Topic archive
Clear explanations of generative AI and language models, including their capabilities, limits, privacy implications, and safe use.
The public SynthID Detector checks images, video and audio for supported watermarks. Here is how to interpret a result without jumping to conclusions.
Anthropic has released Haiku 5.5. Here is how its two pricing tiers work, what changed for Sonnet cache reads and who can claim monthly API credits.
Claude can now work directly inside Google Workspace files. We explain the add-on, its access requirements, edit approvals and how it differs from connectors.
OpenAI is changing how ChatGPT presents answers. Here is what interactive responses offer, who gets access and why a Pro subscription differs from Pro reasoning.
MAI-Transcribe-2-Streaming recognizes live speech, while MAI-Voice-2.1 and Flash generate it. Pricing, language support and preview limits explained.
Google starts Gemini 4 Argon access with Fairwind partners. We explain announced pricing, the output limit and how CodeMender fits.
New Muse connections target everyday small-business work. We explain how permissions matter and suggest a concrete first assignment.
DevDay brought a new Sol model, a separate Ultrafast tier and cloud-based Codex tasks. Here are the availability details and practical limits.
Dots have their own cloud computer and can handle ongoing tasks. Here is what to know about availability, permissions, pausing and deleting an assistant.
A closer look at the voluntary White House agreement: four layers of oversight, external reviews and the questions that remain unanswered.
The new Sonnet has arrived. Understand token pricing versus task costs, choosing effort levels and the changes that matter for API migration.
ChatGPT voice conversations now work with available plugins. We explain Chat versus Work, permissions and how to check the results.
