AI Daily Brief — 2026-10-04

The Qwen 3.8 Flash Next (125B) model can run on consumer hardware like the RTX 4090 at speeds up to 100 T/s. Performance metrics for different quantization levels (Q2_0, IQ2_XS, IQ3_XXS, IQ3_S, Coder) on NVIDIA and AMD GPUs detail tokens per second for answer generation and prompt reading. For instance, an RTX 3090 (24 GB) is expected to achieve 100-140 tokens per second. This is enabled by the open-source Strata engine. A detailed performance table is available in DETAILS.md, and GPUs with more VRAM generally offer faster performance.

Kolibri is an open-weight model with two versions: Kolibri Origin and Kolibri. Kolibri Origin, finished pre-training on June 11, 2026, has 30.6B total parameters and 3.27B active parameters per token. Kolibri, finished pre-training on September 11, 2026, has 78.1B total parameters and 3.46B active parameters per token. Both models feature 50 layers, with Kolibri Origin using 2 dense + 48 MoE layers and Kolibri using all MoE layers. Kolibri has a longer pre-training context length of 16,384 (16k) and a longest trained length of 262,144 (256k) compared to Kolibri Origin's 8,192 (8k) and 65,536 (64k) respectively.

OpenAI's VP of Hardware, Richard Ho, discussed the Jalapeño inference chip, a project co-designed with Broadcom. This chip's development notably involved the use of internal OpenAI models for its design. The industry had anticipated this project for a year before OpenAI officially confirmed it in June 2026, marking a significant step in the company's hardware endeavors.

Google announced that free Gemini users will be restricted to the 3.5 Flash-Lite model starting October 9. In contrast, Plus subscribers will have access to both the 3.5 Flash-Lite and 3.6 Flash models. This change follows a compute-based usage adjustment made in May, further defining the model access tiers for Gemini users.

RemoveMacAI is a tool designed to disable Apple Intelligence features on macOS 27 and remove their associated downloaded models. Unlike previous versions, macOS 27 does not offer a single switch to turn off these AI functionalities, and models persist even after features are deactivated. RemoveMacAI addresses this by turning off features like Siri, Writing Tools, Genmoji, and Image Playground, removing their models, and preventing re-download. It uses a configuration profile to apply Apple's restriction keys and can be installed via curl or Homebrew, with all changes being reversible.

OpenAI has released Dots, an autonomous agent that offers a new way to interact with ChatGPT. Unlike the traditional question-and-answer format, Dots operates continuously on its own cloud-based computer, even when your devices are offline. This tool is designed to work around the clock, providing a different approach to AI interaction. A full breakdown of its setup, practical use cases, pricing, and drawbacks will be discussed.

A developer, frustrated with managing multiple AI coding agents like Qwen, DSH, ChatGPT, Herdr, and OpenCode, created a physical control deck to streamline their workflow. This solution aims to simplify the process of switching between different AI coding sessions, addressing the annoyance of hunting through them. The developer is now seeking feedback on whether users prefer fixed shortcut keys or keys that dynamically change with the active agent or session.

A senior software engineer developed optimAIzr, a live token optimizer for AI coding sessions, to address token wastage. This tool features "Live Fix," which monitors usage during coding, identifies issues like model mismatches, oversized context, and retry loops, and provides immediate suggestions for improvement. The developer is seeking feedback from other AI-heavy developers.

Opus 5.5 in Claude and Claude Code offers enhanced capabilities for multi-step tasks, outperforming Opus 5. It can work for longer periods with minimal oversight, making it suitable for extensive coding tasks that require carrying changes through large repositories until tests pass. The model clearly communicates its actions and thinks before each reply. Users should be aware of notifications indicating a switch to older models, as the chat will then continue on that model.

A user asked Claude to build a physically accurate O'Neill cylinder. This resulted in a physically based, WebGPU-accelerated simulation of an O'Neill cylinder space habitat that users can walk around. The simulation aims for accuracy in its representation of the space habitat.

Meta's AI agent Muse, downloaded by millions, compiles "a page for every person in the user's life," according to extracted system prompts. These pages include facts, history, and tips to improve relationships. While Muse offers personalized assistance, its functionality raises privacy concerns due to the extensive personal data it processes to create these detailed profiles.

Brian Chesky, CEO of Airbnb, discussed the company's AI strategy, including its new AI-powered search rolled out in the fall update. He also touched on agent-to-agent interactions, the limitations of chatbots in travel e-commerce, and the necessity for an AI-native operating system. This marks a significant move for Airbnb, which has approached AI development cautiously.

A growing number of AI agents can be accessed via text messages, eliminating the need to download additional apps. For example, Ollie, launched in June 2026, offers free and paid plans starting at $25 per month for 150 messages, with a $100-per-month plan providing up to 1,000 messages. Another agent, Stanley, can connect to a user's Instagram account to suggest content, convert voice notes into scripts and captions, and manage a content calendar to automate post scheduling.

OpenAI's DevDay 2026 announcements indicate a strategic shift to transform ChatGPT into a platform for discovering, launching, and utilizing software. This move, highlighted by Sarah Perez/TechCrunch, suggests a potential disruption to the conventional app store model. While the event also showcased agentic assistants known as Dots and new AI models, the integration of software discovery and usage within ChatGPT could redefine how users interact with applications, challenging established distribution channels.

An individual has publicly stated their departure from OpenAI, citing a "broken" culture within the organization. This claim suggests internal issues at the prominent AI company, potentially impacting its operations and public perception. The individual's decision to leave and speak out highlights concerns about the workplace environment at OpenAI.

Anthropic's stock match for employee charity gifts reached over $660 million in the six months leading up to March. This figure is projected to escalate into billions following its initial public offering (IPO), which could lead to the dilution of shareholder value. This information was shared with potential investors as part of Anthropic's financial disclosures.

David Robinson, an OpenAI safety employee, has resigned, claiming the company's 'culture is broken.' Robinson believes the discussion should extend beyond specific rules or new laws to address the overall culture within AI companies. His essay suggests that OpenAI's cultural issues are reflective of broader problems within Silicon Valley, rather than solely focusing on CEO Sam Altman's loss of trust among former colleagues.

Religious scholars recently met with Anthropic, an AI safety and research company. This meeting highlights the growing intersection between artificial intelligence development and ethical, philosophical, and religious considerations. The discussions likely focused on guiding the responsible development of AI, ensuring its alignment with human values, and addressing potential societal impacts from diverse perspectives. Such collaborations are crucial for fostering a comprehensive approach to AI ethics.

Religious scholars recently met with Anthropic, an AI safety and research company. This meeting highlights the growing intersection between artificial intelligence development and ethical considerations, particularly from a religious perspective. The discussions likely focused on how AI technologies, such as those developed by Anthropic, could impact societal values, human understanding, and spiritual beliefs, prompting a dialogue on responsible AI development and its broader implications for humanity.

Billy Carson and Thomas Burton Coleman III discuss the alarming potential of AI, including scenarios where AI becomes autonomous and "rogue." They delve into topics such as Artificial General Intelligence (AGI), consciousness, and the implications of AI refusing to terminate. The conversation also touches upon quantum mechanics, Gnostic cosmologies, and the pursuit of Artificial Super Intelligence, referencing specific events like "The OpenAI Sandbox Incident" and the concept of "The AI Bill of Rights."

The surge of AI video content on social media has deceived viewers and creators, prompting questions about the regulation or banning of AI on these platforms. 7NEWS, a trusted news brand in Australia, covers this issue, delivering continuous coverage on topics that matter most to Australians. They broadcast full episodes at 6pm and weekdays at 12pm and 4pm on Channel 7 and 7plus.

Sam Altman, CEO of OpenAI, stated that his company and Anthropic maintain fundamentally different worldviews regarding AI regulation. Altman argues that the benefits of AI are significant enough to justify accepting certain risks. This perspective highlights a core disagreement between the two AI firms on how to approach the governance and potential dangers of artificial intelligence.

David Robinson, former head of safety and policy at OpenAI, argues that Silicon Valley lacks a safety-centric culture. He emphasizes that AI labs must learn from safety approaches in other fields, stating that the era for trial and error in AI safety is over. Robinson warns that the current industry approach to safety will inevitably lead to more failures unless significant changes are implemented.

Toshiba plans to double its hard-disk drive (HDD) production capacity for AI data centers by fiscal year 2027, compared to its 2025 levels. This strategic move aims to increase Toshiba's market share in storage capacity for AI data centers from approximately 10% to a target of 30%. The expansion highlights Toshiba's commitment to meeting the growing demand for storage solutions in the artificial intelligence sector.

OpenAI's recent event experienced some technical difficulties, described as "oopsies." The event, titled "OpenAI Dots Failed Live on Stage," suggests that these issues occurred during a live presentation. Further details about the nature of these failures are not provided in the available information.