Zocdoc is opening its digital health platform to AI agents through a partnership with Instinct, a healthcare AI company. Instinct's agents can now reach Zocdoc's network of providers and patients, aiming to deliver more personalized care recommendations. Separately, Athenahealth launched a tool that lets patients pull their health records into Anthropic's Claude AI, giving them personalized answers based on their own clinical data. Both moves reflect a broader trend of healthcare companies connecting patient data to AI assistants.
Biossil Labs, a Canadian startup backed by OpenAI, agreed to buy Theriva Biologics' antibiotic degrader portfolio for up to $91 million. The deal includes an upfront payment of $350,000 and potential milestone payments. The drug compounds are designed to protect patients from further infections during antibiotic treatment, and Biossil plans to fund and advance these deprioritized candidates.
HSBC announced it will launch a Hong Kong AI Institute in 2027, focused on expanding workforce training in artificial intelligence. The initiative aims to prepare employees with AI skills across the organization. The plan was reported on November 2, 2026.
New York lawmakers are advancing strict AI safety rules planned for 2027. State Sen. Andrew Gounardes (D) is leading the effort with a bill called RAISE 2.0, which would limit how quickly the largest AI models can improve themselves. The bill also requires companies to report and monitor security incidents, following events like the Hugging Face cyberattack.
Researchers auditing test-time scaling in world-action models on Cosmos Policy and RoboCasa found that a learned value selector performed 1.66 percentage points worse than random on clean data. Token pruning cut per-forward FLOPs by up to 47% but only achieved a 1.15x end-to-end speedup. The team also discovered and fixed a silent rendering corruption that worsened along compute-heavy planning paths.
A separate study introduced a lightweight forecasting model that tracks how retail customers change preferences over time. Using triangular fuzzy numbers and a Mamba-type selective state space model, a 107-parameter version reached a mean absolute error of 0.154, outperforming both a constant-inertia model and a naive forecast. The model can also identify when a customer permanently switches to a competitor.
Elliptic found no evidence that AI agents directed the laundering of $6.4 billion stolen in the Binance hack. The FBI attributed the October 29 theft to North Korea's TraderTraitor group, which used about 20 automated services to move funds in under three days. About $2.85 billion, or 44 percent of the stolen funds, has been frozen or seized, while hackers still control roughly $3.28 billion across more than 950 wallets.
Shaun Shenouda was named Divisional Chief Operating Officer and AI Transformation Officer of Alliant Underwriting Solutions. He will lead AI and automation efforts while remaining president of SES Risk Solutions, a property and casualty insurance provider. Shenouda holds a master's degree in global finance from Pepperdine University's Graziadio Business School.
Bellingcat studied more than 300 online video ad campaigns backed by AI industry Super PACs ahead of the US 2026 midterm elections. Groups like Leading the Future and Public First Alliance funded ads on economic issues, immigration, and Trump without mentioning AI or their policy agenda. About $52 million has been disbursed by LTF to other Super PACs, and experts warn voters may not realize the interests behind the messaging shaping their choices.
Key Takeaways
- Zocdoc partnered with Instinct to let AI agents connect to its provider and patient network for personalized care recommendations.
- Athenahealth launched a tool allowing patients to bring health records into Anthropic's Claude AI for personalized clinical answers.
- Biossil Labs, backed by OpenAI, agreed to buy Theriva Biologics' antibiotic degrader portfolio for up to $91 million.
- HSBC plans to launch a Hong Kong AI Institute in 2027 focused on workforce AI training.
- New York's RAISE 2.0 bill, led by State Sen. Andrew Gounardes, targets strict AI safety rules for 2027, partly prompted by the Hugging Face cyberattack.
- Researchers found that token pruning cut FLOPs by up to 47% in world-action models but only achieved a 1.15x speedup, while a value selector underperformed random on clean data.
- A new fuzzy forecasting model using Mamba-type architecture achieved a mean absolute error of 0.154 in predicting customer preference shifts.
- Elliptic found no evidence AI agents directed laundering of $6.4 billion in the Binance hack; $2.85 billion has been seized while $3.28 billion remains under hacker control.
- Shaun Shenouda became AI Transformation Officer at Alliant Underwriting Solutions while continuing as president of SES Risk Solutions.
- Bellingcat reported that AI industry Super PACs disbursed about $52 million on campaign ads that did not disclose AI interests to voters.
Zocdoc opens platform to AI agents in healthcare push
Zocdoc is letting AI agents connect to its digital health platform. The company partnered with Instinct, a healthcare AI company. Instinct's AI agents can now reach Zocdoc's network of providers and patients. This aims to give patients more personalized care recommendations.
Athenahealth links health records to Anthropic's Claude AI
Athenahealth launched a tool that lets patients bring their health records into Anthropic's Claude AI. Patients can get personalized answers based on their own clinical data. This move follows a trend of healthcare companies opening records to AI assistants. The tool helps patients take a bigger role in managing their health.
Biossil buys Theriva antibiotic degrader portfolio for up to $91M
Biossil Labs, a Canadian startup backed by OpenAI, agreed to buy Theriva Biologics' antibiotic degrader portfolio for up to $91 million. The deal includes an upfront payment of $350,000 and potential milestone payments. The candidates are drug compounds designed to protect patients from further infections during antibiotic treatment. Biossil plans to fund and advance these deprioritized drug candidates.
HSBC plans Hong Kong AI Institute for 2027
HSBC announced it will launch a Hong Kong AI Institute in 2027. The institute will focus on expanding workforce training in artificial intelligence. The plan was reported on November 2, 2026. HSBC aims to prepare its employees with AI skills through this new initiative.
Study audits when test-time search helps world action models
Researchers audited test-time scaling in world-action models on Cosmos Policy and RoboCasa. They found that a learned value selector performed 1.66 percentage points worse than random on clean data. Token pruning cut per-forward FLOPs by up to 47% but only achieved a 1.15x end-to-end speedup. The team also found and fixed a silent rendering corruption that worsened along compute-heavy planning paths.
Fuzzy Model Predicts How Customer Preferences Change
Researchers built a lightweight forecasting model that tracks how retail customers change their preferences over time. It uses triangular fuzzy numbers and a Mamba-type selective state space model to keep predictions clear and interpretable. A 107-parameter version reached a mean absolute error of 0.154, beating both a constant-inertia model and a naive forecast. The model can also show when a customer permanently switches to a competitor.
New York Democrats Plan Strict AI Rules for 2027
New York lawmakers are moving toward strict AI safety rules planned for 2027. State Sen. Andrew Gounardes (D) is leading the effort with a bill called RAISE 2.0. The bill aims to limit how quickly the largest AI models can improve themselves. It also requires companies to report and monitor security incidents, following events like the Hugging Face cyberattack.
No Evidence AI Helped Launder $6.4B Binance Hack Funds
Elliptic found no proof that AI agents directed the laundering of $6.4 billion stolen in the Binance hack. The FBI attributed the October 29 theft to North Korea's TraderTraitor group, which used about 20 automated services to move funds in under three days. About $2.85 billion, or 44 percent of the stolen funds, has been frozen or seized. The hackers still control roughly $3.28 billion across more than 950 wallets.
Shaun Shenouda Becomes AI Transformation Officer at Alliant
Shaun Shenouda was named Divisional Chief Operating Officer and AI Transformation Officer of Alliant Underwriting Solutions. He will lead efforts in artificial intelligence and automation, focusing on innovation and enterprise execution. He will also remain president of SES Risk Solutions, a property and casualty insurance provider. Shenouda holds a master's degree in global finance from Pepperdine University's Graziadio Business School.
AI Super PACs Secretly Shape Campaign Advertising
Bellingcat studied more than 300 online video ad campaigns backed by AI industry Super PACs ahead of the US 2026 midterm elections. Groups like Leading the Future and Public First Alliance funded ads on economic issues, immigration, and Trump without mentioning AI or their policy agenda. About $52 million has been disbursed by LTF to other Super PACs running political ads. Experts say voters may not realize the interests behind the messaging shaping their choices.
Sources
- Zocdoc opens infrastructure to AI agents, launches partnership with Instinct
- Athenahealth launches tool to bring patient health records into Anthropic's Claude
- OpenAI-funded Biossil snags Theriva's unwanted antibiotic degrader portfolio for up to $91M
- HSBC to Launch Hong Kong AI Institute in 2027, Expand Workforce Training
- When Does Test-Time Search Help World Action Models? An Empirical Audit of Candidate Selection and Simulator Integrity
- Fuzzy selective state model for forecasting the dynamics of consumer preferences
- New York Democrats Plan Country’s Toughest AI Limits in 2027
- Elliptic finds no evidence of AI use in laundering $6.4B Binance loot
- Shaun Shenouda Named Divisional Chief Operating Officer and AI Transformation Officer of Alliant Underwriting Solutions
- Inside the Secretive World of Campaign Ads Funded by AI Super PACs
Comments
Please log in to post a comment.