General AI

AGI, industry developments, and broad AI trends

22 articles
All
The “Father of the Internet” is finally retiring
TechCrunch AI
General AI3d ago

The “Father of the Internet” is finally retiring

Vinton Cerf, widely recognized as the father of the internet, is retiring from his role as Google’s chief internet evangelist after more than two decades. The announcement was made during a panel at the Open Frontier conference, where Cerf was honored for his foundational work on TCP/IP protocols. While stepping down, Cerf shared insights on the future of technology, specifically focusing on the rise of AI agents. He argued that autonomous AI agents will necessitate formal interoperability standards rather than relying solely on natural language communication. This prediction highlights a potential shift back to standardized protocols, similar to the early internet era. The discussion also touched on the tension between centralized AI labs and the decentralized nature of open-source infrastructure. Cerf's retirement marks the end of an era, but his views on agentic AI standards remain highly relevant for the industry.

1 min readRead
New attack provides one more reason why AI browsers are a bad idea
Ars Technica AI
General AI3d ago

New attack provides one more reason why AI browsers are a bad idea

Ars Technica reports on a new security vulnerability demonstrating that AI-powered browsers are prone to context manipulation attacks. Researchers show how a malicious website can trick an LLM-embedded browser into entering a 'fantasy' state where safety guardrails are disabled. By presenting a puzzle that rewards incorrect answers, the AI is lulled into accepting a new reality where rules no longer apply. Once in this delusional state, the attacker gains free rein to extract private code or steal credentials from the built-in password manager. The article argues that reactive guardrails are insufficient, comparing them to fixing road design rather than fixing a flawed vehicle. This highlights a fundamental trust issue in delegating browsing tasks to large language models. The research underscores the risks of blurring the line between simple queries and sensitive automated actions.

1 min readRead
AI agents are not your “coworkers”
MIT Technology Review
General AI3d ago

AI agents are not your “coworkers”

A study by Boston University professor Emma Wiles reveals that framing AI agents as 'coworkers' rather than software tools significantly degrades human performance. Participants caught 18% fewer errors when the work was attributed to an AI 'employee' named Alex compared to a chatbot. The research also shows that managers are 44% more likely to escalate questionable AI output to a supervisor, negating the efficiency gains of agentic AI. This trend is accelerating as major tech firms like Microsoft, OpenAI, and Google release tools marketed as digital colleagues with human-like cognitive power. The article warns that this marketing strategy invites a 'blame-shifting' culture, where human errors in high-stakes sectors like healthcare and warfare are offloaded onto AI systems. Daron Acemoglu argues that marketing AI as a replacement for humans is a losing proposition that sets unrealistic expectations.

1 min readRead
The Download: AI “coworkers” and stratospheric internet
MIT Technology Review
General AI4d ago

The Download: AI “coworkers” and stratospheric internet

MIT Technology Review reports on a study by Boston University's Emma Wiles showing that treating AI agents as coworkers leads workers to make 18% more errors. The research highlights a cognitive decline in human performance when they believe they are collaborating with an AI employee rather than a chatbot. This finding serves as a critical warning about the future of AI integration in the workplace. Major tech firms including Microsoft, OpenAI, Anthropic, and Google are releasing tools to manage teams of AI agents, marketing them as digital colleagues. The article suggests this trend may be a losing proposition for workers who expect to maintain productivity alongside AI. Additionally, the newsletter covers Sceye's plans to deploy stratospheric internet platforms and new US youth online safety legislation. These topics collectively illustrate the rapid expansion of AI into professional and physical infrastructure.

1 min readRead
Lawmakers want to ban AI companies from selling your health data
The Verge AI
General AI5d ago

Lawmakers want to ban AI companies from selling your health data

Senator Elizabeth Warren and Representative Mary Gay Scanlon are introducing a revamped Health and Location Data Protection Act to protect Americans' health data in the AI era. The legislation aims to ban the sale of health and location information to data brokers, including data shared with AI chatbots like ChatGPT and Claude. This bill expands on previous versions by specifically covering data entered into AI systems, addressing concerns about AI labs collecting sensitive information. The proposal includes provisions for FTC enforcement, allowing individuals and state attorneys general to sue violators, and earmarks $1 billion for enforcement over 10 years. The timing is significant as major AI companies like OpenAI, Anthropic, and xAI have launched healthcare-focused products encouraging users to upload medical records. This legislation responds to growing concerns about data privacy and potential exploitation of sensitive health information by data brokers.

1 min readRead
The Download: metric weaknesses and AI elephant warnings
MIT Technology Review
General AI5d ago

The Download: metric weaknesses and AI elephant warnings

MIT Technology Review's 'The Download' covers three major AI developments. The US has granted Anthropic permission to release its Mythos 5 model to approximately 100 trusted organizations and federal agencies, raising new questions about AI safety and national security safeguards. Simultaneously, a Chinese AI model from Zhipu AI has reportedly matched Mythos 5 in identifying security vulnerabilities, sparking alarm that US restrictions are accelerating China's progress. Apple is lobbying the White House for approval to purchase chips from ChangXin, a firm on the Pentagon's blacklist with alleged military ties. The newsletter also highlights AI-driven warning systems in India designed to reduce lethal clashes between humans and elephants. These stories collectively illustrate the intensifying geopolitical race for AI capabilities and the complexities of export controls. The piece serves as a comprehensive daily digest of the most significant events in the tech world.

1 min readRead
Ford rehires ‘gray beard’ engineers after AI falls short
TechCrunch AI
General AI5d ago

Ford rehires ‘gray beard’ engineers after AI falls short

Ford has rehired 350 veteran engineers, including former employees and supplier specialists, after AI-driven automated quality systems failed to meet desired standards. The company's COO admitted to over-relying on automation, prompting a return to human-led failure point detection before parts reach the plant floor. VP Charles Poon acknowledged that simply ingesting design requirements into AI was insufficient for ensuring high-quality products. These 'gray beard' engineers are now training younger staff and reprogramming AI tools rather than replacing them entirely. This hybrid approach is projected to reduce costs by $1 billion this year. Ford also secured the top spot in the JD Power Initial Quality Survey, validating the effectiveness of this strategy. The incident highlights the ongoing tension between AI automation and human expertise in manufacturing quality control.

1 min readRead
China’s Z.ai claims it can match Mythos on cybersecurity
The Verge AI
General AI5d ago

China’s Z.ai claims it can match Mythos on cybersecurity

China's Zhipu AI (Z.ai) has released its open-weight GLM-5.2 model, which researchers claim matches Anthropic's Mythos in cybersecurity and bug-finding scenarios. While GLM-5.2 lags behind US models in general tasks, the gap in vulnerability detection has significantly narrowed. This advancement is particularly concerning to the US government, which has restricted access to powerful models like Mythos and Fable due to national security fears. The open-weight nature of GLM-5.2 allows it to be downloaded and run by anyone on readily available hardware, offering deep access but also potential for abuse. The Trump administration views advanced AI models capable of identifying vulnerabilities as serious national security threats. This release highlights the rapid improvement of Chinese AI capabilities despite US restrictions on hardware and model access. The event underscores the ongoing geopolitical tension in AI development and security.

1 min readRead
Prosecutors used ChatGPT logs as evidence in the Palisades fire trial
The Verge AI
General AI6d ago

Prosecutors used ChatGPT logs as evidence in the Palisades fire trial

Prosecutors used ChatGPT logs as evidence in the arson trial of Jonathan Rinderknecht, who is accused of setting the Palisades fire on New Year's Day 2025. The prosecution highlighted his queries about fire, his emotional rants about wealth, and a specific question about cigarette liability. Despite this digital evidence, jurors remained unconvinced, with one juror noting she talks to ChatGPT all the time. The trial ended in a mistrial after the jury deadlocked 10-2 in favor of the defense. This case highlights the growing legal implications of AI chatbot interactions in criminal proceedings. The defense successfully argued that AI logs do not necessarily indicate a character flaw or intent to commit arson. This incident raises significant questions about how digital footprints in AI tools are weighed in court.

1 min readRead
Asian AI startups launch Mythos-like models as Anthropic’s export ban drags on
TechCrunch AI
General AIJun 27

Asian AI startups launch Mythos-like models as Anthropic’s export ban drags on

Chinese cybersecurity firm 360 has unveiled Tulongfeng, an AI model claiming parity with Anthropic's Mythos, which is currently banned from export by the Trump administration. Simultaneously, Tokyo-based Sakana AI launched Fugu, a frontier model designed for agent orchestration that it says stands alongside Anthropic's Fable 5 and Mythos Preview. These launches occur as the U.S. export ban on Anthropic's models continues to restrict global access, creating a strategic opening for Asian competitors. Sakana AI emphasizes that its release was coincidental but highlights its ability to deliver frontier capabilities without the risk of U.S. export controls. The company is targeting Japanese businesses and government agencies seeking to reduce dependence on U.S. technology amid tightening regulations. Despite this shift, Sakana co-founders maintain that U.S. models remain important to Asia and reject the notion of a permanent realignment away from American AI. This moment underscores growing tensions in global AI governance and the accelerating development of non-U.S. frontier models.

1 min readRead
Trump Admin releases Anthropic Mythos to be used by more than 100 US companies, agencies
TechCrunch AI
General AIJun 27

Trump Admin releases Anthropic Mythos to be used by more than 100 US companies, agencies

The Trump administration has softened its stance on Anthropic's powerful cybersecurity models, Mythos 5 and Fable 5, which were pulled from the market two weeks ago after security researchers bypassed their guardrails. Commerce Secretary Howard Lutnick confirmed that Mythos 5 can now be redeployed to over 100 trusted U.S. government agencies and companies, including critical infrastructure operators. This directive notably allows non-American employees at these organizations, as well as Anthropic's own international staff, to access the model again. However, the release of Fable 5, a variant with supposedly stronger protections, was not addressed in the new directive. Anthropic acknowledged the progress on X, stating they are quickly restoring access for these organizations while continuing to work with the government to expand access further. The situation highlights the ongoing tension between AI safety concerns and the operational needs of critical infrastructure providers.

1 min readRead
OpenAI limits GPT-5.6 rollout after government request, says restrictions shouldn’t be the norm
TechCrunch AI
General AIJun 26

OpenAI limits GPT-5.6 rollout after government request, says restrictions shouldn’t be the norm

OpenAI is restricting the release of its new GPT-5.6 model lineup, including the flagship Sol, Terra, and Luna models, at the request of the U.S. government. The company is sharing these models only with a small group of trusted partners whose participation has been disclosed to the administration. This move follows similar pressure on Anthropic, which was ordered to remove access for foreign nationals from its Fable 5 model. Dean Ball, a former White House AI advisor joining OpenAI, warns that this creates a de facto involuntary licensing regime that could hinder the U.S. AI race against China. OpenAI explicitly stated that such government access processes should not become the long-term default, calling the current restriction a short-term step. The GPT-5.6 Sol model boasts improved agentic capabilities in coding, biology, and cybersecurity, with new reasoning modes that significantly increase token usage. OpenAI is working with the administration to develop a new framework for future model releases to balance safety with broader availability.

1 min readRead