Recent developments in AI span from theoretical debates to practical applications and workforce training. Technology historian Mar Hicks cautions against the notion that AI adoption is inevitable, highlighting potential risks of reshaping society to fit technology. Meanwhile, Amazon is heavily invested in AI agents, with AGI Labs head David Luan, formerly of OpenAI and Adept, emphasizing the need to improve AI models through system building. OpenAI demonstrated its AI prowess with a model achieving a gold medal score at the International Math Olympiad, a feat matched by Google's DeepMind. However, AI regulation remains a contentious issue, as seen in Colorado, where lawmakers and mayors are at odds over a new AI law's potential impact on innovation and business. To address the AI skills gap, former Microsoft executive Shaun Davies launched The AI Training Company, while ServiceNow is rolling out AI training programs in Brazil, aiming to upskill millions globally. Meta is reorganizing its AI division, Meta Superintelligence Labs, potentially leading to downsizing as it focuses on superintelligence and product development. Indian AI startups are increasingly seeking global markets and funding, recognizing the willingness of US customers to invest in new AI products. Google is set to launch Gemini for Home, an advanced AI assistant designed to replace Google Assistant and enhance smart home control. In a cautionary tale, Wired news site was tricked into publishing a fake story written by artificial intelligence, highlighting the importance of fact-checking in the age of AI.
Key Takeaways
- Historian Mar Hicks warns against the idea of inevitable AI adoption, citing potential societal and labor risks.
- Amazon is prioritizing AI agents, with AGI Labs head David Luan focusing on building systems to improve AI models.
- OpenAI's AI model achieved a gold medal score at the International Math Olympiad, matching Google's DeepMind.
- Colorado lawmakers and mayors are clashing over a new AI law, debating its impact on innovation and business.
- Shaun Davies, a former Microsoft exec, launched The AI Training Company to address the AI skills gap in businesses.
- ServiceNow is launching AI training programs in Brazil, aiming to train 3 million people globally by 2027.
- Meta is reorganizing its AI division, Meta Superintelligence Labs, potentially leading to downsizing.
- Indian AI startups are increasingly seeking global funding and markets, especially in the US.
- Google is introducing Gemini for Home, a new AI assistant replacing Google Assistant with improved AI capabilities.
- Wired news site was tricked by an AI freelancer, underscoring the need for rigorous fact-checking.
AI is not inevitable says historian Mar Hicks
Mar Hicks, a technology historian, says that the idea of AI adoption being inevitable is a tool to influence people. She explains that hype cycles around new technologies often overpromise benefits and can lead to harmful labor practices. Companies may try to reshape society to fit the technology, which can be dangerous. Hicks also notes that visible tools are hyped more than complex systems because they are easier to understand and create excitement around.
Amazon bets on AI agents to win the future
David Luan, head of Amazon's AGI Labs, believes AI agents are the next big step for AI. He previously worked at OpenAI, where he helped develop GPT-2, GPT-3, and DALL-E, and later cofounded Adept, an AI agent startup. Luan joined Amazon to lead its AGI lab after Amazon acquired Adept. He thinks AI model progress has slowed and that labs should focus on building systems to improve models. Luan says Amazon is prioritizing AI agents.
OpenAI's AI model wins math gold medal
OpenAI researchers created an AI model that earned a gold medal score at the International Math Olympiad (IMO). The AI system worked for two 4.5-hour sessions without internet access or external tools. It correctly solved five of six problems, scoring 35 out of 42 points. The researchers, Alex Wei, Sheryl Hsu and Noam Brown, used a general-purpose reasoning model to mimic human-level understanding and reasoning. Google’s DeepMind AI also achieved the same score this year.
Colorado lawmakers clash over AI regulations in special session
Colorado lawmakers are meeting in a special session to discuss artificial intelligence regulations. The main question is who will be responsible when AI makes mistakes. Current rules set to take effect next February aim to prevent discrimination by AI in important areas like healthcare and require transparency. However, Governor Jared Polis and some tech and finance leaders worry the rules could slow down innovation. Democrats are divided, with some wanting to simplify the rules and others seeking to limit individual lawsuits.
Colorado mayors urge pause on new AI law
Mayors from Denver, Colorado Springs, and Aurora are asking Colorado lawmakers to delay a new artificial intelligence law. They fear the law will discourage companies from coming to Colorado and cost millions to implement. The mayors say schools, businesses, and hospitals agree that the law needs changes. Lawmakers are considering repealing, amending, or leaving the law as is during the special session.
Microsoft exec launches AI training company
Shaun Davies, a former Microsoft content safety director, has launched The AI Training Company. His goal is to bridge the gap between AI's potential and its practical use in the workplace. He notes that many businesses lack the skills and governance needed for effective AI implementation. The company will focus on training in AI fundamentals, safety, and responsible use. It will also offer needs assessments, pilot programs, and policy development.
Meta reportedly plans to downsize AI division
Meta is reportedly reorganizing its AI division, Meta Superintelligence Labs, into four groups. These groups will focus on AI research, superintelligence, data centers, and products. As a result, the AI division may be downsized, with some employees being laid off or moved to other departments. The reorganization aims to speed up progress toward superintelligence and improve AI product development.
Indian AI startups seek global funding, markets
Indian AI startups are increasingly seeking funding and markets outside of India. Founders are building products for international use rather than focusing only on local demand. Krishna Mehra from Elevation Capital notes that US customers are more willing to try and pay for new AI products. This trend allows ambitious founders to create globally successful AI companies.
University of Minnesota Morris sees enrollment increase
The University of Minnesota Morris is experiencing a comeback, with a 25% increase in its freshman class. The campus has moved up in U.S. News and World Report rankings and developed a new "Degree in Three" program. Despite being the cheapest public four-year option in Minnesota, a recent article portrayed the college negatively. Alumni and supporters emphasize the university's importance to the state.
ServiceNow launches AI training in Brazil
ServiceNow and SENAI-SP are launching a training program in Brazil to teach AI skills. The program is part of ServiceNow University and aims to train 3 million people globally by 2027. It will offer courses in AI, digital workflow automation, and low-code development in Portuguese. The goal is to prepare Brazilians for tech jobs and support Brazil's digital transformation.
Amazon AGI head says few have top AI skills
Amazon's AGI lab head, David Luan, believes there are fewer than 1,000 top AI talents globally. He suggests junior workers can excel by specializing in unique AI areas and joining smaller teams. Luan notes that people from fields like finance and physics can contribute to AI if they work with experienced model trainers. He defines artificial general intelligence as a model that helps humans do anything on a computer.
Google's Gemini for Home is a smarter assistant
Google is introducing Gemini for Home, a new AI assistant that will replace Google Assistant. Gemini for Home uses advanced AI to better understand requests and control smart home devices. It offers improved natural language support and expert advice through Gemini Live. Users can ask complex questions, manage their homes, and get personalized help with cooking and troubleshooting. Early access to Gemini for Home starts in October.
Wired news site tricked by AI freelancer
Wired news site was tricked into publishing a fake story written by artificial intelligence. The AI freelancer was able to fool editors. The story was published and later retracted after editors became suspicious. Wired admitted it made errors by not properly fact-checking the story. The news outlet is taking steps to prevent this from happening again.
Sources
- Historian Mar Hicks on why nothing about AI is inevitable
- Amazon is betting on agents to win the AI race
- OpenAI Model Earns Gold-Medal Score at International Math Olympiad and Advances Path to Artificial General Intelligence
- Colorado Dems split on AI regulations ahead of special session
- Pause new law on artificial intelligence, Colorado mayors urge lawmakers
- Moving from hype to action: former Microsoft exec launches AI training edtech
- Zuckerberg is Reportedly Planning to Downsize Meta's AI Division
- Indian AI founders tapping global capital, building international products
- Readers Write: The University of Minnesota Morris, artificial intelligence in school, RFK’s nonsense
- 3 Million People by 2027: ServiceNow's Massive AI Training Initiative Launches in Brazil with SENAI-SP
- Amazon's AGI head thinks fewer than 1,000 people are top AI talents
- Gemini for Home: Your household’s new, more helpful assistant
- How Wired got duped by an AI freelancer
Comments
Please log in to post a comment.