UNIT.City — місце, де люди працюють... КРАЩЕ! Обирай свій простір просто зараз 👉
Олег ОнопрієнкоAI Eng
11 February 2026, 12:38
2026-02-11
Former GitHub CEO creates startup Entire, which integrates AI agents into Git
Thomas Domke’s new company has officially emerged from stealth mode with the ambitious goal of reimagining the entire software development lifecycle (SDLC) that has become outdated in the AI era. The startup has already raised $60 million in investment.
Thomas Domke’s new company has officially emerged from stealth mode with the ambitious goal of reimagining the entire software development lifecycle (SDLC) that has become outdated in the AI era. The startup has already raised $60 million in investment.
Entire founder Thomas Domke argues that with the advent of powerful agent models such as Claude Code (Opus 4.6), GPT-5.3-Codex from OpenAI, and Cursor Composer 1.5, the rules of the game have changed. Today, development is moving to the terminal, where engineers control an army of agents. However, the infrastructure (Git, Pull Requests, task tracking systems) remains in the era of cloud technologies designed for human-to-human collaboration.
“We still have a development lifecycle built for the cloud era… Git repositories have never scaled to version everything developers create in the AI era. Pull Requests simply don’t scale for large mono-repositories,” the company’s manifesto states .
Entire positions itself as the next platform for developers. Its architecture is based on three components:
A Git-compatible database that unifies code, intent, and logic into a single system.
Semantic reasoning layer for coordinating multi-agent systems.
AI-native SDLC, adapted for collaboration between humans and machines.
Today, the company has released its first open-source tool, Entire CLI with Checkpoints. This solution solves the problem of context loss. Typically, when an AI agent ends a session, its “train of thought” disappears, leaving only the finished code.
How Checkpoints work:
They store the agent context (prompts, session transcript, tokens used, tool calls) as first-class data in Git.
Each commit generated by the agent gets a "checkpoint" attached to it.
This allows developers to see not only what changed in the code (diff), but also why (agent logic).
At launch, the tool supports Anthropic Claude Code and Google Gemini CLI, with Codex and Cursor support coming soon.
“Just as car companies replaced artisanal manufacturing with assembly lines, we must rethink the entire software development cycle for a world where machines are the primary producers of code,” Entire states.
The $60 million investment round was led by Felicis Fund with support from Madrona, M12 (Microsoft's venture capital fund), 20VC, and others. Among the angel investors are well-known names in the tech community: Jerry Yang (co-founder of Yahoo), Harry Tan (president of Y Combinator), Olivier Pommel (CEO of Datadog), and Gergely Oros, author of the popular blog The Pragmatic Engineer.
Чи може AI замінити кодерів: ось які завдання готові віддати штучному інтелекту українські айтішники
GitHub запровадив розширення Copilot, що за допомогою штучного інтелекту може писати код вслід за вже написаним рядком. І це не перша розробка в цьому полі.
Ми запитали розробників, керівників, кодерів та всіх, хто долучений до цієї теми, чи може штучний інтелект замінити людину, що можна довірити машині, та чи не підштовхне це працедавців до звільнень. На опитання відгукнулись 25 респондентів і їх відповіді нас зацікавили.
Штучний інтелект допомагатиме розробникам кодувати. GitHub запустив сервіс Copilot на AI за $10 на місяць
Спеціалісти GitHub від Microsoft розробили та запустили сервіс Copilot, що за допомогою штучного інтелекту додає пропозиції коду на основі попередньої строки або коментаря. Версія загальнодоступна для всіх розробників за $10 на місяць або $100 на рік.