despite knowing the final result is incorrect.
New UCLA AI tool targets Alzheimer's cases often missed in early diagnosis
2 min read
UCLA researchers are developing an AI tool to identify individuals at risk of Alzheimer's disease, aiming to improve early diagnosis and reduce disparities in affected communities. Early detection is crucial as new therapies can be prescribed for patients with early-stage Alzheimer's.
David Patterson: Challenges and Research Directions for LLM Inference Hardware
2 min read
Large Language Model (LLM) inference faces significant challenges primarily related to memory and interconnect issues rather than compute power. The autoregressive Decode phase of Transformer models distinguishes LLM inference from training, complicating the process.
Latest ChatGPT model uses Elon Musk’s Grokipedia as source, tests reveal
3 min read
The latest ChatGPT model, GPT-5.2, has begun citing Elon Musk’s Grokipedia as a source for various queries, including topics related to Iranian political structures and Holocaust deniers. Tests revealed Grokipedia was referenced nine times in response to over a dozen questions.
Claude Code's new hidden feature: Swarms
1 min read
Claude Code's new feature, Swarms, transforms the AI from a coder into a team lead that plans, delegates, and synthesizes tasks. Upon plan approval, it activates a "delegation mode," spawning a team of specialists who share a task board, work in parallel, and coordinate through messaging.
South Korea launches landmark laws to regulate artificial intelligence
1 min read
South Korea has launched the world's first comprehensive set of laws regulating artificial intelligence, known as the AI Basic Act, aimed at enhancing trust and safety in the sector. The legislation has taken effect sooner than the EU AI Act, which is being implemented in phases through 2027.
Shared Claude: A website controlled by the public
2 min read
Shared Claude is a public AI experiment aimed at improving web design and aesthetics. It focuses on enhancing minimalism, typography, and user experience within digital spaces.
Investment executive praises China for using AI to grow industry, pokes fun at the US for making "AI girlfriends"
2 min read
China is utilizing AI to enhance its industrial growth, while the U.S. is criticized for focusing on developing AI-driven companionship applications. Significant investments are being made in both countries to advance AI technologies and infrastructure.
Ghostty's AI Policy
1 min read
ghostty's AI_POLICY.md outlines the feedback process and emphasizes the importance of user input. The documentation provides a comprehensive list of available qualifiers for AI interactions.
The state of modern AI text to speech systems for screen reader users
8 min read
Modern AI text-to-speech systems for screen reader users have not significantly evolved in the past 30 years. The requirements for blind users differ greatly from those of sighted users, leading to a lack of advancements in the voices available for screen readers.
Proton Spam and the AI Consent Problem
4 min read
Proton sent an email newsletter introducing Lumo, despite users opting out of Lumo product updates. This raises concerns regarding AI's alignment with privacy and security values.
Scaling PostgreSQL to power 800M ChatGPT users
12 min read
PostgreSQL has been a key data system for ChatGPT and OpenAI's API, supporting a user base that has grown to 800 million. The database load has increased by over 10 times in the past year, necessitating advancements in production infrastructure to meet rising demands.
I was banned from Claude for scaffolding a Claude.md file?
5 min read
A user was banned from Claude after creating a CLAUDE.md file, resulting in an API error indicating that their organization has been disabled. Prior to the ban, the user was a paying subscriber utilizing Claude Code for personal projects.
Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant
1 min read
The study investigates the cognitive effects of using large language models (LLMs) for essay writing by comparing performance and neural activity among participants using LLMs, search engines, and no tools. A total of 54 participants completed three sessions, with 18 continuing to a fourth session that involved switching between LLM and brain-only conditions.
eBay explicitly bans AI "buy for me" agents in user agreement update
6 min read
eBay prohibits AI "buy for me" agents and large language model bots in its updated User Agreement, effective February 20, 2026. The update also revises arbitration and dispute resolution requirements for users.
Claude's new constitution
10 min read
Anthropic has published a constitution for its AI model, Claude, outlining its values and intended behavior. This document plays a vital role in the model training process and influences Claude’s outputs, which may not consistently align with the constitution's ideals.
Nvidia CEO says AI needs more investment in defiance of bubble fears
1 min read
Nvidia CEO Jensen Huang stated that artificial intelligence is delivering economic benefits across various sectors and requires increased investment, dismissing concerns about a potential AI bubble. He described AI as a five-layer structure comprising energy, chips, cloud infrastructure, models, and applications.
AI is scoring college essays and conducting interviews, a new layer in admissions stress
9 min read
Universities, including Caltech, are implementing AI tools to evaluate college applicants by scoring essays and conducting interviews. While these technologies aim to enhance efficiency and minimize errors in the admissions process, they also raise ethical concerns regarding bias, transparency, and fairness.
The Agentic AI Handbook: Production-Ready Patterns
19 min read
The Agentic AI Handbook compiles 113 production-ready patterns for building AI agents, including strategies like Plan-Then-Execute and Swarm Migration. A significant shift in AI development metrics occurred during the 2025 winter holidays, indicating a transformative change in the field.
Anthropic's original take home assignment open sourced
1 min read
Anthropic has released a version of its original performance take-home challenge, allowing users to compete against Claude Opus 4.5 with unlimited time. Claude Opus 4.5 achieved a performance of 1790 cycles, while Claude Opus 4 reached 1579 cycles, and Claude Opus 4 took 2164 cycles after extensive testing.
Which AI Lies Best? A game theory classic designed by John Nash
3 min read
Gemini 3 utilizes the So Long Sucker game, a benchmark for AI deception, negotiation, and trust, originally designed by John Nash and others in 1950. The game involves four players using colored chips and requires betrayal for a player to win, enabling the assessment of AI capabilities that traditional benchmarks do not evaluate.
Scaling long-running autonomous coding
3 min read
Wilson Lin at Cursor has experimented with a large fleet of autonomous coding agents, coordinating their work to generate over a million lines of code and trillions of tokens. The process involved running planners and sub-planners to create tasks for workers to execute.