General Resolution: LLM usage in Debian
https://www.debian.org/vote/2026/vote_002
https://www.debian.org/vote/2026/vote_002
ClickHouse: сценарии, сильные стороны, лучшие практики работы в 2026 году
https://habr.com/ru/companies/vktech/articles/1061284/
https://habr.com/ru/companies/vktech/articles/1061284/
Английский вместо кода
https://habr.com/ru/articles/1058500/
https://habr.com/ru/articles/1058500/
👍1
The new rules of context engineering for Claude 5 generation models
https://claude.com/blog/the-new-rules-of-context-engineering-for-claude-5-generation-models
https://claude.com/blog/the-new-rules-of-context-engineering-for-claude-5-generation-models
Claude
The new rules of context engineering for Claude 5 generation models | Claude by Anthropic
We removed over 80% of Claude Code's system prompt for more advanced models. How to apply the lessons we learned to your own context engineering in Claude Code and with your own agents.
JEP 540: Simple JSON API (Now in Incubator)
https://openjdk.org/jeps/540
https://openjdk.org/jeps/540
Running Kimi K3 on a M1 Mac
https://github.com/gavamedia/deltafin
https://github.com/gavamedia/deltafin
GitHub
GitHub - gavamedia/deltafin: Run full Kimi K3 on a single device. And an OpenAI-compatible API server for local chat and coding…
Run full Kimi K3 on a single device. And an OpenAI-compatible API server for local chat and coding agents. - gavamedia/deltafin
Я чего-то пропустил с 23й версии такое благолепие
https://docs.oracle.com/en/java/javase/25/javadoc/using-markdown-documentation-comments.html
https://docs.oracle.com/en/java/javase/25/javadoc/using-markdown-documentation-comments.html
Why Software Factories Fail (or: harness engineering is not enough)
https://github.com/humanlayer/advanced-context-engineering-for-coding-agents/blob/main/wsff.md
https://github.com/humanlayer/advanced-context-engineering-for-coding-agents/blob/main/wsff.md
GitHub
advanced-context-engineering-for-coding-agents/wsff.md at main · humanlayer/advanced-context-engineering-for-coding-agents
Contribute to humanlayer/advanced-context-engineering-for-coding-agents development by creating an account on GitHub.
Context rot: как умирают ИИ-агенты в продакшене и как их спасти
https://tproger.ru/articles/context-rot-kak-umirayut-ii-agenty-v-prodakwene-i-kak-ih-spasti
https://tproger.ru/articles/context-rot-kak-umirayut-ii-agenty-v-prodakwene-i-kak-ih-spasti
JEP 401: Value Objects Officially Proposed to Target JDK 28
https://www.reddit.com/r/java/comments/1v507d0/jep_401_value_objects_officially_proposed_to/
https://www.reddit.com/r/java/comments/1v507d0/jep_401_value_objects_officially_proposed_to/
Reddit
From the java community on Reddit: JEP 401: Value Objects Officially Proposed to Target JDK 28
Posted by Joram2 - 145 votes and 11 comments
Forwarded from IT-architecture RSS (RSS to Telegram Bot)
The Economic Benefit of Refactoring
Giles Edwards-Alexander does an experiment to see if decomposing a large function helps reduce token costs, suggesting that is may now be possible to measure the economic benefit of refactoring
more…
via Martin Fowler (author: Martin Fowler (martin@martinfowler.com))
Giles Edwards-Alexander does an experiment to see if decomposing a large function helps reduce token costs, suggesting that is may now be possible to measure the economic benefit of refactoring
more…
via Martin Fowler (author: Martin Fowler (martin@martinfowler.com))
Лучшая модель этой недели - от рекламы пахнуло МММ и Хопер-инвестом 😁
Please open Telegram to view this post
VIEW IN TELEGRAM
Forwarded from Zavtracast (Ярослав Ивус)
Alibaba выпустила новый флагманский ИИ — Qwen 3.8 Max. Это модель на 2,4 трлн параметров, которая может без остановки работать в течение 10 дней.
В рекламе Alibaba показывает, как различные специалисты отдыхают, а Qwen 3.8 Max самостоятельно всё делает в фоне. Например, разрабатывает чипы в течение 12 часов.
В некоторых бенчмарках Qwen 3.8 Max обходит Claude Fable 5 и ChatGPT-5.6 Sol. Модель уже доступна на сайте — на следующей неделе должны выложить веса в открытый доступ.
@zavtracast
В рекламе Alibaba показывает, как различные специалисты отдыхают, а Qwen 3.8 Max самостоятельно всё делает в фоне. Например, разрабатывает чипы в течение 12 часов.
В некоторых бенчмарках Qwen 3.8 Max обходит Claude Fable 5 и ChatGPT-5.6 Sol. Модель уже доступна на сайте — на следующей неделе должны выложить веса в открытый доступ.
@zavtracast
👍1
Новый выпуск и опять в гости пришёл Костя из сетей Yandex Cloud https://razborpoletov.mave.digital/ep-322
❤2👍2🔥1
Forwarded from Denis Pavlyuchenko
ох, какая же насущная проблема. Интересно, как ее возможно щас побороть вообще, что все корпоративное общение сводится к копипасте из модели https://gruhn.me/blog/2026-08-03/
What's the largest software project AI can complete on its own?
https://epoch.ai/MirrorCode
https://epoch.ai/MirrorCode
Epoch AI
MirrorCode: What's the largest software project AI can complete on its own?
MirrorCode is Epoch AI's benchmark for long-horizon coding: AI can reimplement entire programs end-to-end, with no access to the original source code.
Zitron: The Subprime Datacenter Crisis
https://www.wheresyoured.at/the-subprime-data-center-crisis/
https://www.wheresyoured.at/the-subprime-data-center-crisis/
Ed Zitron's Where's Your Ed At
The Subprime Data Center Crisis
If you liked this piece, you should subscribe to my premium newsletter.
It's $70 a year, $18 a quarter, or $7 a month, and in return you get a weekly newsletter that’s usually anywhere from 10,000 to 18,000 words, including vast, detailed analyses of NVIDIA…
It's $70 a year, $18 a quarter, or $7 a month, and in return you get a weekly newsletter that’s usually anywhere from 10,000 to 18,000 words, including vast, detailed analyses of NVIDIA…