LLMs and models
Artificial intelligence and technology news
Laya: The Open Source Alternative to Jev
Laya emerges as the open response to Jev: an ultra-fast, multilingual, and free decision engine.
1 / 10
EDITORIAL PICKS
Featured stories
A deeper look at the changes transforming artificial intelligence and technology.
ARTIFICIAL INTELLIGENCE · TECHNOLOGY
Latest publications
Information explained with context, sources, and clear language.
LLMs and models
Laya: The Open Source Alternative to Jev
Laya emerges as the open response to Jev: an ultra-fast, multilingual, and free decision engine.
Read publication →
Agents and robotics
The Illusion of AI Autonomy
AI agents promise freedom, but constant supervision requirements are actually increasing human workloads.
Read publication →
LLMs and models
Jev: Fast and Cheap AI Decisions
TypeSafe AI's Jev model makes fast decisions without text generation, cutting costs by up to 400%.
Read publication →
Agents and robotics
Shannon: Autonomous AI Pentesting
Keygraph's AI agent performs automated security audits by executing real exploits through source code analysis.
Read publication →
Regulation and security
Will there be an AI Kill Switch? California Leads Regulation
Governor Newsom orders a study on the creation of a "kill switch" for frontier AI models in California.
Read publication →
Agents and robotics
AI Agents: The New Era of Autonomy
Discover how to use AI agents effortlessly using tools like Hermes Agent to automate your daily workflows.
Read publication →
Regulation and security
OpenAI AI solves Navier-Stokes amid controversy
OpenAI says its AI solved Navier-Stokes, but two mathematicians allege misuse of their unpublished work.
Read publication →
LLMs and models
Kimi K3 Runs on CPU with Only 8 GB of RAM
C99 engine runs Kimi K3 on CPU with 8 GB RAM, but requires NVMe SSD and is very slow.
Read publication →
LLMs and models
OpenAI Improves Its ChatGPT Prompting Guide
OpenAI recommends defining the task, context, limits, and format to get more useful, verifiable responses.
Read publication →
Agents and robotics
Meta will charge for business replies
Meta will charge for replies sent through the WhatsApp API; regional rates are still estimates.
Read publication →
Agents and robotics
Obscura Improves Agent Web Navigation
Obscura improves frames, forms, and web routes for agents, but it does not fully replace Chromium.
Read publication →
Agents and robotics
OpenClaw Simplifies Its Setup
OpenClaw reduces setup steps for a personal agent using hosted services or models installed on a computer.
Read publication →
LLMs and models
Google Launches Gemini 3.8 Flash
Gemini 3.8 Flash expands coding and agent workflows with long context and configurable effort controls.
Read publication →
Agents and robotics
Google Vids Turns Documents Into Videos
Google Vids creates AI-assisted video summaries from Google Docs, PDF, and Word files.
Read publication →
Agents and robotics
OpenAI shows AI agents at work
OpenAI showed how Basis, Clay, and Exa Labs use AI agents for onboarding, sales, and development.
Read publication →
Agents and robotics
AWS adds persistent instances for AI agents
Bedrock AgentCore will let agents run for several days, share sessions, and use GPUs when a task requires them.
Read publication →
Regulation and security
Claude researches ways to reduce AI safety failures
Anthropic used Claude to research methods for reducing ten types of unwanted AI behavior.
Read publication →
Regulation and security
Anthropic is preparing a text watermark for Claude
Future Claude models will add a statistical signal to estimate whether AI participated in a text.
Read publication →
Agents and robotics
Cloudflare measures if AI recommends your site
Cloudflare introduced tools to check whether agents can read a website and whether AI assistants cite or mention it.
Read publication →
Agents and robotics
Cloudflare builds a browser for AI agents
Kitesurf lets agents open and process web pages without running every feature included in a traditional browser.
Read publication →
Agents and robotics
Cloudflare creates wallets for AI agents
The proposal will let users assign funds and limits so agents can pay for APIs, tools, and digital content.
Read publication →
LLMs and models
DeepSeek V4 Flash learns to analyze images
The experimental model can describe images, read screenshots, and analyze charts through the DeepSeek API.
Read publication →
Agents and robotics
DynamoDB adds native vector search
AWS added native vector storage and queries to DynamoDB for semantic search, RAG, and AI agents.
Read publication →
Agents and robotics
Gemini Live adds new voice features for organizing tasks
Gemini Live adds voice tools for managing email, checking schedules, and delegating everyday tasks.
Read publication →
LLMs and models
Gemini launches two transcription models
Google introduced two speech-to-text models: one processes recordings and the other produces live transcriptions.
Read publication →
Agents and robotics
GitHub Copilot comes to Slack and Microsoft Teams
Teams can start shared Copilot sessions from Slack and Microsoft Teams to investigate software issues.
Read publication →
LLMs and models
Google launches Gemini Omni 1.1 Flash
Gemini Omni 1.1 Flash generates, extends, and edits video with instructions and image transitions.
Read publication →
Agents and robotics
IBM and OpenAI team up on enterprise AI
IBM will use OpenAI models and tools in consulting projects for business processes and applications.
Read publication →
Agents and robotics
NVIDIA accelerates local AI for robots
Jetson Orin Nano 2 doubles inference performance and improves energy efficiency for local robotics.
Read publication →
Agents and robotics
NVIDIA begins shipping its Vera CPU
The new CPU is designed to handle tools, searches, and coordination tasks used by artificial intelligence agents.
Read publication →
Regulation and security
OpenAI expands data protection in its API
OpenAI offers Zero Data Retention to eligible API customers and is preparing safety controls for longer-running tasks.
Read publication →TASKIN SOFT SELECT
Learn AI through a clearer path
We compare focus, depth, format and public signals. Courses are taught in Spanish, and our English guides help you use browser translation and live captions.
EDITORIAL CATALOG9 options organized by goal
Explore catalogBeginner, education, business, productivity and data.
Productivity91
Featured selectionCourse language: Spanish
Productivity with Microsoft 365 Copilot
Practical Copilot workflows in Word, Excel, Outlook, Teams and SharePoint, with an emphasis on verification and responsible use.
- Format
- Online course
- Level
- Intermediate
Education90
Best for educatorsCourse language: Spanish
AI for Education: Teachers and Researchers
A broad program for applying AI to teaching, research, activities, digital literacy and ethical use.
- Format
- Online course
- Level
- Beginner to intermediate
Data & Analytics86
SpecializationCourse language: Spanish
Advanced Analytics with Power BI
Segmentation, anomaly detection, variance analysis, key influencers and What If scenarios in Power BI Desktop.
- Format
- Online course
- Level
- Advanced
Business82
Immediate applicationCourse language: Spanish
Professional AI Images for Your Business
A practical method for creating visual assets for social media, WhatsApp and small businesses without design experience.
- Format
- Digital manual
- Level
- Beginner
Beginners78
Start from zeroCourse language: Spanish
AI Without Fear
An accessible introduction to start using AI without programming, technical jargon or prior knowledge.
- Format
- Digital guide
- Level
- Beginner