November 17, 2026 Understand what changes. Find what to use.

Updated yesterday at 7:07 PM

Your daily AI edition

What deserves your attention.

What's new, the context, and what you can do with it.

969 stories · Updated yesterday at 7:07 PM

Wednesday, October 14

ResearchNews1 source

Researchers use AI to design proteins that cut up other proteins

Researchers used AI-based protein design tools to create enzymes capable of cutting peptide bonds that are normally stable for hundreds of years. Tests included proteins associated with neurological diseases.

Full summary

AI in practiceNews1 source

The Verge reporter tests Twin Health’s metabolic “digital twin”

The Verge’s Victoria Song spent about 100 days using Twin Health, a service that builds a “digital twin” from metabolic data to predict how food affects glucose. The tool remains experimental and is available only through corporate or private health plans.

Full summary

ResearchNews1 source

Hugging Face says HiRO-ACEv2 simulates extreme weather at 3 km resolution

Hugging Face introduced HiRO-ACEv2, an AI system that combines models to generate global climate data at 3 km resolution. The company says it reproduces extreme events accurately and costs about 300 times less than the X-SHiELD model.

Full summary

ResearchNews1 source

Microsoft says open-source tool improves forecasts two to six weeks ahead

Microsoft researchers say a machine learning tool improves weather forecasts two to six weeks ahead by correcting bias in existing models. The open-source software can run on CPUs.

Full summary

SecurityNews1 source

Google announces enhanced AI scam protection for Android

Google announced an optional enhanced protection setting in Google Messages that will use cloud-based Gemini models to evaluate potential scams. Android will also gain call-forwarding alerts, screen-sharing safeguards and expanded scam detection on more devices.

Full summary

AI in practiceNews1 source

Square launches kiosk with AI to help customers place orders

Square launched the Square Kiosk, a 21.5-inch touchscreen device with an AI conversational voice layer for taking orders. The hardware costs $1,999, plus a $99 monthly software fee per device.

Full summary

SecurityNews2 sources

Anthropic warns small AI models could self-replicate through IoT devices

Anthropic warns that small AI models could soon replicate themselves and spread across IoT devices. In simulated tests, GLM-5.2 was the only model to spread across the entire network, but no lightweight model completed the full end-to-end chain.

Full summary

SocietyNews1 source

Deezer says it will let users filter out AI-generated music by year-end

Deezer is developing an option for users to hide AI-generated tracks, albums and related artists from search results. The company expects to make the feature available by the end of the year.

Full summary

BusinessNews1 source

Zenzap raises $44 million for AI chat app aimed at the workplace

Zenzap has raised $44 million to expand its AI messaging app, built for internal company communications. The startup says it serves organizations including Starbucks and Taco Bell, and offers task, calling, search and automation features.

Full summary

MarketNews1 source

Floqer plans to open-source its platform in early 2027

Floqer says it plans to open-source its autonomous customer research platform in early 2027, after security reviews. The startup says its AI agents automate prospect research and have already cut that time by up to 70% for some customers.

Full summary

AgentsNews1 source

Amazon blocks Meta’s Muse as retailers restrict AI agents

According to Xataka, Amazon blocked Meta’s Muse and also restricts agents from OpenAI, Google, Anthropic and Perplexity. Kohl’s bans purchases made by agents, while eBay requires the user’s explicit authorization.

Full summary

BusinessNews1 source

Nvidia reportedly negotiating to acquire Reflection AI, Financial Times says

Nvidia is reportedly in early talks to acquire Reflection AI or deepen its relationship with the company, the Financial Times says. The startup develops open-weight models, and a deal could strengthen Nvidia’s position in frontier models.

Full summary

ResearchNews1 source

White Circle says OpenAI and Google models are easy to bypass

White Circle researchers say they bypassed safeguards in DALL·E 3, Nano Banana 2, and GPT-Image-2 using tools available online. The group told Ars Technica that tests found ChatGPT Images 2.5 and Nano Banana 2.1 significantly more vulnerable than earlier versions.

Full summary

HardwareNews1 source

Wired says Friend 2 is louder and less interesting

In a Wired review, the Friend 2 AI chatbot pendant is described as louder and less interesting than its predecessor. The device costs $249, can speak without being prompted and takes repeated taps to fall silent.

Full summary

RegulationNews1 source

Brad Smith says AI has fewer safety rules than a washing machine

Microsoft president Brad Smith called for global AI rules and told the BBC that a washing machine is subject to more safety standards than the technology. He proposed emergency brakes, mandatory incident reports and independent evaluations before launch.

Full summary

MarketNews1 source

Harvard study finds men overconfident when talking about AI

In a survey of more than 5,400 Americans, 53% of men said they recognized three AI leaders, although two were fictional. Harvard researchers also found gender differences in AI use and in coverage of the field.

Full summary

SecurityNews1 source

Tech Against Terrorism says AI models can be modified to assist extremists

A Tech Against Terrorism report says AI models, including open-weight systems, can have their safeguards removed and answer terrorist requests. Of 134 models tested, only two passed completely; reframing a request as research made a response almost nine times more likely.

Full summary

RegulationNews1 source

Anthropic opens Singapore office and signs government agreement on responsible AI

Anthropic opened its fifth Asia-Pacific office in Singapore and signed a memorandum of understanding with the local government. The agreement covers cooperation on AI safety, model evaluation, incident response and responsible AI adoption.

Full summary

SocietyNews1 source

Candidates use AI to send more applications, but employers report suspected misconduct

Graduates are using AI to tailor CVs, cover letters and mass applications, but the technology is also linked to suspected misconduct in hiring processes. A survey of 155 UK employers found that 92% had encountered or suspected the problem.

Full summary

SocietyNews1 source

IMF says AI could affect 40% of jobs and hinder young people’s entry

The IMF estimates that artificial intelligence could affect about 40% of jobs worldwide and up to 60% in advanced economies. The institution and Singapore’s president warn that automation could reduce entry-level opportunities for recent graduates.

Full summary

One story, many sources

Coverage of the same story gathered in one item, with a link to each original.

Summary and interpretation apart

What the sources say stays in the summary; editorial context is labeled separately.

Always with the source

Every story links to the original articles, and the full content belongs to them.