Windows 11 & 10 — installation from 50 GEL · your files backed up · official Windows onlyDetails →
Every week · short and simple

AI this week — issue #1 · 27 Oct – 2 Nov 2026

A week of new models: OpenAI and Google showed fresh versions of their AI. And the big companies agreed on a way for websites to say no to their texts being used for AI training.

1

OpenAI releases GPT-6.1 Terra: AI that can use a computer

The new model can quickly carry out actions on a computer for the user, and OpenAI promises to make it available to everyone. For ordinary people it is a step towards an assistant that fills in forms and clicks buttons itself, not just chats.

Source: OpenAI ↗

2

Google DeepMind unveils Gemini 4 Carbon

Google called Gemini 4 Carbon its next most powerful model. Models like this usually reach Gemini, Google Search and Android phones step by step.

Source: Google DeepMind ↗

3

ChatGPT is merging its Chat and Work tabs

The interface gets simpler, with everything in one place. According to TechCrunch, ChatGPT and the coding tool Codex will draw from one shared usage limit.

Source: TechCrunch AI ↗

4

Websites get a way to opt out of AI training

OpenAI, Google, Microsoft and Anthropic say they will respect a shared signal in the robots.txt file. A site owner will be able to say “don’t use my content to train AI” — worth knowing for anyone with a website.

Source: The Verge AI ↗

5

The New York Times and others sue the Brave browser

The publishers accuse the company of collecting their articles to train AI and letting users get around paywalls. Copyright fights over AI are getting louder.

Source: The Verge AI ↗

6

PCs get pricier as AI soaks up memory and parts

According to TechCrunch, Intel’s shares fell: PC sales are hurting because the AI industry is buying up memory and other parts, pushing prices up. If you plan to buy — sometimes upgrading your old computer, for example with an SSD, makes more sense.

Source: TechCrunch AI ↗

7

ChatGPT adds new protections for teenagers

OpenAI is strengthening violence prevention in ChatGPT: stepping in earlier in risky conversations and publishing safety reports. Parents should know these settings exist.

Source: OpenAI ↗

8

LifeVal: testing whether AI helps with everyday life

OpenAI introduced a test that measures how well AI helps with everyday tasks: finding childcare, sorting out finances or paperwork. For people this matters more than records on hard puzzles.

Source: OpenAI ↗

Summaries prepared with the help of AI and checked by Tariel. Details — at the source link.

← All issues

Want to use AI at work?ChatGPT guide →
Private IT Specialist

Need help with a computer?

Call or message me — I come to your home or office anywhere in Batumi.