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?
939 practical events in the selection 72 active sources Updated

Practical radar

What deserves your attention

Browse the archive →
worth noting Robotics

Hugging Face has published a guide to training and deploying locomotion control for the Microduck robot

The guide from Hugging Face connects locomotion training for the Microduck robot in Isaac Lab and Newton with ONNX export for the physical robot. It includes configurations for flat and uneven terrain; simulation and training require Linux and a supported NVIDIA GPU.

What you get out of it

When deploying a custom control model for the Microduck robot, all 61 input values must retain the same meaning and order, and the scaling of the 14 outputs must remain the same as during training.

1 source
Details and sources →
worth noting AI agents

Bar-Ilan University pilots an AI study agent on the Moodle platform

Bar-Ilan University made an AI study agent available in 71 courses for approximately 2 400 students. According to Microsoft, it draws exclusively on materials provided by instructors and supports searching, summarising and practice through quizzes.

What you get out of it

Students participating in the pilot can look up material they missed and have practice questions prepared from their course materials. They do not have to rely solely on notes from classmates.

1 source
Details and sources →
worth noting Coding

AWS describes a process for deploying and comparing models in Amazon SageMaker JumpStart

AWS published a process for deploying a model in Amazon SageMaker JumpStart, running repeatable tests using SageMaker Python SDK and then deleting the endpoint. The comparison tracks latency, throughput and output quality.

What you get out of it

Developers get a concrete process for repeatedly comparing candidate models when choosing one for their own application.

1 source
Details and sources →
worth noting Coding

Geoffrey Huntley describes a pre-commit hook to prevent bypassing a diagnostic library

Geoffrey Huntley describes a pre-commit hook that checks git diff and blocks the commit when it finds a direct call to observability modules outside the diagnostic library. The AI agent then receives an instruction to use the required library.

What you get out of it

When working with an AI agent, a developer can use automated feedback to correct a specific rule violation instead of repeatedly adding instructions.

1 source
Details and sources →
GitHub worth noting Coding

A guide from GitHub shows how to adapt GitHub Copilot to project rules

GitHub describes how to provide GitHub Copilot with project rules: store shared instructions in the repository, define rules for specific files using applyTo, and describe recurring workflows or specialized tasks using agent skills and custom agents.

What you get out of it

A developer can store a frequently repeated convention directly in the repository and observe whether GitHub Copilot uses it during the next task.

1 source
Details and sources →
worth noting AI agents

Author uses ML Intern to create a smaller model for rewriting prompts

The author used ML Intern to create a prompt-rewriting model with 0.8 billion parameters that, according to the author, runs on a CPU. The compute costs of the project amounted to 16 USD; the author reports 99.7 % valid outputs and approximately a quarter of the token usage.

What you get out of it

A developer who needs to rewrite prompts for Qwen-Image 2.1 gets a description of how to create a smaller variant, along with published prompts to use as a starting point for their own experiment.

1 source
Details and sources →
worth noting Tools and apps

Qlik described source retrieval and answer checking in Qlik Answers

According to Qlik, Qlik Answers retrieves source material from unstructured content using Amazon OpenSearch Service and adds citations. Amazon Bedrock Guardrails then compares generated answers with the sources used.

What you get out of it

Users of Qlik Answers receive citations with answers based on unstructured content, allowing them to locate the source material and assess a specific conclusion.

1 source
Details and sources →
worth noting Business and investment

The McDonald's chain in the US uses AI to recommend prices to individual restaurants

The McDonald's chain in the US uses a machine learning system that recommends prices based on transactions, local willingness to pay, and competitors' prices. According to the company, franchisees decide; the algorithm's influence on specific price differences has not been proven.

What you get out of it

A customer of McDonald's restaurants in the United States cannot assume the same price for the same item even at nearby locations.

1 source
Details and sources →
Microsoft worth noting Tools and apps

Approximately 70% of reviews did not complete during the Copilot Code Review service outage

The Copilot Code Review service experienced an outage on 28 September from 19:23 to 22:08 UTC, during which, according to GitHub, approximately 70% of requested reviews did not complete. Failed reviews were not automatically retried; users must request them again.

What you get out of it

If you requested a code review during the stated period, it may have remained incomplete even after the service was restored.

✓ official
Details and sources →
Amazon worth noting update AI agents

Amazon blocked the shopping agent Muse from Meta from accessing its store

Since 20 September 2026, Amazon has been blocking the shopping agent Muse from Meta from accessing its store — according to Amazon, the agent does not disclose that it is AI and stores data without consent. Amazon previously imposed similar restrictions on agents from Perplexity, Google and OpenAI.

New: The agent does not identify itself as AI while browsing the platform; It appears to store customer data without consent; Muse became the most downloaded free app in the US App Store within a week of launch (8 September 2026); Amazon also blocks agents from Perplexity, Google and OpenAI; Meta and Amazon remain business partners with a billion-scale deal for AI chips

What you get out of it

Users of the Muse app from Meta currently cannot shop on Amazon.com through the agent, and according to Amazon, their data may be stored without proper consent, without this being apparent during use.

✓ 2
Details and sources →
OpenAI worth noting update Business and investment

OpenAI expands ads in ChatGPT to 31 European countries

From 26 August 2026, OpenAI is expanding ad testing in ChatGPT to 31 European countries (Germany, France, Spain, Italy and others). Only users of the Free version and the Go plan at 200 Kč/month will see ads.

New: Expansion of testing to 31 European countries planned for 2026-08-26; Ads are shown only to users of the Free version and the Go plan (200 Kč/month); Plus, Pro and Enterprise users are protected from ads; 9 named countries included: Germany, France, Spain, Italy, Sweden, Norway, Denmark, the Netherlands, Austria; Ad Manager (Ads Manager) for Europe will be available later in the summer

What you get out of it

Users of the free version and the Go plan of ChatGPT in Europe will start seeing paid ads directly in conversations from the end of August 2026; according to OpenAI, these are labelled messages that do not influence the content of responses.

✓ 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 →