Terms we explain
The companies, products, people and ideas in our stories, each explained in a few plain sentences, with its sources. Background from ai notis, not part of the news.
Companies
- AkamaiAkamai Technologies is a company that provides security, cloud computing and delivery solutions. Its delivery solutions are mainly for web and mobile performance and for delivering media such as video streaming and software downloads. It was incorporated in Delaware in 1998, and its shares trade on the Nasdaq under the symbol AKAM.
- AMDAMD, short for Advanced Micro Devices, is a company that makes semiconductors, the chips used in computers and data centres. Its products include EPYC server CPUs, Instinct GPUs for data centres and Ryzen processors for desktops. It was founded on 1 May 1969.
- AnthropicAnthropic is an AI company that describes itself as an AI safety and research company. It is a public benefit corporation, a company whose stated purpose includes benefit to the public. It makes Claude, an AI assistant.
- GitHubGitHub is a platform for building software. It builds on Git, a version control system that tracks changes to files, by hosting Git projects in the cloud and adding planning and collaboration tools for teams.
- GoogleGoogle is the largest of the businesses in Alphabet, the company that owns it. Alphabet's annual report says its mission is to organize the world's information and make it universally accessible and useful. Google's core products include Search, YouTube, Android, Chrome, Gmail, Google Maps and Gemini.
- OpenAIOpenAI is a company that builds AI models, which developers can use through the OpenAI API. It makes ChatGPT and Codex, its coding agent for software development tasks.
Products
- AI OverviewsAI Overviews are a Google Search feature that writes a short summary of a topic with AI, with links to learn more. Google says they are for when you want a quick answer without having to piece together all the information yourself.
- ChatGPTChatGPT is a product from OpenAI. You use it in chats, exchanging messages with it, and its documentation says you can start with a goal, idea or task, and ChatGPT can gather context, take action and produce something useful. It is available on the web, in a desktop app and on mobile.
- ClaudeClaude is an AI assistant made by Anthropic. People use it to draft documents, analyse data from several sources and work with files and connected apps. It runs on the web, in phone and desktop apps, and through an API for developers.
- Claude CodeClaude Code is a coding tool from Anthropic. It reads a project's code, edits files and runs commands on a developer's behalf. It works in a terminal, in code editors, in a desktop app and in a web browser.
- CodexCodex is OpenAI's coding agent for software development tasks. It runs in the ChatGPT desktop app, as a command line tool, in code editor extensions, on the web, in cloud environments and through an SDK.
- DockerDocker is an open platform for developing, shipping and running applications. It packages an application and what it needs into a container, a lightweight, loosely isolated environment. The same container can run on a laptop, in a data centre or in the cloud.
- GitHub CopilotGitHub Copilot is an AI coding assistant built into GitHub. It suggests code as a developer types, answers questions in a chat and explains code. It can also be given a GitHub issue to work on as an agent.
- Google GeminiGemini is the name Google uses for its AI models and for an app. Google's developer documentation says the Gemini API lets developers build Gemini and other generative models into applications to generate text and images, analyze multimodal inputs and build conversational agents. Google DeepMind's page for the models also lists a Gemini app, where you can chat with Gemini.
- KimiKimi is an AI assistant and a family of AI models made by Moonshot AI. Developers can use the models through the Kimi API, which works with the OpenAI and Anthropic API formats. The models cover general tasks, coding and understanding images.
- SynthIDSynthID is a watermarking tool made by Google DeepMind. It adds a mark that people cannot see or hear to AI-generated images, video, audio and text, so the content can later be identified as AI-made or altered. Google uses it in products including Gemini, where an image, video or audio clip can be uploaded to check for it.
Ideas
- benchmarkIn AI, a benchmark is a standard test used to measure how well an AI model or system performs. Results let people compare different models on the same test. One example is SWE-bench Verified, which Epoch AI describes as measuring models' ability to resolve realistic GitHub issues.
- context windowA context window is the amount of text an AI language model can look back on and use when it writes a reply. It works like the model's working memory. It is separate from the large body of data the model was trained on.
- Model Context ProtocolMCP is an open-source standard for connecting AI applications to outside systems, such as files, databases, search engines and other tools. Its own documentation compares it to a USB-C port: one standard connection instead of a different one for every device. Claude, ChatGPT and Visual Studio Code are among the tools that support it.
- sandboxingSandboxing means running software in a restricted, controlled environment, so that it cannot reach system resources it is not authorised to use. NIST's glossary says an application in a sandbox is usually restricted from accessing the file system or the network.
- tokenA token is the small unit of text that an AI language model reads and writes. It can be a whole word, part of a word, a single character or even a byte. For Claude, one token is about 3.5 English characters, though that varies with the language.