Channel
@figure_media
On this record: Topic · Growth · Engagement · Reactions · Posts · Cite this entry
725subscribers
-10 since we began measuring on 7 August 2026
Risers and fallers across the register · movement among entries of Under 1,000.
| Telegram ID | -1001866435566 |
|---|---|
| Type | Channel |
| Username | @figure_media |
| Created | Between 1 October 2022 and 30 September 2023 — estimated from Telegram’s id allocation, not measured. How this range is calculated. |
| First recorded | 8 August 2026 |
| Last confirmed live | 4 October 2026 |
| Measurements held | 6 |
| Confirmed unchanged | 3 times, most recently 4 October 2026 |
| On Telegram | t.me/figure_media |
Technology — a classification, not a measurement. An on-box language model (Qwen3.6-35B-A3B-FP8, prompt version 1) read this channel’s own recent posts on 20 September 2026 and assigned it the closest of 31 fixed categories, at 100% confidence. This is a model’s judgement about what the channel is likely to be about, not a fact this register measured the way a subscriber count or a view count is measured — it can be revised on a later pass, and it carries no weight anywhere else on this page. How this classification works, and why it has no browse page of its own yet.
| Measured (UTC) | Subscribers | Change |
|---|---|---|
| 10 Sept 2026, 11:00 | 725 | -4 |
| 31 Aug 2026, 13:54 | 729 | -3 |
| 24 Aug 2026, 15:39 | 732 | -2 |
| 16 Aug 2026, 03:09 | 734 | -1 |
| 8 Aug 2026, 02:03 | 735 | no change |
| 7 Aug 2026, 14:04 | 735 | first reading |
20 posts held, back to 6 May 2024 — the reader has not yet reached the start of this channel’s public history, so older posts may sit further back, unread. Read across 1 page of Telegram’s post history, 20 posts per page.
Nothing published in the last 30 days. ERR and ER are rolling 30-day measures, so there is nothing to compute — we hold 20 posts for this entry, the most recent from 16 February 2026. An engagement rate over an empty window would be a number about nothing.
34 reactions across 6 posts, in 7 distinct kinds. The most used accounts for 52.9% of them.
| Reaction | Count | Share | Share, drawn |
|---|---|---|---|
| ❤ | 18 | 52.9% | |
| 👍 | 8 | 23.5% | |
| 👎 | 3 | 8.82% | |
| 🦄 | 2 | 5.88% | |
| 🎉 | 1 | 2.94% | |
| 💯 | 1 | 2.94% | |
| 🔥 | 1 | 2.94% |
No sentiment is inferred, and none should be read in. This table is ordered by count and by nothing else. Emoji do not carry stable meaning across languages or communities — 🙏 is thanks in one channel and mourning in another — so we publish which ones were pressed and how often, and pass no judgement on what an audience meant by them.
Precision. Telegram publishes reaction counts per emoji and short-forms each one — 4.34K, 1.2M — so any single kind at or above 1,000 reaches us at three significant figures, and only counts below 1,000 are exact. The shares above are ratios of those figures and carry the same error. This is also why the total here can differ slightly from a reaction total printed elsewhere on the page: both are sums of the same rounded parts, taken over samples with different edges.
Coverage. Reactions were read on 6 of the 20 sampled posts in this sample. Summed by Telegram’s own count on each post — not by adding up the per-emoji breakdown above — those same posts carry 34 reactions in total: the kind of figure the paragraph above means by “a reaction total printed elsewhere on the page”.
Measured over the 20 most recent posts we hold, published 6 May 2024 to 16 February 2026, using the newest reading held for each. Telegram Stars are excluded: they are a payment, not a reaction, and they have their own section.
NVIDIA добавила AutoDeploy в TensorRT‑LLM (beta) — автооптимизация инференса без ручного переписывания моделей. — компилирует PyTorch/Hugging Face модели в оптимизированные графы — автоматом решает KV‑кеш, шардирование, фьюзинг, выбор kernels — заявлена поддержка 100+ LLM (ранняя поддержка VLM/SSM) Почему важно: быстрее выводить новые/«длиннохвостые» архитектуры в прод без недель инженерии. Источник + подробнее: h…
AWS вывела Claude Sonnet 4.5 в GovCloud: Claude Code можно подключить к Bedrock для регулируемых проектов Коротко: • Claude Sonnet 4.5 стал доступен в AWS GovCloud (US) через Amazon Bedrock (в т.ч. cross‑Region inference). • Claude Code можно настроить на Bedrock и использовать в терминале/IDE — для команд с требованиями FedRAMP High / DoD IL4/IL5 / ITAR. • В материале — шаги настройки и рекомендации по орг-внедрени…
NVIDIA: GPU помогают «рулить» научными экспериментами в реальном времени — анализ сократили с месяцев до часов • Кейсы: обсерватория Vera C. Rubin (LSST) и SLAC LCLS‑II • В статье заявляют ускорение анализа до 9 месяцев → ~4 часа за счёт GPU и распределённой обработки Почему важно: когда обработка становится «почти реальной», можно менять параметры эксперимента по ходу, быстрее ловить редкие события и эффективнее и…
NVIDIA: Isaac Lab — open-source GPU-симулятор для масштабного мультимодального обучения роботов Единый стек (физика/рендер/сенсоры) на GPU, тысячи параллельных сред и sim-to-real workflow. https://figure.media/nvidia-isaac-lab-open-source-gpu-simulyator-dlya-masshtabnogo-multimodalnogo-obucheniya-robotov/
Meta и Hugging Face показали, как тестировать агентные ИИ «в проде»: OpenEnv + Calendar Gym Ключевая мысль: главные провалы агентов — многошаговые процессы, двусмысленность и ошибки в аргументах/порядке действий. https://figure.media/meta-i-hugging-face-pokazali-kak-testirovat-agentnye-ii-v-prode-openenv-calendar-gym/
AWS объяснила, как запускать «долгие» MCP-задачи: Bedrock AgentCore + Strands Agents AWS разобрала два паттерна для long-running задач в MCP: progress/heartbeat через context messaging и async-подход с task ID, фоновым выполнением и восстановлением состояния между сессиями. https://figure.media/aws-obyasnila-kak-zapuskat-dolgie-mcp-zadachi-bedrock-agentcore-strands-agents/ Материал подготовлен с помощью ИИ.
Hugging Face сделала skill для агентов: они пишут CUDA-ядра под PyTorch (и это работает) Коротко: Hugging Face показала skill «cuda-kernels», который помогает кодовым агентам (Claude/Codex) генерировать CUDA-ядра и биндинги для PyTorch + бенчмарки. В демо — ускорения RMSNorm до ~2.5× и ~6% end-to-end на пайплайне. https://figure.media/hugging-face-sdelala-skill-dlya-agentov-oni-pishut-cuda-yadra-pod-pytorch-i-eto-r…
AWS усилила браузер для AI-агентов в Bedrock AgentCore: прокси, профили и расширения Теперь можно: маршрутизировать трафик через корпоративные прокси, сохранять cookies/local storage между сессиями и подключать Chrome-расширения. https://figure.media/aws-usilila-brauzer-dlya-ai-agentov-v-bedrock-agentcore-proksi-profili-i-rasshireniya/
Google обновила Gemini 3 Deep Think — режим для сложных задач в науке и инженерии Ключевое: крупное обновление «режима рассуждений», доступ в Gemini для подписчиков AI Ultra + заявка на ранний доступ через Gemini API. https://figure.media/google-obnovila-gemini-3-deep-think-rezhim-dlya-slozhnyh-zadach-v-nauke-i-inzhenerii/
llama.cpp b7974: в CMake добавили опцию не устанавливать тесты при упаковке. https://figure.media/llama-cpp-b7974-v-cmake-dobavili-opcziyu-ne-ustanavlivat-testy/
llama.cpp выпустили релиз b7981: поправили крайний кейс в chat-шаблонах и подчистили логи. https://figure.media/llama-cpp-reliz-b7981-ispravlyaet-obrabotku-shablonov-chata/
Google и YouTube обновили инструменты безопасности для детей и подростков https://figure.media/google-i-youtube-obnovili-instrumenty-bezopasnosti-dlya-detej-i-podrostkov/
Showing the 12 most recent of 20 posts we hold for @figure_media. View and reaction counts are the latest single reading for each post, not a live figure, and a recent post is still accumulating both. A view count marked ≈ was rounded by Telegram before we ever saw it — t.me prints views in full below 1,000 and to three significant figures above, so ≈1,200,000 means somewhere between 1,150,000 and 1,249,999. Unmarked counts are exact. Text is reproduced from the public post preview and truncated for length.
A live page changes as we take new readings, so a citation should name the measurement it is based on, not just the URL. The line below cites the subscriber count as measured 10 September 2026 — this entry's latest reading, not the date you are reading this.
“Фигура — всё о digital” (@figure_media), 725 subscribers as measured 10 September 2026. Telegram Register, tgregister.com/channel/figure_media.
Full measurement history, CC BY 4.0. Every reading this register holds for this entry, not just the latest one, as a dated, downloadable record: CSV · JSON. Free to use with attribution to tgregister.com. Each file carries its own generation timestamp, which is the figure to cite for exactly when the data was retrieved.