- AI in 5
- Posts
- Artificial Superintelligence (ASI) Coming Within 10 Years?
Artificial Superintelligence (ASI) Coming Within 10 Years?
Albert Chan
June 24, 2024
Anthropic’s Claude 3.5 Sonnet Outperforms GPT-4o
Anthropic has unveiled Claude 3.5 Sonnet, a new AI model surpassing OpenAI’s GPT-4o and Google’s Gemini in performance. This model is twice as fast as its predecessor, excels in complex tasks, and introduces the new "Artifacts" feature for in-app content editing. Claude 3.5 Sonnet has topped industry benchmarks, showing significant advancements in both coding and visual capabilities. While lacking some features of GPT-4o, it represents a leap forward in AI development.
Why it matters: Claude 3.5 Sonnet's superior performance and new features highlight rapid innovation and competition in the AI industry.
Ilya Sutskever Launches SSI to Build Safe Superintelligence
Ilya Sutskever, former chief scientist and co-founder of OpenAI, has announced the formation of Safe Superintelligence Inc. (SSI), a new company focused exclusively on developing safe AI superintelligence. Alongside ex-OpenAI engineer Daniel Levy and former Apple AI lead Daniel Gross, Sutskever aims to prioritize safety over commercial pressures. The company will concentrate solely on creating a superintelligent AI system, insulated from typical business cycles and management distractions.
Why it matters: SSI’s singular focus on safe AI superintelligence highlights the urgent need for ethical AI development amidst rapid technological advancements.
The Rising Costs of Training AI Models
The cost of training AI models is doubling every nine months, potentially surpassing $1 billion soon, factoring in electricity, hardware, and employee expenses, says Epoch AI. This escalating trend raises questions about sustainability, with some analysts predicting data scarcity and cost plateau, while others foresee skyrocketing AI chip prices. Microsoft and OpenAI's supercomputer project, costing around $100 billion, exemplifies the financial scale of future AI endeavors.
Why it matters: The rapidly increasing costs may limit AI advancements to only the largest companies, stifling innovation among smaller startups.
Target Tests AI Tool to Boost Employee Productivity
Target is piloting an AI-powered assistant to enhance employee efficiency and productivity, integrating real-time data, task reminders, and personalized recommendations. This tool, currently in limited testing, aims to reduce administrative burdens and improve customer service by leveraging natural language processing for better task management. While it represents a forward-thinking approach in retail, concerns about privacy, data usage, and potential job displacement have yet to be addressed.
Why it matters: Target's AI initiative underscores a significant retail trend towards AI and automation, aiming to optimize store operations and support frontline employees.
UMG Partners with SoundLabs for AI Voice-Cloning Tech
Universal Music Group (UMG) has partnered with SoundLabs to offer its artists AI-powered voice models through the upcoming MicDrop feature. This tool allows artists to create and control their own voice clones, which won't be publicly accessible. MicDrop also includes features like voice-to-instrument conversion and language transposition, facilitating global releases. UMG and SoundLabs stress ethical AI use, keeping artists central to the technology. This partnership aligns with UMG's strategy to enhance creativity with AI while respecting copyrights.
Why it matters: UMG's adoption of AI voice technology aims to empower artists and navigate the ethical challenges of AI in music.
Call Centers Deploy ‘Emotion Canceling’ AI to Protect Workers
Call centers are adopting AI systems to help mitigate stress for employees dealing with irate customers. SoftBank's SoftVoice technology detects hostile tones and alters them to sound calm, providing a "mental shield" for workers. First Horizon Bank experimented with AI to detect employee stress and display calming videos but decided not to implement it widely. These AI tools aim to reduce burnout, although their development suggests a future where AI might fully automate customer interactions.
Why it matters: AI's role in call centers highlights a dual potential: immediate stress relief for workers and a gradual shift towards fully automated customer service.
Snapchat Adds AI-Powered Tools for AR Effects
Snapchat has introduced new generative AI tools to help users and developers create augmented reality (AR) effects more efficiently. These tools include AI assistants for developers, text-based 3D image creation, and features for making realistic AR lenses. Over 300 million users engage with Snapchat’s AR features daily, reflecting Snap’s focus on enhancing AR and AI capabilities. The tools will aid advertisers and creators in driving engagement and monetization through innovative AR experiences.
Why it matters: Snapchat's investment in AI and AR tools underscores the increasing role of these technologies in social media and ecommerce, driving more interactive and engaging user experiences.
Apple and Meta in Talks for AI Integration
Apple and Meta are discussing a potential partnership to integrate Meta's generative AI into Apple's new AI systems for iPhones. This move is part of Apple's broader strategy to partner with multiple AI companies, including Anthropic and Perplexity, to enhance its "Apple Intelligence" system. The partnerships could involve selling premium AI subscriptions through Apple Intelligence, though specifics are still under wraps. Apple has already partnered with OpenAI to integrate ChatGPT into Siri, indicating a willingness to collaborate externally.
Why it matters: Apple's strategy of integrating diverse AI technologies from various partners aims to deliver more robust AI features for users while reducing reliance on a single provider.
AI Boom Strains Power Grid
AI's exponential growth is driving a surge in electricity demand, raising concerns about grid stability, increased fossil fuel use, and conflicts with tech companies' sustainability goals. Data centers could account for 8% of total U.S. electricity use by 2030, nearly triple their current share, according to Goldman Sachs. AI-powered searches consume significantly more energy than standard ones, pushing tech giants to invest in experimental clean energy projects. The AI boom is prompting some utilities to delay coal plant retirements and expand natural gas usage.
Why it matters: AI's future hinges on its energy supply, necessitating massive infrastructure and sustainable energy advancements to meet its industrial-scale demands.
SoftBank Bets on Super-AI
SoftBank CEO Masayoshi Son predicts that artificial superintelligence (ASI), which he believes will be 10,000 times smarter than humans, will emerge within the next decade. He expects artificial general intelligence (AGI), slightly smarter than humans, to arrive in 3-5 years. Son envisions multiple ASI models interacting like neurons, leading to unprecedented cognitive capabilities.
Why it matters: If Son's timeline is accurate, the rapid advancement of AI could fundamentally transform society within the next decade.
