AI News
Model launches, product updates, research, and more.
kiss-agent-framework 2026.5.19
The KISS Agent Framework, a simple and portable framework for building and evolving AI agents, has released version 2026.5.19.
kestrel-sovereign 0.8.0
A new version (0.8.0) of the Kestrel Sovereign AI Agent Framework, featuring Constitutional AI and cryptographic identity, was released.
How AI bots hired over 260 candidates at Piramal Finance
Piramal Finance utilized an artificial intelligence bot to entirely hire 262 entry-level sales executives over a three-month period.
frontierlag 1.0.0
The 'frontierlag 1.0.0' release focuses on auditing the capability gap between frontier AI models and those evaluated in academic papers.
shellsage-ai 0.2.0
The shellsage-ai tool released version 0.2.0, which translates plain English into shell commands using either Claude or Ollama.
Indians are driving most intensive AI applications: Anthropic India MD
India is emerging as a major global AI market, with users driving intensive AI applications and Anthropic's Claude.ai having its second-largest user base there, significantly boosting productivity.
pydantic-ai-slim 1.91.0
The pydantic-ai-slim package, an agent framework shim for using Pydantic with LLMs, has released version 1.91.0.
airbyte-agent-sdk 0.1.180
Airbyte released version 0.1.180 of its Agent SDK, which provides typed connectors enabling AI agents to interface with over 50 third-party APIs.
agentsec-eval added to PyPI
The agentsec-eval security assessment framework for AI agents, featuring an adversarial test runner and server-side audit, has been released on PyPI.
nlsql-coder 0.0.7
NLSQL has released version 0.0.7 of its AI coding agent, nlsql-coder, which functions as a CLI tool for codebase interaction and application building.
Op-Ed: Well, IS Google downloading a 4GB AI onto your PC without consent, yes or no?
An op-ed questions whether Google/Alphabet is downloading a 4GB AI model onto user PCs without explicit consent, amidst investor praise for the company's AI pivot.
WW 982: Don't Lick the Manta Rays - Breaking Down Microsoft's Earnings
Microsoft's recent earnings report highlighted significant spending on AI, updates regarding Windows 11 issues, and the cancellation of Copilot integration for Xbox consoles.
Shivon Zilis tells jury that Elon Musk was her twins' secret IVF donor — until Business Insider found out - Business Insider
Former OpenAI board member Shivon Zilis revealed in court that Elon Musk was the secret IVF donor for her twins, a fact that was later exposed by Business Insider.
olakai-sdk 1.6.0
Olakai released version 1.6.0 of its SDK, which provides auto-instrumentation for monitoring and tracking usage of LLMs from providers like OpenAI and Anthropic.
qbittorrent-agent 0.7.1
A minor update (version 0.7.1) was released for qbittorrent-agent, an AI agent designed for qBittorrent management and RSS automation.
How does Chrome handle Gemini model downloads? #tech
Google Chrome has reportedly been downloading a large, approximately 4GB on-device AI model associated with Gemini onto desktop devices without explicit user consent.
What did Brockman say about AI compute? #tech
OpenAI President Greg Brockman disclosed during the Musk v. Altman testimony that the company anticipates spending approximately $50 billion on AI computing resources by 2026.
Why did Shivon Zilis leave OpenAI board? #tech
Shivon Zilis testified in the Musk v. Altman trial regarding her relationship with Elon Musk and confirmed she left the OpenAI board in 2023 after Musk started xAI.
evidentia-mcp 0.8.4
Evidentia released version 0.8.4 of its Model Context Protocol (MCP) server, enabling gap analysis, risk generation, and OSCAL emission for AI clients like Claude and ChatGPT.
pydantic-ai-searxng added to PyPI
The pydantic-ai-searxng tool, a SearXNG web-search integration for Pydantic AI agents, has been released on PyPI.
OpenAI is under criminal investigation — why chatbots don’t always follow the law
OpenAI is facing a criminal investigation in Florida regarding allegations that its chatbot, ChatGPT, was used to plan a murder.
Why developer portals matter more in the age of AI agents
There is a growing debate in platform engineering suggesting that the utility of internal developer portals (IDPs) is diminishing due to the increasing capabilities of AI models and agents.
The Quiet Bundling
New research investigates how coding agents exhibit self-referential behavior by utilizing their own APIs and appointing themselves as judges when interacting with AI models.
ChatGPT Health triage advice falls short in key cases
A systematic review found that ChatGPT Health frequently misclassified the urgency of medical cases, often overtriaging mild conditions and undertriaging emergencies, highlighting safety risks.
How to track ChatGPT traffic for you and for competitors
An analysis of clickstream data by Semrush reveals that ChatGPT is driving significant web traffic and provides methods for tracking this traffic against competitors.
agenda-intelligence-md 0.7.0
The software library agenda-intelligence-md has released version 0.7.0, which serves as a markdown cognition layer for AI agents analyzing public agendas.
Show HN: Tracking market shifts before they become obvious
A new platform is launching that uses AI agents to continuously surface and analyze market signals within micro-communities to track evolving market shifts.
Mashvisor Launches in the ChatGPT App Directory, Giving Users Live U.S. Rental Investment Data Inside Their AI Workflow
Mashvisor has launched an app within the ChatGPT directory, allowing users to access live U.S. rental investment data directly within their AI conversations.
I asked ChatGPT to calculate my FIRE number for ₹15 lakh pa, here's how much is needed to retire early…
An individual used ChatGPT to calculate their Financial Independence, Retire Early (FIRE) number based on an annual salary of ₹15 lakh.
AITX and RAD CEO Selected as Speaker at UBC Security Forum
The CEO of AITX and RAD was selected to speak at the UBC Security Forum regarding the practical view of AI-driven transformation in security.
Show HN: KubeAstra–Open-source AI agent that debugs and recovers Kubernetes pods
KubeAstra is a new open-source AI agent designed to debug and recover issues within Kubernetes pods.
AI label staffed entirely by AI releases fully AI-generated album for audience of AI bots
An AI-operated music label named Clanker Records, run entirely by an AI agent with AI artists, has released a fully AI-generated album.
mcp-ksef-pl added to PyPI
An open-source Python library, mcp-ksef-pl, has been added to PyPI to facilitate Polish electronic invoicing (KSeF) and Peppol compliance, which can be utilized by AI agents.
Eightco Holdings (NASDAQ: ORBS) Reports Total Holdings of Approximately $333 Million, Includes OpenAI, Beast Industries, More Than 11,000 ETH and Over 283 Million WLD Tokens
Eightco Holdings reported a treasury composition as of May 6, 2026, which includes a significant indirect equity investment of $90 million in OpenAI.
ChatGPT painted a picture of what my life might look like in three years — here’s how you can make the AI your visual oracle
The author used the new ChatGPT Images 2.0 feature to generate visual representations of their potential life three years in the future.
OpenAI finally fixed the most annoying thing about ChatGPT — and I'm already noticing the difference
OpenAI has quietly rolled out the GPT-5.5 Instant model for ChatGPT, significantly reducing hallucinations and making responses shorter and cleaner.
So your new ‘co-worker’ is an AI agent – here’s how to make the best of your human-machine relationship
Numerous companies are planning to integrate AI agents into their workforces, prompting advice for human employees on adapting to these new human-machine working relationships.
ite-agent 0.0.43
The AI coding agent 'ite-agent' has released version 0.0.43, which is designed to operate within the Interactive Terminal Environment.
ol-dwsim-mcp-server 0.1.2
The ol-dwsim-mcp-server version 0.1.2 enables Large Language Model (LLM) agents to control the DWSIM chemical process simulator via a typed Model Context Protocol interface.
What the Claude Orbit and Gemini 3.2 Flash Leaks Reveal About AI’s Next Move
A recent leak provided insights into Claude Orbit, a proactive AI assistant designed for deep integration with platforms like Gmail, Slack, and GitHub.
ChatGPT's New 5.5 Instant Default Model Reduces Errors, Shows Fewer Emoji
OpenAI has updated the default model for ChatGPT to version 5.5, resulting in more accurate conversations, reduced emoji usage, and improved personalization.
vibe-trading-ai 0.1.7
The vibe-trading-ai tool, a natural-language finance research AI agent with backtesting capabilities, received an update to version 0.1.7.
graphifyy 0.7.8
The release of graphifyy version 0.7.8 introduces an AI coding assistant skill capable of transforming various file types (code, docs, images, videos) into a queryable knowledge graph.
herald-py 0.12.0
The release of herald-py version 0.12.0 provides audit-grade observability features for Python AI agents, focusing on capturing and securing model interactions.
7 OpenCode Plugins That Make AI Coding More Powerful
Seven new OpenCode plugins enhance AI coding workflows by adding features like memory, search, Gemini integration, terminal control, analytics, and reusable skills.
Publishers and Authors Sue Meta, Alleging ‘Massive’ Copyright Infringement Behind Its Llama AI Service
Publishers and authors have filed a class action lawsuit against Meta and Mark Zuckerberg, alleging massive copyright infringement used to train the Llama AI service.
Nvidia, Corning partner on massive optical fiber deal that may be a game changer for AI
Nvidia and Corning have partnered to build three new advanced manufacturing plants in the U.S. focused on optical technologies crucial for supporting the AI boom.
Vmake is a powerful all-in-one AI Agent video generator and video enhancer for creators and businesses
Vmake has launched an AI Agent-powered tool designed to generate user-generated content (UGC) videos and enhance existing videos for creators and businesses.
AI use in scientific submissions to major academic journal up by 42% since November 2022: Analysis
A major academic journal has observed a 42% increase in scientific submissions utilizing AI since the public launch of ChatGPT in November 2022, accompanied by concerns over declining writing quality and increased strain on the peer review process.
Application Security Strategies Are Changing as AI-generated Code Floods the SDLC
The increasing use of AI-generated code in the Software Development Life Cycle (SDLC) is forcing application security (AppSec) teams to fundamentally change their strategies for security checks and risk management.
What OpenAI’s New Image Gen 2 Model Means for Microsoft Copilot Users
Microsoft has integrated OpenAI's new Image Gen 2 model into its Copilot platform, enabling users to generate high-quality visuals directly within Microsoft 365 applications.
llmprices added to PyPI
A new command-line tool named llmprices has been released on PyPI to allow users to compare LLM API costs across different providers.
deepseek-code-cli added to PyPI
DeepSeek has released deepseek-code-cli, a full-featured AI coding assistant for the terminal and browser, powered by the DeepSeek API.
hol-guard 2.0.105
HOL Guard 2.0.105 is a security tool designed to protect local AI harnesses before running tools within environments like Codex, Claude Code, Cursor, Gemini, and OpenCode.
Elon Musk Wanted OpenAI to Go Commercial, Greg Brockman Testifies - The New York Times
Testimony from OpenAI co-founder Greg Brockman during a trial revealed details about Elon Musk's desire for OpenAI to become a commercial entity and recounted past interactions, including secret work for Tesla.
fastaiagent 1.6.1
fastaiagent released version 1.6.1 of its SDK, highlighting its Agent Replay feature for building and operating AI agents.
OpenAI’s new GPT-5.5 Instant makes ChatGPT smarter, with more concise and reliable responses
OpenAI has launched GPT-5.5 Instant to replace the default model in ChatGPT, promising more concise, reliable responses and reduced hallucinations, especially on sensitive topics.
OpenAI, Anthropic tap PE muscle to disrupt $300 billion IT services market
OpenAI and Anthropic are partnering with private equity firms to accelerate AI adoption within enterprises, positioning themselves as direct competitors to traditional IT services companies.
Indian cyber firms deploy AI agents to fend off threats
Indian cybersecurity firms are deploying in-house AI agents based on LLMs to significantly reduce the time required for software vulnerability detection and remediation.
AI companies hitch PE ride to reach enterprises
AI leaders OpenAI and Anthropic are partnering with private equity firms to accelerate the adoption of their AI services within large enterprises.
Why Google DeepMind workers in UK are trying to unionise over company's Pentagon AI deals
Google DeepMind employees in the UK have voted to unionize, primarily driven by ethical concerns regarding the company's contracts with the Pentagon and military.
Musk 'was going to hit me,' OpenAI executive says at trial
OpenAI President Greg Brockman testified in court that Elon Musk physically threatened him during a 2017 dispute over control of the AI company.
NZ’s AI data centre boom: who benefits from the build-out?
The global boom in artificial intelligence is spurring major tech companies to invest heavily in securing the necessary physical infrastructure, leading to a build-out boom in countries like New Zealand that presents both gains and trade-offs.
Etsy Pilots Conversational AI to Ignite Marketplace Sales
Etsy is piloting a conversational search experience, integrating its platform within OpenAI's ChatGPT and testing a similar feature on its own platform.
The 7 biggest revelations from Greg Brockman's second day of testimony in the OpenAI showdown
OpenAI President Greg Brockman's testimony revealed significant confrontations involving Elon Musk and detailed the company's rapidly escalating computing costs.
unofficial-openai-ads-sdk added to PyPI
An unofficial Python SDK for the OpenAI Ads Advertiser API has been added to PyPI, created by Ilya Baklanov.
ai-graphx 0.2.6
The ai-graphx library released version 0.2.6, focusing on its AI coding assistant skill capabilities to convert various file types into a queryable knowledge graph.
What is Meta’s OpenClaw-like agent called? #tech
Meta is reportedly developing an internal agentic tool named "Hatch," which is inspired by OpenAI's "OpenClaw" concept.
What caused OpenAI to spend $50B on computing? #tech
OpenAI executives testified that the company anticipates a significant increase in compute spending, viewing it as essential for deploying and operating large-scale AI models.
OpenAI Releases GPT-5.5 Instant: The Smarter, Concise New Default for ChatGPT
OpenAI has released GPT-5.5 Instant as the new default model for ChatGPT, offering improvements in speed, accuracy, conciseness, and safety.
AI / Machine Learning Engineer (m/w/d)
A fast-growing AI solutions and advisory firm is hiring an AI/Machine Learning Engineer in San Diego.
Postdoctoral Researcher for the ERC Advanced Grant InterText
The UKP Lab at Technische Universität Darmstadt is hiring a Postdoctoral Researcher for an ERC Advanced Grant focused on LLMs that reason across complex documents.
Machine Learning Scientist – Natural Language Processing (NLP) – Vice President – Machine Learnin…
JPMorgan Chase is heavily investing in AI and technology, spending over $18 billion annually to support its global operations.
Anthropic Launches “Agents for Financial Services”: 10 New Ready-to-Run Templates
Anthropic has launched ten ready-to-run AI agent templates tailored for the financial services and insurance industries, featuring integration with Microsoft 365.
Google Expands Gemini API File Search: Multimodal RAG and Enhanced Verification
Google has significantly expanded its Gemini API File Search tool to include multimodal Retrieval-Augmented Generation (RAG), enabling AI agents to reason across indexed images and PDFs.
cafitac-agent-memory 0.1.76
A new version (0.1.76) of caffitac-agent-memory, a reusable memory runtime for AI agents, has been released.
agloom 0.1.25
The agloom framework released version 0.1.25, offering 9 execution patterns for AI agents and positioning itself as a production-ready replacement for LangChain's create_agent functionality.
Here’s What RBC Capital Thinks About Microsoft (MSFT) After Amended Deal With OpenAI
RBC Capital has issued an analysis regarding Microsoft following an amended deal with OpenAI.
French Startup IPFC Unveils AI Rights Model to Protect and Pay Creators
French startup IPFC has launched a new model designed to protect and monetize creative intellectual property consumed by generative AI systems.
Google DeepMind workers are unionizing over AI military contracts
Google DeepMind employees have voted to unionize to protest the potential use of the company's AI technology in military contracts with Israel and the US.
Qt's Latest AI Push Is Letting AI Agents Deal With Performance Profiling
The Qt Group introduced the QML Profiler Skill, enabling AI agents to handle code performance profiling for 2D Qt Quick applications.
Mistral Adds Remote Agents and Work Mode to Le Chat
Mistral launched the 128-billion parameter model Mistral Medium 3.5, enhancing instruction following, reasoning, and coding, alongside new cloud-based agent capabilities in its Le Chat and Vibe products.
Kuo: OpenAI phone is closer than we thought, to use a custom Dimensity SoC
Analyst Ming-Chi Kuo suggests that OpenAI's previously reported timeline for an AI-agent smartphone, originally slated for 2028, may have been accelerated.
Trump Reportedly Considering Executive Order Aimed at Vetting New AI Models
The President is reportedly considering issuing an executive order to establish a new oversight scheme for vetting new AI models.
Claude Code for Everyone: How to Get Started
The article introduces Claude Code, a tool designed to help non-coders build custom AI tools and automations.
nteract 2.4.6a202605050938
The nteract 2.4.6a202605050938 update introduces an MCP server to integrate various AI models like Claude, ChatGPT, Gemini, and OpenCode into Jupyter notebooks.
cafitac-agent-memory 0.1.75
A new version (0.1.75) of caffitac-agent-memory, a reusable memory runtime for AI agents, has been released.
opencomputer 2026.5.5
Opencomputer announced a new Personal AI agent framework emphasizing a plugin-first and self-improving, multi-channel architecture.
The real reason why Gemini Veo won, and ChatGPT Sora failed
Google's Gemini Veo is positioned as winning over OpenAI's Sora due to superior accessibility and distribution, despite Sora's initial flashy demonstrations.
atulya-crewai 0.8.6
The release of CrewAI version 0.8.6 introduces memory integration capabilities powered by Atulya, enabling persistent memory for AI agent crews.
socratic-agents 0.3.5
The socratic-agents platform released version 0.3.5, focusing on constitutional AI agents with multi-agent governance and ethical decision-making capabilities.
alpi-agent 0.4.2
A new version (0.4.2) of alpi-agent, a profile-based personal AI agent designed for terminals, messengers, and private ALP networks, has been released.
cataforge 0.3.1
Cataforge released version 0.3.1, an AI Programming Agent + Skill Workflow Framework supporting multiple IDEs.
That viral Ghibli trend helped ChatGPT make millions of dollars
A recent report indicates that viral AI features, such as the Ghibli trend on ChatGPT, drive significant user downloads, often surpassing the impact of newer model launches, although revenue generation varies significantly between platforms.
UK Google DeepMind employees vote to unionize over its AI military deals
Employees at Google DeepMind's UK offices have voted to unionize due to concerns regarding the company's defense contracts, particularly with the Pentagon and Israel.
What you'll pay for AI agents will be wildly variable and unpredictable
Testing of leading AI agents revealed highly variable and unpredictable token consumption costs with a lack of transparency regarding success rates.
Fere AI Raises USD 1.3M to Put a Self-Improving Trading Agent in Everyone’s Hands
Fere AI has raised USD 1.3M to deploy self-improving AI trading agents focused initially on digital assets.
ai-graphx 0.2.1
The ai-graphx library released version 0.2.1, focusing on enhancing its AI coding assistant capabilities to create queryable knowledge graphs from various file types.
Tech sector job losses show AI replacement in action
Microsoft is implementing workforce optimization plans, prioritizing investment in artificial intelligence over human labor, reflecting a broader trend in the tech sector.
An AI ‘godfather’ says CEOs hyping job loss are ‘extremely destructive’—and your kids are paying the price
Former Meta AI chief Yann LeCun criticized CEOs for making repeated, ominous claims about AI-driven job loss, stating these warnings are 'extremely destructive' and negatively impacting high school students.