AI Daily brief · Wednesday, 4 November 2026
STARHOT
MIT experiment: GPT-6 Astra directly modifies Qwen3.5 weights, mathematical ability improves by about 6 times
MIT's experiment allowed GPT-6 Astra to directly read and edit the weights of Qwen3.5-0.8B, changing only 0.016% of the parameters, without training and gradient descent, MATH-500 score improved from 7.9% to 48.1%, and instruction execution ability also increased 3 times.
Product launches & updates
Google DeepMindOfficial+1 more outletsGoogle Antigravity launches cheminformatics skills, completing drug discovery process in one hour
Google DeepMind releases a set of cheminformatics skills (Cheminformatics Skills) for Google Antigravity, based on RDKit and integrated with ChEMBL and PubChem, where language models are responsible for process orchestration and scientific reasoning, and deterministic chemical tools perform molecular calculations.
The Verge · AI+1 more outlets
Apple has opened up Siri AI capabilities to third-party apps in iOS 27.3
Apple has opened up new Siri AI capabilities to developers with iOS 27.3, allowing third-party apps to access personal context, natural language prompts, and screen awareness to complete tasks. This update was released in Developer Preview form and adds app skills that can be created by developers, combining multiple tasks into a single action.
OpenAI NewsOfficial+2 more outlets
OpenAI launched the Grid Protectors Program
OpenAI's Grid Protectors Program will collaborate with utilities to help future computing campuses support grid reliability and protect consumer electricity costs.
Industry
TechCrunch · AI+3 more outletsFollow-up
Anthropic warns of nearing RSI capabilities in the revised S-1
Anthropic, in the revised S-1A risk factors section submitted on Monday, states that its internal estimates show it may soon meet the definition of 'rapid acceleration of progress brought by AI' in its Responsible Scaling Policy, but has not yet designated any existing models as exceeding that threshold.
- Anthropic's internal AI usage, when calculated at retail prices, has exceeded employee salaries.Hacker News · AI hot posts
Hacker News · AI hot posts+4 more outlets
Google will shut down search services for schools that refuse AI.
According to The New York Times, Google will shut down search services for schools that refuse to adopt its AI. The post on Hacker News received 259 points and 208 comments, with the discussion link pointing to Google's official blog page for educational products.
The Decoder
OpenAI claims GPT-6 Astra designed a hardware device from scratch in less than a week.
OpenAI states its hardware team used GPT-6 Astra to design a device that can communicate with Codex encoding agents as a side project, completed in less than a week. Astra handled all aspects including casing design, PCB layout with component selection and routing, firmware, and user interface. Contract manufacturers produced small batches of prototypes in less than a month, but due to Astra not always considering manufacturing constraints, tolerances, and assembly details, several rounds of modifications were required.
OpenAI NewsOfficial+2 more outletsOpenAI appoints Dan Driscoll as senior advisor for national security modernization
OpenAI appoints Dan Driscoll as senior advisor for national security modernization, responsible for providing advice on responsible AI, cyber resilience, and collaboration with US allies.
Research papers
Anthropic ResearchOfficial+2 more outlets
Anthropic has released the verifiable reasoning system FlashProve, generating proofs at approximately 110 tokens/second on a single H100.
Anthropic announced the verifiable reasoning prototype FlashProve, using zkSNARK to prove that the response indeed comes from the committed model weights. For the 400B parameter Llama 4 Maverick (4-bit quantization), a single H100 generates proofs at approximately 110 tokens/second, with proofs for each 1,024 token request being about 19 MB, and verification on a 24-core CPU taking about 17 seconds.
Hugging Face BlogOfficialThe same model with a different harness: Comparing CUGA, Hermes, and Deep Agents on AppWorld.
The IBM team compared CUGA, Hermes, and Deep Agents using the same open-source weight model GLM-5.3-flash on 90 tasks in AppWorld, with pass rates of 81.1%, 46.7%, and 30.0%, respectively, and 20, 8, and 2 scenarios where all three variants passed.
More worth reading
OpenHands BlogOpenHands and Snowflake introduce a workflow for training models using their own Agent harnessOpenHands partnered with Snowflake AI Research, allowing developers to directly use the existing OpenHands Agent harness for reinforcement learning and training without rebuilding the harness in the training framework.
Where's Your Ed AtOpenAI's funding struggles, the AI industry's capital surge phenomenon is reversingOpenAI completed a $11 billion funding round in early November 2026 with a valuation of $125 billion, despite lukewarm interest from the capital markets, but it has revealed increased difficulty in subsequent funding. This round of investment reflects investors' doubts about OpenAI's revenue growth ($6.7 billion in the second quarter of 2026) and cost control, as the company's Q1 2026 revenue costs reached $12.1 billion.
Tomasz Tunguz2027 AI's financial engineering key elements: GPU costs double and coping strategiesGPU rental prices have doubled in the past 6 months, Anthropic's GPU equivalent hourly cost has increased from $3.10 to $5.65, forcing AI providers to switch to long-term contracts to fix rents between 2 and 4 dollars per hour
ITHome · AITencent will launch an AI-powered search engine for C-end users: minimalist design, enhanced authoritative source ranking- Hacker News · AI hot postsProfessor Kate Devlin talks about AI resistance lists: using complaints to fight against poor AI applications.Professor Kate Devlin discussed the AI resistance list at EMFCamp, calling for public complaints about poor AI applications. She used the UK online supermarket Ocado as an example, where customers complained about AI-generated customer service responses, ultimately leading the company to commit to passing the issues to the technical team. Devlin emphasized that AI responses lack a human touch and contain grammatical errors, while human customer service's apologies and commitments to solving problems satisfied her, believing organizations should respect customer time and avoid using AI responses for real complaints.
- ITHome · AIDevelopers flash Apple iOS 16.5 test firmware pack for Xiaomi REDMI K20 ProDeveloper 'You say you like it but you're lying' has flashed the Apple iOS 16.5 test firmware package for Xiaomi REDMI K20 Pro phone, supporting dual systems and data persistence, requiring Android data to be cleared and TWRP Recovery to be replaced. It can switch systems through normal startup or Magisk module under Root privileges, but the developer emphasizes that this solution is a test version and carries the risk of bricking.
