Section
AI
What changed in artificial intelligence: new models, rules and tools, explained for everyday readers with links to the original sources.
60 POSTSUPDATED
IndexDate · Title · Reading time
-
OpenAI Decisions API now generally available on GPT-6 Luna
OpenAI has opened its fast decision model to all customers, adding GPT-Live voice integration, prompt caching and structured JSON input. TechCrunch reports responses of roughly 100 to 180 milliseconds, yet public indexes still rank TypeSafe's Jev ahead.
-
Canva Craft AI image model is Canva's first built in-house
Canva built and trained Craft itself for design work, not just pictures. It is available in Dream Lab and through the Canva API, with cheaper Craft-fast and Craft-ultrafast versions planned.
-
Prompt engineering techniques: 9 ways to write better LLM prompts
A vendor-neutral guide built on the official OpenAI, Anthropic and Google documentation, with before-and-after examples and about 40 lines of Python to test prompt variants against your own evaluation set.
-
DeepSeek funding round reportedly reaches $15 billion
DeepSeek has reportedly raised about $15 billion, with Tencent and CATL among its largest backers, days after the US ordered federal civilian agencies to remove its technology. The company has not confirmed the round.
-
US banks block AI agents from logging in to customer accounts
JPMorgan Chase, Citigroup, Bank of America and Charles Schwab are shutting out third-party AI agents that log in with customers' passwords. Read-only data sharing survives, but who pays for an agent's mistakes is still unresolved.
-
Claude Duet: Anthropic opens preview of background agent app
Anthropic's Claude Duet sends morning briefs, monitors work apps and runs tasks on its own cloud computer, asking before it pays, sends or signs. Here is who can join the preview and how it compares with rivals.
-
How to build an AI agent in Python with tool calling
A framework-free guide to building an AI agent in Python: describe tools with JSON Schema, write the loop that runs tool calls, switch to a local model via Ollama, and enforce guardrails such as allow-lists, step limits and human confirmation.
-
How to Fine-Tune an LLM with LoRA and QLoRA on a Single GPU
A practical walkthrough of parameter-efficient fine-tuning: format a chat dataset, train 4-bit QLoRA adapters on Llama 3.1 8B with Hugging Face PEFT and TRL, check VRAM needs, evaluate against the base model, merge the adapter and decide when fine-tuning beats RAG or prompting.
-
Barclays job cuts: 5,000 roles to go, Unite says AI is an excuse
About 1,000 roles have already gone and 4,000 more are due by the end of the fourth quarter, finance director Anna Cross said. Chief executive C. S. Venkatakrishnan called AI "the icing on the cake, not the cake".
-
How to Build a RAG App in Python with Ollama and Chroma
A hands-on, fully local retrieval-augmented generation walkthrough: split your documents into chunks, embed them with Ollama, search them in Chroma and make a local LLM answer with citations, plus chunk-size trade-offs and a simple evaluation checklist.
-
How to build an MCP server in Python: a step-by-step guide
A hands-on guide to building a Model Context Protocol server in Python: set up the official SDK, expose a tool and a resource, test with the MCP Inspector, connect it to Claude and apply the security rules the spec requires.
-
How to run a local LLM with Ollama on macOS, Windows or Linux
A hands-on guide to running a private, offline AI model with Ollama: what hardware you need, how to install it, pick a quantized model, save a custom system prompt and call the local API from curl and Python.
-
OpenAI brings GPT-6 models and Codex to Google Cloud
Customers will pay on a single Google Cloud invoice that can draw down existing spending commitments, and Codex users can already route inference through Google. The deal follows Microsoft's switch to a non-exclusive licence in April 2026.
-
TorchTPU e TorchNeuron: o avanço do PyTorch além da CUDA da Nvidia
Google e Amazon lançaram backends nativos do PyTorch para seus chips de IA na mesma semana. Veja o que um backend faz, o que os desenvolvedores já podem tirar das GPUs da Nvidia e por que o software pode pesar mais do que as especificações dos chips.
-
TorchTPUとTorchNeuronを解説：Nvidia CUDAの先を目指すPyTorch
GoogleとAmazonが同じ週に、自社AIチップ向けのネイティブなPyTorchバックエンドを公開しました。バックエンドの役割、開発者が今NvidiaのGPUから移行できるもの、そしてチップの仕様よりソフトウェアが重要になり得る理由を解説します。
-
TorchTPU dan TorchNeuron: dorongan PyTorch melampaui Nvidia CUDA
Google dan Amazon merilis backend PyTorch native untuk cip AI mereka dalam pekan yang sama. Berikut cara kerja backend, apa yang sudah bisa dipindahkan pengembang dari GPU Nvidia, dan mengapa perangkat lunak bisa lebih menentukan daripada spesifikasi cip.
-
TorchTPU y TorchNeuron: el impulso de PyTorch más allá de Nvidia CUDA
Google y Amazon lanzaron backends nativos de PyTorch para sus chips de IA en la misma semana. Explicamos qué hace un backend, qué pueden migrar hoy los desarrolladores desde las GPU de Nvidia y por qué el software puede pesar más que las especificaciones de los chips.
-
TorchTPU and TorchNeuron explained: PyTorch beyond Nvidia CUDA
Both tools plug into PrivateUse1, PyTorch's official extension point for new hardware, so moving a model can be as simple as changing the device setting. Neither is complete yet: Amazon lists operator gaps, including dynamic shapes.
-
Gemini no carro: o que a Personal Intelligence traz ao motorista
O Google está liberando a Personal Intelligence para o Gemini no Android Auto, e os carros com Google integrado vêm depois. Veja o que o Gemini já faz na estrada, o que a nova personalização acrescenta, quem pode usar e quais configurações de privacidade e segurança importam.
-
車のGeminiが個人向けに：Personal Intelligenceで何が変わるか
GoogleはAndroid AutoのGeminiでPersonal Intelligenceの提供を開始し、Google搭載車にも順次広げる予定です。Geminiが運転中にすでにできること、新たな個人向け機能で加わること、利用できる人、重要なプライバシーと安全の設定をまとめました。
