https://promtime.net/c/ru/post/in-claude-code-a-block-hook-could-let-commands-through Claude News ru 2026-10-08T08:30:22.465Z Claude Code 2.1.294 чинит хуки, которые пропускали запреты https://promtime.net/c/ru/post/grok-bot-will-route-tasks-to-claude-opus-5-5-musk-says Claude News ru 2026-10-08T07:00:25.926Z Grok Bot возьмёт Claude Opus 5.5 там, где он лучше https://promtime.net/c/ru/post/claude-s-sdks-run-the-click-loop-but-the-browser-is-yours Claude News ru 2026-10-08T00:14:23.250Z SDK для Claude сами ведут цикл computer use, браузер ваш https://promtime.net/c/ru/post/meta-trims-claude-code-use-yet-spent-over-105m-in-28-days Claude News ru 2026-10-07T21:12:24.805Z Meta и Microsoft отучают своих инженеров от Claude https://promtime.net/c/ru/post/claude-code-2-1-293-switches-its-default-haiku-to-5-5 Claude News ru 2026-10-07T19:42:23.240Z Claude Code 2.1.293 перестал переделывать уже сделанное https://promtime.net/c/ru/post/haiku-5-5-cuts-token-prices-90-on-prompts-under-100k Claude News ru 2026-10-07T18:12:55.197Z Haiku 5.5 решил 39,2% Terminal-Bench, у Haiku 4.5 был ноль https://promtime.net/c/ru/post/anthropic-lobbied-the-pope-on-ai-consciousness-and-lost Claude News ru 2026-10-07T10:44:22.741Z Папа Лев XIV не поверил Anthropic в сознание ИИ https://promtime.net/c/ru/post/claude-code-2-1-292-gives-sub-agents-their-own-effort-dial Claude News ru 2026-10-06T20:38:22.323Z Claude Code 2.1.292 ставит плагин вместе с маркетплейсом https://promtime.net/c/en/post/in-claude-code-a-block-hook-could-let-commands-through Claude News en 2026-10-08T08:30:22.465Z In Claude Code, a "block" hook could let commands through https://promtime.net/c/en/post/grok-bot-will-route-tasks-to-claude-opus-5-5-musk-says Claude News en 2026-10-08T07:00:25.926Z Grok Bot will route tasks to Claude Opus 5.5, Musk says https://promtime.net/c/en/post/claude-s-sdks-run-the-click-loop-but-the-browser-is-yours Claude News en 2026-10-08T00:14:23.250Z Claude's SDKs run the click loop, but the browser is yours https://promtime.net/c/en/post/meta-trims-claude-code-use-yet-spent-over-105m-in-28-days Claude News en 2026-10-07T21:12:24.805Z Meta trims Claude Code use, yet spent over $105M in 28 days https://promtime.net/c/en/post/claude-code-2-1-293-switches-its-default-haiku-to-5-5 Claude News en 2026-10-07T19:42:23.240Z Claude Code 2.1.293 switches its default Haiku to 5.5 https://promtime.net/c/en/post/haiku-5-5-cuts-token-prices-90-on-prompts-under-100k Claude News en 2026-10-07T18:12:55.197Z Haiku 5.5 cuts token prices 90% on prompts under 100k https://promtime.net/c/en/post/anthropic-lobbied-the-pope-on-ai-consciousness-and-lost Claude News en 2026-10-07T10:44:22.741Z Anthropic lobbied the Pope on AI consciousness and lost https://promtime.net/c/en/post/claude-code-2-1-292-gives-sub-agents-their-own-effort-dial Claude News en 2026-10-06T20:38:22.323Z Claude Code 2.1.292 gives sub-agents their own effort dial https://promtime.net/p/ru/post/200-fine-tune-beats-openai-s-decisions-api-on-email-attacks Promtime ru 2026-10-08T17:58:24.692Z В тесте Abnormal дообученная kev-4b обошла фронтирные LLM https://promtime.net/p/ru/post/openai-s-first-category-5-op-hired-unwitting-latin-americans Promtime ru 2026-10-08T17:08:22.736Z OpenAI впервые нашла операцию влияния пятой категории https://promtime.net/p/ru/post/google-s-gemini-agent-can-run-jobs-on-anthropic-s-claude Promtime ru 2026-10-08T15:06:25.616Z Gemini agent от Google гоняет задачи и на моделях Claude https://promtime.net/p/ru/post/openai-closes-in-on-anthropic-in-openrouter-coding-data Promtime ru 2026-10-08T12:32:24.089Z OpenAI добралась почти до половины кодинга на OpenRouter https://promtime.net/p/ru/post/one-sign-error-pulls-three-openai-math-papers Promtime ru 2026-10-08T09:07:24.966Z Одна ошибка в знаке стоила OpenAI трёх статей по математике https://promtime.net/p/ru/post/copilot-on-windows-will-hand-some-tasks-to-local-models Promtime ru 2026-10-07T23:59:12.566Z Copilot на Copilot+ PC начнёт считать часть задач локально https://promtime.net/p/ru/post/claude-haiku-5-5-takes-415-30-seconds-to-say-its-first-word Promtime ru 2026-10-07T23:08:25.833Z Haiku 5.5 думает 415 секунд до первого слова ответа https://promtime.net/p/ru/post/claude-max-subscribers-get-up-to-200-a-month-for-the-api Promtime ru 2026-10-07T19:54:23.569Z В Max 20x входит 200 долларов в месяц на Claude API https://promtime.net/p/ru/post/chatgpt-answers-can-now-be-a-game-or-a-budget-calculator Promtime ru 2026-10-07T19:04:50.074Z ChatGPT начал собирать калькуляторы прямо в ответе https://promtime.net/p/ru/post/short-prompts-to-claude-haiku-5-5-cost-90-less Promtime ru 2026-10-07T18:14:21.525Z Haiku 5.5 подешевела на 90%, но только до 100K токенов https://promtime.net/p/ru/post/openai-s-math-results-cost-about-three-hours-of-pro-each Promtime ru 2026-10-07T00:04:55.102Z OpenAI открыла на GitHub математику невыпущенной модели https://promtime.net/p/ru/post/copilot-wins-github-s-benchmark-ranks-fourth-on-martian-s Promtime ru 2026-10-06T21:32:21.001Z Copilot первый в тесте GitHub и четвёртый у Martian https://promtime.net/p/ru/post/openai-trains-gpt-6-astra-on-real-ironclad-contract-work Promtime ru 2026-10-06T20:42:25.331Z OpenAI тренирует агентов на договорной рутине Ironclad https://promtime.net/p/en/post/200-fine-tune-beats-openai-s-decisions-api-on-email-attacks Promtime en 2026-10-08T17:58:24.692Z $200 fine-tune beats OpenAI's Decisions API on email attacks https://promtime.net/p/en/post/openai-s-first-category-5-op-hired-unwitting-latin-americans Promtime en 2026-10-08T17:08:22.736Z OpenAI's first Category 5 op hired unwitting Latin Americans https://promtime.net/p/en/post/google-s-gemini-agent-can-run-jobs-on-anthropic-s-claude Promtime en 2026-10-08T15:06:25.616Z Google's Gemini agent can run jobs on Anthropic's Claude https://promtime.net/p/en/post/openai-closes-in-on-anthropic-in-openrouter-coding-data Promtime en 2026-10-08T12:32:24.089Z OpenAI closes in on Anthropic in OpenRouter coding data https://promtime.net/p/en/post/one-sign-error-pulls-three-openai-math-papers Promtime en 2026-10-08T09:07:24.966Z One sign error pulls three OpenAI math papers https://promtime.net/p/en/post/copilot-on-windows-will-hand-some-tasks-to-local-models Promtime en 2026-10-07T23:59:12.566Z Copilot on Windows will hand some tasks to local models https://promtime.net/p/en/post/claude-haiku-5-5-takes-415-30-seconds-to-say-its-first-word Promtime en 2026-10-07T23:08:25.833Z Claude Haiku 5.5 takes 415.30 seconds to say its first word https://promtime.net/p/en/post/claude-max-subscribers-get-up-to-200-a-month-for-the-api Promtime en 2026-10-07T19:54:23.569Z Claude Max subscribers get up to $200 a month for the API https://promtime.net/p/en/post/chatgpt-answers-can-now-be-a-game-or-a-budget-calculator Promtime en 2026-10-07T19:04:50.074Z ChatGPT answers can now be a game or a budget calculator https://promtime.net/p/en/post/short-prompts-to-claude-haiku-5-5-cost-90-less Promtime en 2026-10-07T18:14:21.525Z Short prompts to Claude Haiku 5.5 cost 90% less https://promtime.net/p/en/post/openai-s-decisions-api-beats-its-own-engine-by-up-to-10x Promtime en 2026-10-07T03:07:46.536Z OpenAI's Decisions API beats its own engine by up to 10x https://promtime.net/p/en/post/openai-s-math-results-cost-about-three-hours-of-pro-each Promtime en 2026-10-07T00:04:55.102Z OpenAI's math results cost about three hours of Pro each https://promtime.net/p/en/post/copilot-wins-github-s-benchmark-ranks-fourth-on-martian-s Promtime en 2026-10-06T21:32:21.001Z Copilot wins GitHub's benchmark, ranks fourth on Martian's https://promtime.net/p/en/post/openai-trains-gpt-6-astra-on-real-ironclad-contract-work Promtime en 2026-10-06T20:42:25.331Z OpenAI trains GPT-6 Astra on real Ironclad contract work