OpenAI launches GPT-5.6 family with tiered pricing and 54 percent higher token efficiency
OpenAI has released the GPT-5.6 model family, featuring the flagship Sol, the mid-range Terra, and the high-speed Luna models. These models share a 1.5 million-token context window and introduce specialized "max" and "ultra" modes for handling complex, multi-step reasoning tasks. The ultra configuration is particularly notable for its ability to deploy parallel sub-agents to accelerate demanding workflows like advanced coding and biological analysis.
Source
👉@sysadminoff
https://4sysops.com/archives/openai-launches-gpt-5-6-family-with-tiered-pricing-and-54-percent-higher-token-efficiency/
OpenAI has released the GPT-5.6 model family, featuring the flagship Sol, the mid-range Terra, and the high-speed Luna models. These models share a 1.5 million-token context window and introduce specialized "max" and "ultra" modes for handling complex, multi-step reasoning tasks. The ultra configuration is particularly notable for its ability to deploy parallel sub-agents to accelerate demanding workflows like advanced coding and biological analysis.
Source
👉@sysadminoff
https://4sysops.com/archives/openai-launches-gpt-5-6-family-with-tiered-pricing-and-54-percent-higher-token-efficiency/
Starbucks uses AI-assisted coding to replace enterprise software vendors with in-house software
Starbucks is leveraging artificial intelligence to develop in-house software alternatives to applications currently provided by Microsoft and IBM. The initiative specifically targets a Microsoft inventory tracking system and an IBM maintenance management tool for replacement by the end of next year. This shift is part of a strategic $2 billion cost-cutting program aimed at reducing the company's $400 million annual software expenditure.
Source
👉@sysadminoff
https://4sysops.com/archives/starbucks-uses-ai-assisted-coding-to-replace-enterprise-software-vendors-with-in-house-software/
Starbucks is leveraging artificial intelligence to develop in-house software alternatives to applications currently provided by Microsoft and IBM. The initiative specifically targets a Microsoft inventory tracking system and an IBM maintenance management tool for replacement by the end of next year. This shift is part of a strategic $2 billion cost-cutting program aimed at reducing the company's $400 million annual software expenditure.
Source
👉@sysadminoff
https://4sysops.com/archives/starbucks-uses-ai-assisted-coding-to-replace-enterprise-software-vendors-with-in-house-software/
SharePoint Copilot Apps
Microsoft announced the public preview of SharePoint Copilot Apps. This feature, shipping as part of the SharePoint Framework (SPFx) version 1.24 preview, lets developers build interactive user interface components that appear directly inside the Copilot canvas. The Copilot canvas is the chat interface where Microsoft 365 Copilot operates. Instead of receiving only text responses, users can interact with data grids, forms, charts, dashboards, and approval panels without leaving the conversation. This article covers what SharePoint Copilot Apps are, how they work, what you need to deploy them, and the limitations you should be aware of.
Source
👉@sysadminoff
https://4sysops.com/archives/sharepoint-copilot-apps/
Microsoft announced the public preview of SharePoint Copilot Apps. This feature, shipping as part of the SharePoint Framework (SPFx) version 1.24 preview, lets developers build interactive user interface components that appear directly inside the Copilot canvas. The Copilot canvas is the chat interface where Microsoft 365 Copilot operates. Instead of receiving only text responses, users can interact with data grids, forms, charts, dashboards, and approval panels without leaving the conversation. This article covers what SharePoint Copilot Apps are, how they work, what you need to deploy them, and the limitations you should be aware of.
Source
👉@sysadminoff
https://4sysops.com/archives/sharepoint-copilot-apps/
GPT-5.6 Sol challenges Claude Fable 5 with superior coding efficiency and lower costs
OpenAI has introduced GPT-5.6 Sol, a flagship model that nearly matches the performance of Anthropic's Claude Fable 5 while significantly reducing operational expenses. The model achieved a top score on the Coding Agent Index and demonstrated high proficiency in realistic office tasks through the AA-Briefcase benchmark. Although it remains in a restricted preview for vetted organizations, it is already being integrated into development environments like the Codex client.
Source
👉@sysadminoff
https://4sysops.com/archives/gpt-5-6-sol-challenges-claude-fable-5-with-superior-coding-efficiency-and-lower-costs/
OpenAI has introduced GPT-5.6 Sol, a flagship model that nearly matches the performance of Anthropic's Claude Fable 5 while significantly reducing operational expenses. The model achieved a top score on the Coding Agent Index and demonstrated high proficiency in realistic office tasks through the AA-Briefcase benchmark. Although it remains in a restricted preview for vetted organizations, it is already being integrated into development environments like the Codex client.
Source
👉@sysadminoff
https://4sysops.com/archives/gpt-5-6-sol-challenges-claude-fable-5-with-superior-coding-efficiency-and-lower-costs/
OpenAI sunsets Atlas browser to focus on ChatGPT desktop app and agents
OpenAI is discontinuing ChatGPT Atlas, its standalone desktop browser designed to perform web-based tasks on behalf of users. The company has set a target deprecation date of August 9, 2026, for the experimental software. This move marks the end of a short-lived project that first arrived on the Mac in October to test agentic browsing capabilities.
Source
👉@sysadminoff
https://4sysops.com/archives/openai-sunsets-atlas-browser-to-focus-on-chatgpt-desktop-app-and-agents/
OpenAI is discontinuing ChatGPT Atlas, its standalone desktop browser designed to perform web-based tasks on behalf of users. The company has set a target deprecation date of August 9, 2026, for the experimental software. This move marks the end of a short-lived project that first arrived on the Mac in October to test agentic browsing capabilities.
Source
👉@sysadminoff
https://4sysops.com/archives/openai-sunsets-atlas-browser-to-focus-on-chatgpt-desktop-app-and-agents/
📰 Initial Patches Posted For Booting The Apple M4 On Linux
With the Linux 7.2 kernel there is initial support for booting the Apple M3 SoC on Linux but it's not yet functional for end users with just booting to a simple console. There are now Device Tree files posted for booting the Apple M4 on Linux but also not yet useful for any typical Apple Mac/MacBook usage on Linux...
🔗 Source:
#kernel #linux
👉@sysadminoff
https://www.phoronix.com/news/Apple-M4-DT-Linux
With the Linux 7.2 kernel there is initial support for booting the Apple M3 SoC on Linux but it's not yet functional for end users with just booting to a simple console. There are now Device Tree files posted for booting the Apple M4 on Linux but also not yet useful for any typical Apple Mac/MacBook usage on Linux...
🔗 Source:
#kernel #linux
👉@sysadminoff
https://www.phoronix.com/news/Apple-M4-DT-Linux
Phoronix
Initial Patches Posted For Booting The Apple M4 On Linux
With the Linux 7.2 kernel there is initial support for booting the Apple M3 SoC on Linux but it's not yet functional for end users with just booting to a simple console
Testing GPT-5.6 and comparing it with Anthropic Fable
OpenAI’s GPT-5.6 model family demonstrates significant advancements in autonomous task execution through integrated "computer use" and "browser use" capabilities. Testing reveals the model can independently operate desktop applications and web browsers to achieve high-level goals, such as recreating complex software like Excel or Minecraft from scratch. For technical operations, the model successfully executes multi-step workflows including complex DNS record modifications and email management via a single natural language prompt.
Source
👉@sysadminoff
https://4sysops.com/archives/testing-gpt-5-6-and-comparing-it-with-anthropic-fable/
OpenAI’s GPT-5.6 model family demonstrates significant advancements in autonomous task execution through integrated "computer use" and "browser use" capabilities. Testing reveals the model can independently operate desktop applications and web browsers to achieve high-level goals, such as recreating complex software like Excel or Minecraft from scratch. For technical operations, the model successfully executes multi-step workflows including complex DNS record modifications and email management via a single natural language prompt.
Source
👉@sysadminoff
https://4sysops.com/archives/testing-gpt-5-6-and-comparing-it-with-anthropic-fable/
Testing OpenAI GPT-5.6 and comparing it with Anthropic Fable
The latest testing of GPT-5.6 reveals a significant leap in autonomous software engineering, capable of recreating complex applications like Excel and Minecraft through multi-day development cycles. By utilizing "computer use" features, the model can interact directly with desktop environments and browsers to perform technical operations such as modifying DNS records or sorting Gmail. This evaluation highlights a tiered architecture—Luna, Terra, and Sol—that allows for strategic model routing to balance high-level reasoning with cost efficiency.
Source
👉@sysadminoff
https://4sysops.com/archives/testing-openai-gpt-5-6-and-comparing-it-with-anthropic-fable/
The latest testing of GPT-5.6 reveals a significant leap in autonomous software engineering, capable of recreating complex applications like Excel and Minecraft through multi-day development cycles. By utilizing "computer use" features, the model can interact directly with desktop environments and browsers to perform technical operations such as modifying DNS records or sorting Gmail. This evaluation highlights a tiered architecture—Luna, Terra, and Sol—that allows for strategic model routing to balance high-level reasoning with cost efficiency.
Source
👉@sysadminoff
https://4sysops.com/archives/testing-openai-gpt-5-6-and-comparing-it-with-anthropic-fable/
📰 GodDamn Ransomware Uses PoisonX Driver to Disable Endpoint Defenses
Cybersecurity researchers have flagged a new ransomware family called GodDamn that employs the PoisonX kernel driver to neutralize security software as part of its defense evasion strategy.According to a new report published by the Threat Hunter Team from Symantec, the ransomware was first publicly spotted in the wild on May 21, 2026. It's assessed to be a rebrand of the Beast ransomware,.
🔗 Source: https://thehackernews.com/2026/07/goddamn-ransomware-uses-poisonx-driver.html
#kernel
👉@sysadminoff
Cybersecurity researchers have flagged a new ransomware family called GodDamn that employs the PoisonX kernel driver to neutralize security software as part of its defense evasion strategy.According to a new report published by the Threat Hunter Team from Symantec, the ransomware was first publicly spotted in the wild on May 21, 2026. It's assessed to be a rebrand of the Beast ransomware,.
🔗 Source: https://thehackernews.com/2026/07/goddamn-ransomware-uses-poisonx-driver.html
#kernel
👉@sysadminoff
📰 Linux Prepares For New USB-C Security Feature On Lenovo ThinkPads
Newer Lenovo ThinkPad systems feature a security feature called USB-C Security Restricted Mode that is in the process of being wired up for reporting under Linux...
🔗 Source:
#linux
👉@sysadminoff
https://www.phoronix.com/news/ThinkPad-USB-C-Restricted-Mode
Newer Lenovo ThinkPad systems feature a security feature called USB-C Security Restricted Mode that is in the process of being wired up for reporting under Linux...
🔗 Source:
#linux
👉@sysadminoff
https://www.phoronix.com/news/ThinkPad-USB-C-Restricted-Mode
Phoronix
Linux Prepares For New USB-C Security Feature On Lenovo ThinkPads
Newer Lenovo ThinkPad systems feature a security feature called USB-C Security Restricted Mode that is in the process of being wired up for reporting under Linux.
📰 Wireshark 4.6.7 Released with Updated Protocol Support, Bug and Security Fixes
Wireshark 4.6.7 open-source network protocol analyzer is now available for download with updated protocol and capture file support, as well as various bug and security fixes.
🔗 Source: https://9to5linux.com/wireshark-4-6-7-released-with-updated-protocol-support-bug-and-security-fixes
#opensource
👉@sysadminoff
Wireshark 4.6.7 open-source network protocol analyzer is now available for download with updated protocol and capture file support, as well as various bug and security fixes.
🔗 Source: https://9to5linux.com/wireshark-4-6-7-released-with-updated-protocol-support-bug-and-security-fixes
#opensource
👉@sysadminoff
pgrust - клон PostgreSQL на Rust, проходящий все регрессионные тесты
Опубликован проект pgrust с реализацией СУБД PostgreSQL, переписанной на языке Rust c использованием транслятора c2rust и AI-ассистента. Заявлено, что проект пока не готов для использования в рабочих проектах, но уже успешно проходит все регрессионные тесты PostgreSQL, совместим с форматом хранения данных PostgreSQL 18.3 и может использовать data-каталог от PostgreSQL 18.3). Код проекта распространяется под лицензией AGPLv3, в отличие от оригинального PostgreSQL, поставляемого под пермиссивной лицензией PostgreSQL, близкой к лицензиям BSD и MIT. Доступна online-демонстрация возможностей, основанная на сборке в WebAssembly, а также docker-образ для быстрого запуска.
👉@sysadminoff
https://www.opennet.ru/opennews/art.shtml?num=65879
Опубликован проект pgrust с реализацией СУБД PostgreSQL, переписанной на языке Rust c использованием транслятора c2rust и AI-ассистента. Заявлено, что проект пока не готов для использования в рабочих проектах, но уже успешно проходит все регрессионные тесты PostgreSQL, совместим с форматом хранения данных PostgreSQL 18.3 и может использовать data-каталог от PostgreSQL 18.3). Код проекта распространяется под лицензией AGPLv3, в отличие от оригинального PostgreSQL, поставляемого под пермиссивной лицензией PostgreSQL, близкой к лицензиям BSD и MIT. Доступна online-демонстрация возможностей, основанная на сборке в WebAssembly, а также docker-образ для быстрого запуска.
👉@sysadminoff
https://www.opennet.ru/opennews/art.shtml?num=65879
Microsoft pauses Secure Boot certificate updates due to firmware compatibility issues
Microsoft has confirmed that Secure Boot certificate updates are failing or being blocked on a subset of Windows 11 PCs. This issue stems from specific firmware configurations where applying the new 2023 certificates could trigger boot failures or BitLocker recovery screens. As a precaution, the rollout has been paused for affected devices while Microsoft works with hardware manufacturers to develop necessary firmware patches.
Source
👉@sysadminoff
https://4sysops.com/archives/microsoft-pauses-secure-boot-certificate-updates-due-to-firmware-compatibility-issues/
Microsoft has confirmed that Secure Boot certificate updates are failing or being blocked on a subset of Windows 11 PCs. This issue stems from specific firmware configurations where applying the new 2023 certificates could trigger boot failures or BitLocker recovery screens. As a precaution, the rollout has been paused for affected devices while Microsoft works with hardware manufacturers to develop necessary firmware patches.
Source
👉@sysadminoff
https://4sysops.com/archives/microsoft-pauses-secure-boot-certificate-updates-due-to-firmware-compatibility-issues/
📰 MINISFORUM MS-03 is a compact Panther Lake PC with 10 Gigabit LAN and a PCIe slot
Chinese mini PC maker MINISFORUM unveiled its first mini PC with an Intel Core Ultra (Series 3) “Panther Lake” processor during CES: the upcoming MINISFORUM M2 Pro is a small computer with support for up to an Intel Core Ultra X9 388H processor, 12-core Intel Arc B390 graphics, and up to 96GB of LPDDR5x-9600 memory. .
🔗 Source:
#intel
👉@sysadminoff
https://liliputing.com/minisforum-introduces-ms-03-mini-pc-with-intel-panther-lake/
Chinese mini PC maker MINISFORUM unveiled its first mini PC with an Intel Core Ultra (Series 3) “Panther Lake” processor during CES: the upcoming MINISFORUM M2 Pro is a small computer with support for up to an Intel Core Ultra X9 388H processor, 12-core Intel Arc B390 graphics, and up to 96GB of LPDDR5x-9600 memory. .
🔗 Source:
#intel
👉@sysadminoff
https://liliputing.com/minisforum-introduces-ms-03-mini-pc-with-intel-panther-lake/
Liliputing
MINISFORUM introduces MS-03 mini PC with Intel Panther Lake
MINISFORUM introduces MS-03 mini PC with Intel Panther Lake
69% of enterprises allow AI agents to share credentials
A recent survey of 107 organizations reveals that 69% of enterprises allow AI agents to share credentials, such as API keys or service-account tokens. This practice creates a significant security flaw where a single compromised agent grants an attacker the combined permissions of every agent using that key. Furthermore, the use of shared identities eliminates forensic accountability because logs cannot distinguish which specific agent performed an action.
Source
👉@sysadminoff
https://4sysops.com/archives/69-of-enterprises-allow-ai-agents-to-share-credentials/
A recent survey of 107 organizations reveals that 69% of enterprises allow AI agents to share credentials, such as API keys or service-account tokens. This practice creates a significant security flaw where a single compromised agent grants an attacker the combined permissions of every agent using that key. Furthermore, the use of shared identities eliminates forensic accountability because logs cannot distinguish which specific agent performed an action.
Source
👉@sysadminoff
https://4sysops.com/archives/69-of-enterprises-allow-ai-agents-to-share-credentials/
Leaked web‑native Copilot OS and 4,800 Microsoft layoffs
Microsoft is internally testing an experimental, web-native operating system prototype known as Project Aion. This "Copilot OS" replaces the traditional Windows desktop environment and Start menu with an interface built entirely inside the Microsoft Edge browser. The platform-agnostic design focuses on agentic AI and task-based organization rather than local software execution.
Source
👉@sysadminoff
https://4sysops.com/archives/leaked-web%e2%80%91native-copilot-os-and-4800-microsoft-layoffs/
Microsoft is internally testing an experimental, web-native operating system prototype known as Project Aion. This "Copilot OS" replaces the traditional Windows desktop environment and Start menu with an interface built entirely inside the Microsoft Edge browser. The platform-agnostic design focuses on agentic AI and task-based organization rather than local software execution.
Source
👉@sysadminoff
https://4sysops.com/archives/leaked-web%e2%80%91native-copilot-os-and-4800-microsoft-layoffs/
Десятый экспериментальный выпуск среды рабочего стола Orbitiny
Опубликован десятый выпуск среды рабочего стола Orbitiny Desktop, написанной с нуля с использованием фреймворка Qt. Проект пытается совместить некоторые инновационные идеи, которые раньше не встречались в пользовательских окружениях, с традиционными элементами, такими как панель с поддержкой плагинов, меню приложений и рабочий стол, на котором можно размещать ярлыки. Работа пока сосредоточена на запуск в окружениях на базе X-сервера, но в будущем не исключается добавление поддержки Wayland. Код написан на языке C++ и распространяется под лицензией GPL.
👉@sysadminoff
https://www.opennet.ru/opennews/art.shtml?num=65881
Опубликован десятый выпуск среды рабочего стола Orbitiny Desktop, написанной с нуля с использованием фреймворка Qt. Проект пытается совместить некоторые инновационные идеи, которые раньше не встречались в пользовательских окружениях, с традиционными элементами, такими как панель с поддержкой плагинов, меню приложений и рабочий стол, на котором можно размещать ярлыки. Работа пока сосредоточена на запуск в окружениях на базе X-сервера, но в будущем не исключается добавление поддержки Wayland. Код написан на языке C++ и распространяется под лицензией GPL.
👉@sysadminoff
https://www.opennet.ru/opennews/art.shtml?num=65881
Samsung develops Gaia AI accelerator to offload generative tasks in future PCs
Samsung is developing a dedicated AI accelerator codenamed Gaia to enhance the performance of next-generation personal computers. Built on a 4-nanometer process, this neural processing unit is designed specifically to offload generative AI and inference workloads from the primary CPU and GPU. The chip aims to improve on-device processing speeds for real-time language tasks and image generation without relying on cloud-based resources.
Source
👉@sysadminoff
https://4sysops.com/archives/samsung-develops-gaia-ai-accelerator-to-offload-generative-tasks-in-future-pcs/
Samsung is developing a dedicated AI accelerator codenamed Gaia to enhance the performance of next-generation personal computers. Built on a 4-nanometer process, this neural processing unit is designed specifically to offload generative AI and inference workloads from the primary CPU and GPU. The chip aims to improve on-device processing speeds for real-time language tasks and image generation without relying on cloud-based resources.
Source
👉@sysadminoff
https://4sysops.com/archives/samsung-develops-gaia-ai-accelerator-to-offload-generative-tasks-in-future-pcs/
Managing CodexSandboxUser accounts and profile errors in Windows 11
OpenAI Codex creates local security principals such as CodexSandboxUser, CodexSandboxOffline, and CodexSandboxOnline to execute AI-generated commands in an isolated environment. This architecture prevents potentially unsafe code from accessing the primary user's files or system settings by running tasks with limited privileges. These accounts are legitimate security features and should remain on the system as long as the Codex Desktop tool or related CLI utilities are in use.
Source
👉@sysadminoff
https://4sysops.com/archives/managing-codexsandboxuser-accounts-and-profile-errors-in-windows-11/
OpenAI Codex creates local security principals such as CodexSandboxUser, CodexSandboxOffline, and CodexSandboxOnline to execute AI-generated commands in an isolated environment. This architecture prevents potentially unsafe code from accessing the primary user's files or system settings by running tasks with limited privileges. These accounts are legitimate security features and should remain on the system as long as the Codex Desktop tool or related CLI utilities are in use.
Source
👉@sysadminoff
https://4sysops.com/archives/managing-codexsandboxuser-accounts-and-profile-errors-in-windows-11/
📰 KDE Plasma users face a dire omen of change: 6.6.6 arrives
6.7 is now current, and in 6.8 you're getting Wayland whether you like it or not.
🔗 Source:
#kde #plasma #wayland
👉@sysadminoff
https://www.theregister.com/os-platforms/2026/07/09/kde-plasma-users-face-a-dire-omen-of-change-666-arrives/5269267
6.7 is now current, and in 6.8 you're getting Wayland whether you like it or not.
🔗 Source:
#kde #plasma #wayland
👉@sysadminoff
https://www.theregister.com/os-platforms/2026/07/09/kde-plasma-users-face-a-dire-omen-of-change-666-arrives/5269267
theregister
KDE Plasma users face a dire omen of change: 6.6.6 arrives
6.7 is now current, and in 6.8 you're getting Wayland whether you like it or not
📰 Wayland 1.26 RC1 Released With New Event To Help Ensure Correct Pointer Coordinates
In addition to Weston 16 nearing release and its release candidate out today, the Wayland 1.26 release candidate was just issued with a few notable changes on top of the more typical bug fixing...
🔗 Source:
#wayland
👉@sysadminoff
https://www.phoronix.com/news/Wayland-1.26-RC1
In addition to Weston 16 nearing release and its release candidate out today, the Wayland 1.26 release candidate was just issued with a few notable changes on top of the more typical bug fixing...
🔗 Source:
#wayland
👉@sysadminoff
https://www.phoronix.com/news/Wayland-1.26-RC1
Phoronix
Wayland 1.26 RC1 Released With New Event To Help Ensure Correct Pointer Coordinates
In addition to Weston 16 nearing release and its release candidate out today, the Wayland 1.26 release candidate was just issued with a few notable changes on top of the more typical bug fixing.