Skip to content

A practical perspective for individuals

Less monitoring. More value.

We turn new tools and features into one question: what can you do with them today, without unnecessary hype?

  • AvailabilityIs it actually out?
  • BenefitWhat work could this improve?
  • First stepWhat is worth verifying?
938 practical events in the selection 72 active sources Updated

Available now

What you can try right away

A specific item, procedure and conditions verified in the original source.

Amazon context

AWS Quick Flows now lets you create and share automated workflows directly in chat

available aplikace

What you can tryAWS Quick Flows – creating, running, and editing workflows directly in chat with support for multiple agents and tools

HowIn the AWS Quick application, describe a workflow in plain language in chat, have it created as a flow, test it, and then publish and share it with the team; it can be run from web, desktop, or mobile.

TermsRequires access to the AWS Quick work environment (enterprise deployment); the feature is part of the existing AWS Quick product.

Check access at the source ↗
✓ official
Google worth noting

Google removed the Search button from the homepage and replaced it with AI features

available web

What you can tryNew Google Search homepage on desktop with buttons for generating images, analyzing files, and brainstorming with Gemini

HowVisit the Google Search homepage on desktop and try the new buttons below the search field for generating images, analyzing files, or brainstorming with the Gemini product

TermsDesktop version of Google.com search

Check access at the source ↗
1 source
Amazon context

AWS brings Amazon Textract Insights, a generative AI feature for document processing, to general availability

available API

What you can tryAmazon Textract Insights – a generative AI feature for extracting data from documents using blueprints

HowCreate a Textract Insights project with blueprints, upload a document to S3, and call the StartDocumentInsights API to extract fields with confidence scores and grounding.

TermsRequires an AWS account and an Amazon S3 bucket, accessed via the new Amazon Textract API.

Check access at the source ↗
✓ official
Google worth noting update

Google expands AI Mode in Search: global sidebar for threads and temporary chats with no history

available web

What you can tryTemporary threads in the AI Mode tool in Google Search

HowOpen AI Mode in Google Search on desktop or mobile and click the temporary threads icon in the top right corner; a chat starts that is not saved to history or personalization and disappears once closed.

TermsNo change in settings, registration, or paid tier is required; the feature is part of regular Google Search.

RegionThe rollout is happening globally on desktop and mobile in all supported languages

Check access at the source ↗
✓ 2
Amazon context

AWS expands Amazon SageMaker Inference with Inference Gateway, managed NVIDIA Dynamo, and cross-region routing

restricted access API

What you can tryAmazon SageMaker Inference: Inference Gateway with managed NVIDIA Dynamo, cross-region/cross-account routing of endpoints, and Inference recommendations with LoRA support…

HowIn SageMaker AI Studio, open Jobs → Inference optimization, select a use-case profile, an optimization goal, and a model, and run the benchmark; or in SageMaker Python SDK version 3.17.0+, call the benchmarking and deploy the endpoint…

TermsRequires an AWS account and use of Amazon SageMaker AI (Studio or Python SDK 3.17.0 and higher).

Check access at the source ↗
✓ official
Amazon context

Kimi Code CLI gains native support for Amazon Bedrock with AWS login

available API

What you can tryNative bedrock provider in Kimi Code CLI version 2.7.0 for accessing Kimi models (e.g., Kimi K3.1 Pro Highspeed) via Amazon Bedrock using the standard AWS…

HowUpdate Kimi Code CLI to version 2.7.0, log in via AWS CLI (profile or SSO), and set up a provider of type bedrock in the config.toml file according to the guide from the AWS Machine Learning Blog

TermsRequires an AWS account with access to Amazon Bedrock, configured IAM permissions, AWS CLI version 2.32.0 or newer, and Kimi Code CLI installed

Check access at the source ↗
✓ official

Practical radar

What deserves your attention

Browse the archive →
OpenAI major update New models

GPT-6.1 Terra has half the API input and output rates of GPT-6.1 Sol

GPT-6.1 Terra costs 1 USD per million input tokens and 5 USD per million output tokens through the API, half the rates for GPT-6.1 Sol. According to OpenAI, it focuses on office tasks, visual reasoning and automated software operation.

New: GPT-6.1 Sol costs 2 USD per million input tokens.; GPT-6.1 Sol costs 10 USD per million output tokens.; GPT-6.1 Terra has half the input and output token prices of GPT-6.1 Sol.

What you get out of it

Users on supported plans gain another option for technical writing and working with visual materials. According to Ted Sanders of OpenAI, GPT-6.1 Terra may be better suited to clear technical communication than more expensive models.

✓ 4
Details and sources →
Google major update New models

Gemini 4 Argon has a context window of one million tokens, according to Google

Google states that Gemini 4 Argon has a context window of one million tokens. The Decoder also reports up to one million output tokens. Access remains restricted to selected cybersecurity defenders in Fairwind Program.

New: Gemini 4 Argon is now available worldwide to developers, businesses and consumers.; The model is available through the API in AI Studio and on the Gemini platform.; Google AI Ultra subscribers have gained access in Google Antigravity.; AI Ultra and Pro subscribers have gained access in the Gemini app and AI mode in Google Search.; Under the new tier structure, free users and Plus subscribers have access only to Flash models.

What you get out of it

A developer who is not among the selected participants in Fairwind Program cannot yet try Gemini 4 Argon or count on it as an available tool for everyday work. No date for broader access has been announced.

✓ official
Details and sources →
OpenAI major Business and investment

OpenAI cuts prices for GPT-5.6 Luna and Terra models, says optimization of its own infrastructure by Sol made this possible

From 30 July, OpenAI reduced the price of GPT-5.6 Luna by 80 % (to 0.20 $/1.20 $ per million tokens) and Terra by 20 %. According to the company, the cuts stem from savings achieved by optimizing infrastructure with Sol; commentators point to pressure from low-cost competitors.

What you get out of it

Anyone using OpenAI API for their own projects or tools with a low-cost model (Luna) will pay significantly less for the same work than before, as illustrated by Simon Willison, who switched from Gemini to Luna because of the price.

✓ 2
Details and sources →
Meta major update New models

Zuckerberg defends the open strategy of Meta and distillation in an essay; the company also plans to open the weights of Muse Spark 1.2

To accompany the release of the open model Muse Glimmer (30B, Apache 2.0), Meta published an extensive essay by Mark Zuckerberg defending distillation and its open strategy against OpenAI and Anthropic; it also promised to open the weights of Muse Spark 1.2.

New: Mark Zuckerberg published an essay (6000+ words) on AI philosophy, governance and the strategy of Meta; The announcement includes Muse Code, an AI agent for the terminal; Meta strategically positions open-weight models as an alternative to proprietary approaches from OpenAI and Anthropic; The essay criticizes lobbying by competitors and mentions competition from Chinese open-weight models; The new strategy of Meta focuses on differentiation through an open-source approach

What you get out of it

Developers and enthusiasts can run a full-fledged agent for coding or working with files or images directly on their Mac or PC, without depending on the cloud or sending data to a third party.

✓ 4
Details and sources →
Anthropic major New models

Anthropic has released the Claude Opus 4 and Claude Sonnet 4 models, Claude Code has moved to general availability

Anthropic has released the Claude Opus 4 and Claude Sonnet 4 models, moved Claude Code to GA, and added a code execution tool, MCP connector, Files API, and one-hour prompt caching to the API. Prices remain unchanged.

What you get out of it

Developers can now use Claude Code directly integrated into VS Code and JetBrains with in-file edit display, run background tasks via GitHub Actions, and take advantage of extended thinking with tools as well as improved model memory when working on complex, long-running tasks.

✓ official
Details and sources →
OpenAI important update New models

ChatGPT with GPT-6 models adds interactive forms and responses during reasoning

The Intelligent UI feature in ChatGPT also includes interactive forms. GPT-6 models can respond while reasoning, which OpenAI says reduces waiting time by 44 %. The rollout is global and is proceeding gradually by plan.

New: Intelligent UI also supports interactive forms.; OpenAI says it trained GPT-6 to decide when to use interactive visualizations and how to format them.

What you get out of it

When learning, you can use buttons to explore individual parts of an explanatory diagram. Those who prefer text can ask ChatGPT for fewer visual elements.

✓ 4
Details and sources →
Hugging Face important update Tools and apps

Hugging Face launches language model routing through OpenRouter with a "fastest" or "cheapest" option

Hugging Face has launched full production routing of language models through OpenRouter: automatic provider selection based on price/speed, ":fastest" and ":cheapest" suffixes, the end of BYOK for language models (remains for images/video), and billing directly through Hugging Face without a markup.

New: Hugging Face Inference Providers offers :fastest and :cheapest routing modifiers to control distribution; Hugging Face bills directly without a markup from OpenRouter and settles with providers; BYOK (bring-your-own-key) for language models has been discontinued, it remains only for images and video; OpenRouter reports a 46% reduction in capacity errors and a 16.3% increase in throughput; Time-to-first-token improved by 16.9% with the OpenRouter integration

What you get out of it

Developers calling language models through Hugging Face no longer need to provide a custom API key for an individual provider and can control provider selection by appending ":fastest" (highest throughput) or ":cheapest" (lowest price) to the model identifier, or alternatively choose a specific provider by name.

✓ official
Details and sources →
NVIDIA important update New models

NVIDIA released the Granary v2 dataset with 844 thousand hours of audio for Nemotron 3.5 Transcribe

Nemotron 3.5 Transcribe has 30 billion parameters, approximately 4 billion of which are active, and runs on a single GPU according to NVIDIA. The release is accompanied by the Granary v2 dataset with 844 thousand hours of audio and tools for domain adaptation.

New: The model has 30 billion parameters, with approximately 4 billion active in a single step.; The model can run on a single GPU.; The release includes the Granary v2 dataset with 844 000 hours of audio.; The release includes tools for adapting the model to specialized domains.

What you get out of it

For a recorded meeting, text instructions let you choose between a transcript with speaker labels, a translation and an overview of decisions made. Available features vary by language.

✓ official
Details and sources →
OpenAI important update Tools and apps

ChatGPT Work: OpenAI clarified the Cloud and Local versions and internet access

New analyses describe ChatGPT Work, an agentic tool from OpenAI released on 9 July 2025 at $20/month: it comes in Cloud (web) and Local (desktop app, formerly Codex) versions, with code execution in Work Cloud having internet access, unlike regular chat.

New: ChatGPT Work exists in two versions: Cloud (web-based through chatgpt.com) and Local (desktop app, formerly Codex); Code execution in Work Cloud has internet access, while Chat does not have this capability; Exact release date: 9 July 2025; The Local version is a redesigned Codex app for nontechnical users; Offers Sol/Luna/Terra models with reasoning levels from Light to Ultra; separate token limits for Work vs Chat

What you get out of it

A user with a subscription costing $20 per month or more can connect ChatGPT Work to email, Slack, Notion or Figma and let it independently complete multistep tasks, including code execution with internet access – but this means handing control over sensitive personal and work data to the model.

✓ 2
Details and sources →
GitHub important Tools and apps

GitHub Copilot gains eight new MCP connectors in public preview

GitHub has added eight new MCP connectors to Copilot (Supabase, Box, Calendly, Buildkite, and others) in public preview for all Copilot plans. For enterprise plans, administrator approval is required.

What you get out of it

Developers can now work directly within Copilot with the Supabase schema before writing a migration, or identify and fix the cause of a failed build in Buildkite, without leaving the working session.

✓ official
Details and sources →
OpenAI important New models

OpenAI makes GPT-6 Sol and GPT-6 Luna available on the Amazon Bedrock platform

GPT-6 Sol and GPT-6 Luna from OpenAI are now generally available on the Amazon Bedrock platform at lower prices than GPT-5.6; according to OpenAI, GPT-6 Sol makes about half as many factual errors as GPT-5.6 Sol.

What you get out of it

Developers working with the Amazon Bedrock platform gain another model choice based on the task type: assign more demanding tasks to GPT-6 Sol and simple, frequent tasks to GPT-6 Luna, and adjust the reasoning effort level as needed.

✓ official
Details and sources →
important New models

Nemotron 3.5 Lightning model from NVIDIA available in Ollama for local agentic tasks with a context of 1M tokens

NVIDIA has made the open-source model Nemotron 3.5 Lightning (30B parameters, 3B active per token, MoE) available in Ollama for local AI agents. It supports a context of up to 1M tokens, runs on NVIDIA GPUs as well as Apple silicon and can be further trained for custom tasks.

What you get out of it

Developers can download the model through Ollama and run it locally on their own GPU (RTX PC, RTX PRO workstation, DGX Spark or Station, as well as Apple silicon through the MLX variant), including integration with tools such as Claude Code, OpenCode or Hermes Agent, without data leaving the device.

✓ official
Details and sources →

When personal experience proves useful

Does it make sense for a team too?

Look at the same topics through the lens of processes, risks, and implementation within a company.

Go to the company radar →