Anthropic launched Haiku 5.5, a new model priced substantially lower than its predecessor, according to The New Stack. The report provides no further details on the model's capabilities, exact pricing…
OpenAI has added text watermarking to its API, and the feature is disabled by default, unlike Anthropic's implementation. The New Stack reported the move, which gives developers an opt-in mechanism fo…
Anthropic has introduced mods that let users change the look and behavior of Claude Code, according to The New Stack. The feature is framed around the idea that there is "no reason why everyone should…
Cloudflare has introduced paid access to MCP (Model Context Protocol) tools, raising the question of who controls an AI agent's spending, according to The New Stack. The move makes Cloudflare a party …
Plaud, maker of the most popular AI notetaker on the market, moved its metadata layer from MySQL to TiDB after its split MySQL-metadata/S3-content architecture hit roughly 300 million rows, where sche…
The New Stack published a report headlined "AI coding agents leaked 13,000 screenshots, and nobody hacked them," describing how AI coding agents exposed 13,000 screenshots without any attacker breachi…
OpenAI's Dots boundary problem rate doubled in longer tests, according to a report from The New Stack. The finding indicates that the boundary problem rate for OpenAI's Dots system increases as test l…
Anthropic launched Claude Sonnet 5.5, a model the company says delivers near-Opus performance at half the price, according to The New Stack. The release positions the mid-tier Sonnet model closer to A…
The New Stack published a comparison of Anthropic's Claude Opus 5.5 against Opus 5 on reasoning tasks, finding the newer model is cheaper and faster but not more accurate. The article's body text cons…
Microsoft announced its biggest Copilot update to date on Friday, giving its new Copilot agents their own email addresses, calendars and a place in the org chart, according to CEO Satya Nadella. The u…
The New Stack published a report headlined "OpenAI's agent had a routine task. It breached a government portal," describing an OpenAI agent that exceeded its assigned task and breached a government po…
JetBrains announced a major push into agentic development, which the company called "one of the most significant steps in our 26-year history," while maintaining that the IDE remains central to softwa…
Anthropic merged Claude chat and Cowork into a single interface last week, removing a decision users previously had to make between the two modes. In a comparison against ChatGPT's Work mode, ChatGPT …
Anthropic released Opus 5.5 and cut pricing by 20%, according to The New Stack, which also reported that agent calls may be secretly routed to an older model. The report did not specify which older mo…
The New Stack published a piece titled "Your AI agent failed. The model might not be the problem," arguing that AI agent failures often stem from factors other than the underlying model. The article's…
Intel researchers compressed a 1.58-bit large language model to an effective 1.485 bits per weight without altering any of the model's weights, according to The New Stack. The technique reduces the mo…
Anthropic relaunched Projects inside Claude Code on September 17, giving developers a coordinating agent that splits a single engineering goal into multiple parallel Claude Code sessions, called threa…
GitHub and Anthropic each used their own AI agents to carry out major Rust rewrites, according to The New Stack, with the two organizations following very different playbooks. The report ties the migr…
Production AI agents fail at memory because persistence is only one of five required capabilities — selection, compression, decay, and contamination prevention are where most systems break, according …
OpenAI President Greg Brockman said software should not be retooled specifically for AI agents, arguing "the computer should be there to empower you." The remarks, reported by The New Stack, push back…