The ChatGPT Observer Edition 34

The ChatGPT Observer Edition 34

If you're short on time, here's your one-minute summary of this week's key AI news and advancements:

?? Top Story: Anthropic launches Claude 3.5 Sonnet with revolutionary "computer use" feature, allowing AI to interact with computer interfaces and surpassing GPT-4o on SWE-bench with a 49.0% score. Claude.ai launches new analysis tool for data processing and real-time insights, supporting various business functions.

?? Key Statistics: KPMG reports 88% of US businesses see improved profitability from digital transformation, with 39% having AI use cases delivering business value.

?? ChatGPT & OpenAI News: OpenAI disbands AGI Readiness team, introduces sCM for 50x faster image generation, and expands Advanced Voice features to EU users. OpenAI's upcoming 'Orion' model set for December release, expected to be 100x more powerful than GPT-4.

?? Corporate AI Developments: Meta partners with Reuters for real-time news integration, xAI releases Grok API, and Microsoft unveils autonomous agent capabilities in Copilot Studio.

?? AI Innovations: Midjourney tests new image editing features, Runway introduces Act-One for character animation, and ElevenLabs launches Voice Design for custom AI voices.

?? AI in Business: Zoom partners with Suki for AI medical note-taking, Harvard transforms medical education with AI, and SlideTeam launches free AI presentation maker.

? Responsible AI: Dow Jones sues Perplexity AI over copyright issues, U.S. issues AI National Security Memorandum, and IndiaAI Mission selects 8 projects for ethical AI development.

?? For more details, read the full newsletter!


Anthropic Launches Claude 3.5 Sonnet with Revolutionary "Computer Use" Feature

Source: Anthropic

Anthropic has unveiled the upgraded Claude 3.5 Sonnet, setting new benchmarks for AI performance, especially in coding and tool use. With a 49.0% SWE-bench score, Claude 3.5 Sonnet now surpasses all publicly available models, including GPT-4o and Gemini 1.5 Pro. This makes it a top choice for developers, particularly for tasks involving complex software engineering and multistep reasoning.

One of the most exciting updates is the introduction of the “computer use” capability. This experimental feature allows Claude 3.5 Sonnet to interact with computer interfaces—moving cursors, clicking buttons, and typing—based on user instructions. By analysing screenshots, it can perform tasks like filling out forms and navigating applications. Though still in beta, this feature marks a significant step forward in AI-driven automation, offering promising applications in software testing and process automation.

While the new computer use capability is still being refined, it has the potential to revolutionise how AI integrates with user interfaces. Anthropic has taken steps to ensure safety with new classifiers and pre-deployment testing by the US AI Safety Institute and UK AI Safety Institute.

Claude 3.5 Sonnet is now available via Anthropic’s API, Amazon Bedrock, and Google Cloud’s Vertex AI.

Read more: Anthropic

Watch the demos:


Claude.ai Introduces Analysis Tool for Data Processing and Real-Time Insights

Source: Anthropic

Anthropic has launched a new analysis tool for Claude.ai , allowing users to run JavaScript code directly within the platform. This built-in feature functions like a code sandbox, enabling Claude to clean, explore, and analyze data with mathematical precision. The tool expands Claude’s capabilities, supporting tasks such as analyzing CSV files and generating accurate, reproducible results.

The analysis tool benefits various teams:

  • Marketers can identify conversion opportunities from customer interaction data.
  • Sales teams can gain insights into country-specific performance.
  • Product managers can use customer engagement data for sprint planning.
  • Engineers can optimize server resource utilization.
  • Finance teams can create dashboards for financial trend analysis.

Read more: Anthropic


  • 88% of US businesses report improvements in profitability and performance from digital transformation efforts, a significant increase from 45% in 2023.
  • 81% of respondents find it challenging to keep up with the rapid pace of technological change.
  • 71% of US companies ensure their technology investments target sustainability and social responsibility goals.
  • 39% have AI use cases that deliver business value, and 35% are scaling AI with achieved ROI.
  • 72% of respondents continuously evaluate the business value of their technology investments, slightly above the global average.
  • 84% predict a shift from knowledge creation to knowledge sharing due to AI's impact on knowledge work over the next decade.

Read more: 2024 KPMG US technology survey report: The digital dividend


Advanced Voice Now Available for Plus Users in the EU and Surrounding Countries

OpenAI has expanded Advanced Voice features to Plus users in the European Union, Switzerland, Iceland, Norway, and Liechtenstein. This upgrade allows users to experience enhanced voice capabilities, making interactions more dynamic and engaging.

Read more: OpenAI on X


Accelerating AI: Introducing sCM for Faster, High-Quality Sampling

Source: OpenAI

OpenAI has unveiled sCM (simplified Continuous-time Consistency Models), a new approach that simplifies, stabilises, and scales the training of continuous-time consistency models. This method achieves comparable sample quality to leading diffusion models but with just two sampling steps, resulting in a ~50x speedup. In simpler terms, sCM allows AI models to generate realistic images, audio, and video much faster, reducing the time needed from hundreds of steps to just two, without compromising quality. This breakthrough makes real-time AI generation practical for everyday applications, such as design tools, music, and video effects. By training with 1.5 billion parameters on ImageNet, sCM opens up new possibilities for efficient and high-quality generative AI.

Read more: OpenAI | The Decoder


OpenAI Disbands AGI Readiness Team Amid Leadership Changes

OpenAI has dissolved its "AGI Readiness" team, which assessed the company's ability to manage advanced AI and global preparedness. Miles Brundage, the team's senior advisor, resigned to pursue AI policy research externally, citing the need for more impactful work. This follows the May disbandment of the "Superalignment" team, which aimed to control superintelligent AI.

Read more: CNBC


OpenAI Prepares to Launch New AI Model ‘Orion’ by December

OpenAI plans to release its next major AI model, Orion, by December, initially providing access to select partners instead of a wide ChatGPT rollout. Engineers at Microsoft, a key OpenAI partner, are preparing to host the model on Azure as early as November. While seen as a successor to GPT-4, it’s uncertain if Orion will be branded as GPT-5. Orion is expected to be up to 100 times more powerful than GPT-4, with development completed in September.

Read more: The Verge


Ex-OpenAI CTO Mira Murati Seeks Funding for New AI Startup

Mira Murati, former Chief Technology Officer at OpenAI, is raising capital for a new AI venture aimed at developing products using proprietary models. Sources suggest the startup could secure over $100 million in funding. Murati’s departure from OpenAI in late September, alongside researcher Barret Zoph, marks another high-profile exit as the company undergoes governance changes. Murati was a leading figure behind projects like ChatGPT and DALL-E during her six-year tenure at OpenAI.

Read more: Reuters


Asana Unveils No-Code Tool for Creating AI Agents

Asana has launched AI Studio, a no-code platform for designing AI agents that act as virtual teammates. The tool allows users to create customised AI agents to automate tasks, streamline workflows, and enhance productivity without needing programming skills. This move adds Asana to the growing list of companies providing AI agent builders, making it easier for teams to integrate AI capabilities into their daily operations.

Read more: ZDNet


Perplexity Launches Native Mac App for AI-Powered Search

AI-focused search startup Perplexity has introduced a native Mac app, positioning itself as a competitor to AI tools like ChatGPT and Quora's Poe. The app is free, with a Pro subscription available for $20/month, offering over 300 in-depth searches. The Mac app supports keyboard shortcuts for quick searches, voice mode, file uploads, and voice dictation. Users can also track conversation history and ask follow-up questions. Perplexity is reportedly seeking $500 million in funding, potentially doubling its valuation to $8 billion.

Read more: TechCrunch


Apple Expands AI Capabilities in iOS 18.2 Developer Beta

Apple's latest iOS 18.2 developer beta introduces new AI features, including Genmoji for creating custom emojis, Visual Intelligence for real-time object recognition, and ChatGPT integration for Siri and writing tools. The Genmoji tool allows users to generate unique emojis based on text descriptions or photos. Visual Intelligence enables iPhone 16 users to solve math problems or identify objects using the camera. ChatGPT now assists with tasks like generating itineraries and composing text. These features aim to enhance AI-driven user experiences across Apple’s devices.

Read more: Engadget | Watch WSJ Interview with Apple's Craig Federighi


Meta Partners with Reuters for Real-Time AI News Integration

Meta has signed a multi-year agreement with Reuters to provide real-time news content for its AI chatbot on platforms like Facebook, Instagram, WhatsApp, and Messenger. Starting Friday, U.S. users can access news-related answers sourced from Reuters, with links to original coverage. This marks Meta's first major news deal in the AI era, aiming to enhance the chatbot's ability to deliver reliable information amid a shift away from traditional news in main feeds. Compensation details remain undisclosed, and it's unclear if Reuters' content will be used for AI training.

Read more: Axios


xAI, Elon Musk's AI Startup, Launches API for Grok Model

Elon Musk's AI company, xAI, has released an API for its generative AI model, Grok. The API, starting with the "grok-beta" model, is priced at $5 per million input tokens and $15 per million output tokens. It supports function calling to connect Grok with external tools and hints at future vision capabilities. Grok, known for its "rebellious streak" and unfiltered responses, is integrated into X (formerly Twitter) and powers features like image generation and news summarization.

Read more: TechCrunch


Midjourney Tests New Image Editing and Re-Texturing Features

Midjourney is testing a new external image editor and re-texturing tool, allowing users to upload images for scene modifications, including cropping, repainting, expanding, and altering lighting and materials. The re-texturing mode can adjust surfaces for a refreshed look, while all edits can be controlled via text prompts and region selection. These tools integrate with advanced features such as style references, character references, and personalized models.

A new V2 AI moderator is also being introduced to provide more nuanced content moderation, examining prompts, images, and edits to reduce false positives. Access to these features is limited during the initial rollout to yearly subscribers, members subscribed for at least 12 months, and those who have generated over 10,000 images.

Read more: Midjourney | Midjourney on X

Watch Midjourney's NEW Powerhouse: Deep Dive & Tips for Editor!


Runway Gen-3 Alpha Introduces Act-One for Character Animation

Runway's Gen-3 Alpha model offers improvements in fidelity, consistency, and speed for creating generative videos. The new Act-One feature animates character images using "driving performance" videos to influence expressions and movements. Users upload a video to guide the animation and a reference image for the character, following best practices to ensure high-quality results.

Read more: Runway Help Center


Ideogram Canvas Launches with Magic Fill and Extend Tools

Ideogram has introduced Ideogram Canvas, an infinite creative platform for image organization, generation, editing, and combining. The new tools include Magic Fill for inpainting, allowing users to edit image regions, fix imperfections, and change backgrounds, and Extend for outpainting, expanding images while preserving style. Available with any paid plan, Ideogram Canvas supports advanced text rendering and precise prompt adherence, making it ideal for graphic design. Users can access unlimited canvases, and developers can integrate these features using the Ideogram API.

Read more: Ideogram | Ideogram on X


Stable Diffusion 3.5 Launches with Enhanced Customisability and Diversity

Stability AI has released Stable Diffusion 3.5, featuring three new models: 3.5 Large (8 billion parameters), 3.5 Large Turbo (a faster variant), and 3.5 Medium (launching October 29). These models offer improvements in image diversity, quality, and prompt adherence, with outputs covering various styles from 3D art to line drawings. The models aim for broader inclusivity, generating diverse images without extensive prompting. Stability AI continues its permissive licensing policy, allowing free use for non-commercial purposes and businesses under $1M in revenue.

Read more: Stability AI | TechCrunch


Canva Launches Droptober with New Features and Wellness Challenge

In October 2024, Canva introduced Droptober, featuring new tools to boost creativity and a wellness initiative. Key updates include the AI-powered Dream Lab for image generation, enhanced Magic Write, Custom Mockups, and Interactive Charts for better data visualization. While these features aim to improve user experience, they are mostly limited to Pro subscribers, sparking debates about accessibility.

Read more: Canva | Newsroom


Microsoft Copilot Studio Unveils Autonomous Agent Capabilities

At its AI Tour in London, Microsoft announced new autonomous agent capabilities in Copilot Studio, set for public preview at Microsoft Ignite 2024. These agents automate tasks across business functions, acting independently to initiate events and execute processes. Organizations can build, customise, and manage agents using a low-code interface and secure infrastructure, with prebuilt connectors and the latest AI models, including the OpenAI o1 series.

Early adopters like McKinsey and Pets at Home have already leveraged these agents for workflow automation, showing significant time and cost savings. The launch aims to help companies embrace AI-first approaches and optimize business processes.

Read more: Microsoft


IBM Granite 3.0: Open-Source AI Models Built for Business

IBM has launched Granite 3.0, a family of open-source AI models under the Apache 2.0 license, designed specifically for enterprise applications. These models, available in 8B and 2B parameter versions, focus on tasks like Retrieval Augmented Generation, coding assistance, and risk detection. Granite 3.0 features enhanced data training, greater flexibility, and specialized variants for instruction and security.

The models are accessible through platforms like Hugging Face, GitHub, and IBM's Watsonx.ai , supporting programming across 116 languages and advanced enterprise use cases.

Read more: IBM Granite | ZDNet


Apple Expands AI Capabilities with ChatGPT Integration in iOS 18.2

Apple's iOS 18.2 beta introduces significant AI enhancements, including integration with ChatGPT. The update allows Siri to utilize ChatGPT for complex queries and text generation, enabling tasks like creating original content and generating images. Users can choose to bypass Siri's permission prompts for ChatGPT, except when sharing files. This integration supports devices with Apple Intelligence, such as iPhone 16 and 15 Pro models.

ChatGPT can be accessed without an OpenAI account, although users can link one for personalized features. The public release of iOS 18.2 is anticipated by mid-December.

Read more: CNBC | 9to5Mac


ElevenLabs Introduces Voice Design for Custom AI-Generated Voices

ElevenLabs has launched Voice Design, a new AI tool that allows users to create unique, custom voices from scratch by adjusting parameters like gender, age, and accent, or using text descriptions. Each voice is generated with built-in randomness for a unique output. This tool is ideal for various industries, including audiobooks, gaming, content creation, and corporate communications, providing new possibilities for lifelike voice applications.

Voice Design currently supports English with a focus on American and British accents.

Read more: Voice Design Overview | Voice Design Update | Watch examples from MattVidPro AI


SlideTeam Launches Free Online AI Presentation Maker

SlideTeam has introduced a free AI-powered tool that creates professional presentations in minutes. Trained on a variety of designs and content, the AI generates decks with minimal input, though more details lead to better customization. Users can easily modify templates, colours, and text before downloading in PPTX format for editing in PowerPoint, Google Slides, or Apple Keynote.

Read more: SlideTeam


Genmo Unveils Mochi 1: An Open-Source Video AI to Rival Proprietary Models

AI startup Genmo has launched Mochi 1, a 10-billion-parameter open-source video generation model. The new model offers capabilities on par with or surpassing proprietary solutions like Runway’s Gen-3 Alpha and Kling. Mochi 1 features high-fidelity motion and precise prompt adherence, aiming to democratize video generation. The model is free under Apache 2.0, though it requires powerful hardware (4 Nvidia H100 GPUs). A 480p preview is available now, with a 720p upgrade expected later this year.

Read more: VentureBeat


Speechify Enhances Reading with Text-to-Speech Capabilities

Speechify is a text-to-speech app designed for efficient reading. It supports listening to various documents (PDFs, books, emails) on the go, offers speed reading at up to 4.5x, and provides 200+ human-like voices in 60+ languages. The AI-powered tool also includes instant summaries, a scan-and-listen feature, and voice cloning for personalized experiences. Accessible across iOS, Android, Chrome, Mac, and Edge, Speechify caters to students, professionals, seniors, and those with learning differences.

Read more: Speechify


AI Algorithm Decodes Pig Sounds to Detect Emotions

Researchers have developed an AI tool capable of interpreting pig vocalizations, potentially allowing farmers to monitor animal emotions. The algorithm, created using thousands of recorded pig sounds, distinguishes between positive and negative emotions based on vocal cues. Short grunts indicate positive feelings, while long grunts and high-pitched squeals signal stress or discomfort.

The tool, which aims to improve animal welfare, could eventually be integrated into mobile apps for farmers and even help label farm practices for consumers.

Read more: Sky News


ChatGPT Analyzed 170,000 Lines of Code in Seconds

Using ChatGPT, a user was able to analyze over 170,000 lines of G-code from two 3D printer test files to identify differences in print speed. The AI revealed that faster prints used higher feed rates, larger layer heights, and more aggressive acceleration settings. ChatGPT also confirmed the code was likely generated by a slicer rather than hand-coded, providing insights that saved hours of manual work.

This case illustrates the potential of AI tools for complex data analysis, while reminding users to critically evaluate AI-generated responses.

Read more: ZDNet


Generative AI Is Transforming Medical Education at Harvard

Harvard Medical School (HMS) is integrating generative AI into its curriculum to prepare future doctors for a tech-driven medical landscape. Changes include a new one-month course on AI in healthcare for first-year students, a PhD program focused on AI in Medicine, and grants for AI-based education projects. The aim is to equip students with AI skills essential for clinical practice, including data analysis and decision-making.

AI is also being tested in clinical settings, such as automating documentation and aiding in diagnosis, with a focus on improving patient interactions and medical training.

Read more: Harvard Magazine


Zoom Partners with Suki for AI-Powered Medical Note-Taking

Zoom has teamed up with AI medical scribe startup Suki to offer AI-driven note-taking during telehealth consultations. The collaboration aims to help doctors reduce administrative time by automatically generating medical records from patient visits. This partnership comes as Zoom shifts its focus towards integrating AI tools across workplaces. Suki, chosen by Zoom over other AI scribe providers, recently closed a $70 million Series D funding round.

The move follows a broader trend, with companies like One Medical and Microsoft’s Nuance also investing in AI for healthcare documentation.

Read more: TechCrunch


Funlab Eyes AI for Enhanced Customer Engagement

Entertainment venue operator Funlab, known for brands like Strike, Holey Moley, and Archie Brothers, is set to explore AI-driven customer engagement. Speaking at a Salesforce event, Tracy Tanti, head of customer relationships and retention, highlighted plans to experiment with AI to integrate digital elements into physical experiences. The company aims to use AI to support venue and customer service teams.

Funlab is currently optimizing its data in preparation for AI adoption, following insights from Norths Collective's transformation journey. As it expands in the U.S., the company seeks to boost customer engagement and grow its event sales.

Read more: iTnews


AI-Powered Facial Recognition Revolutionises Sea Turtle Tracking

Green Heroes, an Australian non-profit, is using AI technology, typically employed for passport facial scanning, to monitor sea turtles. Through the Ocean ID project, AI analyses the unique facial scales of turtles, similar to human fingerprints, enabling non-intrusive tracking. Citizen scientists, divers, and snorkellers contribute by uploading photos of turtles to a database.

The method offers a cost-effective alternative to invasive flipper tagging or expensive satellite tracking. Experts believe the technology could extend to other marine species with distinctive physical markings. Green Heroes plans to expand the initiative into a mobile app to boost conservation efforts.

Read more: ABC News


European Commission Pilots GPT@EC for Policy Drafting

The European Commission has launched GPT@EC, an AI tool designed to assist staff with drafting policy documents, summarising content, brainstorming, and generating code. This pilot project, led by DG DIGIT, marks the Commission's first foray into deploying generative AI, akin to tools like ChatGPT, for internal use. Previously, AI was mainly used for translations and document classification within the Commission.

The tool, which builds on GPT@JRC tested by the Joint Research Centre, offers access to models like Meta's Llama and OpenAI’s ChatGPT. It operates securely on the EU's data centre, adhering to compliance standards to ensure data protection.

Read more: POLITICO


AI Robot Boosts Aluminum Can Recycling in Chicago

Chicago's main recycling facility, The Exchange, uses an AI-powered robot to recover aluminum cans that would otherwise end up in landfills. The AI system quickly identifies cans in just 10-15 milliseconds using an optical scanner, enabling a robotic arm to retrieve them. This initiative has helped capture over a million cans since last year.

Aluminum can recycling is highly energy-efficient, requiring only 5% of the energy needed to produce new cans. However, the U.S. recycles just 45% of aluminum cans. Improved legislation, such as deposit programs, could significantly increase recycling rates.

Read more: Axios


-

Dow Jones and New York Post Sue Perplexity AI Over Copyright Issues

News Corp’s Dow Jones & Co. and the New York Post filed a lawsuit against AI startup Perplexity on October 21, 2024, accusing the company of massive copyright infringement. The lawsuit alleges that Perplexity's AI, which offers a platform for information discovery, unlawfully copies copyrighted material from publishers like the Wall Street Journal and New York Post, diverting revenue and readers away from the original sources. The suit seeks damages and aims to stop Perplexity from using the plaintiffs' content without authorization.

Background and Allegations:

  • The lawsuit claims that Perplexity’s AI allows users to “skip the links” to original content, thus repurposing valuable copyrighted works without compensation.
  • In July 2024, Dow Jones and the Post notified Perplexity about these issues and proposed a licensing discussion, but Perplexity did not respond.
  • News Corp CEO Robert Thomson criticized Perplexity for copying “copious amounts” of material and presenting it as a substitute for original journalism. He labeled the startup’s actions as “content kleptocracy.”

Broader Context:

  • The lawsuit comes amid rising concerns about copyright and intellectual property rights related to AI-generated content. News Corp had previously reached a $250 million licensing deal with OpenAI, whereas Perplexity has not secured similar agreements.
  • Other publishers, such as The New York Times, have also taken legal actions against AI companies, including Perplexity, OpenAI, and Microsoft, for copyright infringement.

Read more: Variety

Perplexity’s Response:

  • Perplexity responded by stating that the lawsuit is an attempt to stifle technological innovation and dismissed allegations of wrongdoing. They argued that their technology aligns with fair use principles and facilitates user access to information. Read more: Perplexity's detailed response here .
  • Additionally, Perplexity’s CEO expressed frustration with the lawsuit in an interview with TechCrunch, claiming that the lawsuit reflects an attempt to prevent technological progress. More details can be found here .


OpenAI Supports U.S. AI National Security Efforts with New Framework

OpenAI endorses the U.S. National Security Memorandum on AI, emphasizing democratic AI leadership and responsible use. The memorandum aims to boost AI development through increased access to chips, power, and data centers.

OpenAI’s national security work aligns with its values, focusing on democratic principles, safety, responsibility, and accountability. Current partnerships include collaborations with DARPA on cybersecurity and with USAID for administrative efficiency, with plans to expand projects in scientific research and logistics.

Read more: OpenAI


U.S. Issues AI National Security Memorandum to Lead in AI Development

The White House released a National Security Memorandum on AI, emphasising U.S. leadership in safe and responsible AI use for national security. The memo outlines plans to advance AI capabilities, secure AI infrastructure, and establish safeguards for human rights and safety. Key goals include boosting domestic AI development, protecting AI intellectual property, and collaborating with allies on AI governance. The memo also calls for improved AI risk management, governance frameworks, and talent acquisition to maintain a technological edge over strategic competitors.

Read more: The White House


‘Blade Runner 2049’ Producer Sues Elon Musk’s Tesla and Warner Bros. Over AI Images

Alcon Entertainment has filed a lawsuit against Tesla and Warner Bros. Discovery, accusing them of using AI-generated images from Blade Runner 2049 for Tesla’s robotaxi promotion without authorization. The lawsuit claims that the promotional image showcased during a recent event was a replica of a scene from the film, created by feeding movie stills into an AI generator. Alcon, citing ongoing projects, expressed concerns about associating the movie with Musk due to his controversial views and seeks damages and a court order to prevent further use of the materials.

Read more: Hollywood Reporter


IndiaAI Mission Selects 8 Projects to Advance Ethical AI Development

The Indian government has chosen eight projects under the “Safe and Trusted AI Pillar” of the IndiaAI Mission to foster responsible AI development. Out of 2,000 proposals, the selected initiatives focus on areas like AI bias mitigation, machine unlearning, privacy-enhancing tools, and ethical AI frameworks. The INR 10,372 Cr mission, approved earlier this year, aims to promote transparent AI governance and tackle AI bias, privacy, and accountability challenges. The announcement follows the formation of an advisory group to establish AI regulation frameworks.

Read more: Inc42


AI Leaders Discuss Responsibility, Regulation, and Text’s Future

At the TIME100 Impact Dinner in London, AI leaders discussed the challenges and responsibilities surrounding AI development. Panelists Jade Leung (U.K. AI Safety Institute), Victor Riparbelli (Synthesia), and Abeba Birhane (Trinity College Dublin) highlighted risks posed by AI, including bias, data sourcing, and safety practices. They stressed the need for thorough evaluations of AI’s capabilities and regulation based on empirical evidence. Riparbelli predicted that AI-generated videos could soon make text obsolete, while Birhane warned against overhyping AI's current capabilities.

Read more: TIME


Robot & Technology Exhibitions - GITEX 2024 and IROS 2024 Highlight Innovations in AI and Mobility

The recent GITEX 2024 in Dubai and IROS 2024 in Abu Dhabi showcased cutting-edge advancements in robotics, AI, and transportation. GITEX featured groundbreaking innovations like the Nissan Hyperforce, an all-electric supercar with 1,360 horsepower and a 0-100 km/h acceleration in just 2 seconds, illustrating the future of high-performance electric mobility. Meanwhile, IROS focused on robotic agility with events such as obstacle courses for quadruped robots and humanoid robots demonstrating advanced navigation and athleticism.

The exhibitions emphasised trends like integrating various mobility solutions, enhancing human-robot interaction, and using AI in security applications.


OpenAI Scientist Unveils ‘System Two Thinking’ AI at TED AI Conference

At the TED AI conference in San Francisco, OpenAI scientist Noam Brown discussed the potential of OpenAI's new o1 model to transform AI through "system two thinking" — a slower, more deliberate reasoning approach. Brown compared this method to scaling AI models by 100,000x, revealing that giving AI more time to think can lead to major performance improvements without vastly increasing data or compute requirements.

The recently launched o1 models demonstrate high accuracy in complex tasks like mathematics, strategic decision-making, and scientific research, achieving 83% on the International Mathematics Olympiad exam. Brown suggested that system two thinking could enhance decision-making in fields like healthcare and energy, reshaping the AI landscape by focusing on deep reasoning over sheer processing speed.

Read more: VentureBeat


Pearson and ServiceNow Partner to Enhance Workforce Development with AI

Pearson and ServiceNow have announced a global, multi-year collaboration aimed at transforming workforce development and employee experiences using AI. The partnership will leverage ServiceNow's AI-powered Now Platform and Pearson's Credly for skills accreditation, targeting the growing skills gap highlighted by the World Economic Forum, which estimates that a billion people will need reskilling by 2030.

Key initiatives include:

  • Research and Insights: Joint analysis of emerging technologies’ impact on the workforce, identifying skills needed for future roles.
  • Skills Accreditation: Integration of Pearson's Credly into ServiceNow to streamline accreditation and skills tracking.
  • Enhanced Learning Experiences: Use of ServiceNow's AI capabilities to improve employee productivity, talent retention, and training.

This collaboration underscores a shared commitment to preparing workers for an AI-driven future, with a focus on reskilling and upskilling across industries.

Read more: PR Newswire


The Upside of AI: A Vision of a Transformed World By Dario Amodei, October 2024

Dario Amodei, CEO of Anthropic, argues that while AI poses significant risks, its potential benefits could radically improve life if managed properly. His essay outlines how powerful AI could create a better future, emphasising five key areas of transformation:

1?? Biology and Health:

  • AI could accelerate scientific progress by 10x, potentially compressing 50-100 years of medical advances into a decade.
  • Expected achievements include eradicating infectious diseases, significantly reducing cancer mortality, curing genetic disorders, and extending human lifespans.
  • AI’s role will be to design and execute biological experiments, akin to a virtual biologist, enhancing discoveries like CRISPR or mRNA vaccines.

2??Neuroscience and Mental Health:

  • AI could vastly improve understanding and treatment of mental illnesses like depression, PTSD, and schizophrenia.
  • Technological advances could lead to fine-grained neural measurements and AI-driven behavioral interventions.
  • Everyday cognitive enhancements and better treatment options would improve overall mental well-being.

3??Economic Development and Poverty:

  • AI could help bring technological and health benefits to developing nations, lifting billions out of poverty.
  • Scenarios like a second “Green Revolution” in agriculture and significant GDP growth in poorer regions are envisioned.
  • While automation poses challenges, AI's spread could boost global economic equality if efforts are made to ensure widespread access.

4??Peace and Governance:

  • Democracies could use AI to promote freedom, counter authoritarianism, and enhance transparency in legal systems.
  • AI-driven governance could also improve service delivery and increase trust in public institutions.
  • An “entente strategy” is suggested, where democratic nations collaborate to gain a technological edge over authoritarian states.

5?? Work and Meaning:

  • As AI takes over economic roles, human work may shift towards pursuits offering meaning beyond economic output.
  • Economic structures may need significant reform, potentially involving universal basic income or new reward systems.
  • The transition will require adapting to new forms of societal organization, similar to historical shifts like industrialisation.

Amodei acknowledges that while these outcomes are optimistic, they are achievable through collaboration, innovation, and strategic efforts to minimize AI risks.

Read more: Dario Amodei blog


??? We'd Love to Hear from You!

Did you find this edition insightful? Like, share your thoughts, and ask us anything.

Stay curious, and until next time, enjoy exploring the vast possibilities AI offers.


???? Subscribe on LinkedIn The ChatGPT Observer Newsletter


As we explore the captivating world of AI technologies, it's important to note that the content of this article is a product of human effort, enhanced by AI tools and additional resources. We have made sure to properly attribute and link all referenced articles to their original sources. For the visuals:

  • Images sourced from external sites are credited within the article.
  • Other images were generated using DALL-E, MidJourney and IdeoGram with enhancements made in Canva. These AI-generated and enhanced images are designed to visually represent each section's key themes.


要查看或添加评论,请登录

社区洞察

其他会员也浏览了