🗞️ AI News Recap

Hey TEDchats Fam!

Happy Monday! Ever feel like you're living in a sci-fi movie? With the pace of AI innovation these days, it's hard not to! Welcome back to our weekly AI recap, where we dive into the latest developments shaping our futuristic world. Let's buckle up for another thrilling ride through the past week of AI!

We have partnered with MIT's business school to bring you today's newsletter. Apply for their online AI short course → here

🐇 QUICK BITES — THE PAST WEEK

Let's Start with the MOST Groundbreaking News: OpenAI's GPT-4o

OpenAI has unveiled GPT-4o, a revolutionary AI model integrated into ChatGPT across all plans. GPT-4o, short for "omni", excels in text, speech, and video understanding, promising GPT-4-level intelligence with enhanced versatility. Unlike its predecessors, GPT-4o offers real-time emotional responses and seamless integration of voice, text, and vision, showcased in a live demonstration by CTO Mira Murati. This breakthrough makes interactions with AI more natural and assistant-like, enabling dynamic conversations and instant, contextually rich responses. With multilingual support and faster processing speeds, GPT-4o is accessible and practical for users worldwide (see OpenAI’s announcement + more GPT-4o demos).

Strava Takes on Cheaters with an AI Boost

Strava is an exercise tracking app that features a live leaderboard ranking Strava athletes. Now, the company has upped its game against leaderboard cheaters by implementing advanced AI to automatically detect and remove dubious activities. This cutting-edge technology aims to preserve the integrity of Strava's competitive segments and ensure fair play for all users (see Strava’s announcement).

The U.S. Senate Eyes a $32B Yearly AI Investment!

A Senate Working Group under the majority leader Chuck Schumer has proposed a massive $32 billion annual investment in AI programs. This sweeping recommendation covers everything from advanced infrastructure and AI grand challenges to bolstering national security and modernizing federal IT systems (read the U.S. Senate’s official report).

OpenAI's Chief Scientist Departs + OpenAI Researcher Emphasizes that Safety has "Taken a Backseat to Shiny Products"

Ilya Sutskever, co-founder, and former chief scientist of OpenAI, has officially departed the company (see Sutskever’s post). This departure comes in the wake of the 2023 OpenAI turmoil surrounding the removal of CEO Sam Altman. Despite this tumultuous history, Altman publicly lauded Sutskever as one of the greatest minds of our generation, highlighting his brilliance, vision, and compassion. Jakub Pachocki, the director of research, will step into the role of chief scientist (see OpenAI’s announcement). Following Sutskever's exit, Jan Leike, a prominent OpenAI researcher, also resigned, and has since expressed concerns in a series of X posts: "Building smarter-than-human machines is an inherently dangerous endeavor. OpenAI is shouldering an enormous responsibility on behalf of all humanity" & "But over the past years, safety culture and processes have taken a backseat to shiny products" (see Leike’s posts here and here).

It's Official — Reddit and OpenAI Have Partnered!

OpenAI will begin using Reddit's data to train its AI models, and will also integrate Reddit content and products into ChatGPT. Additionally, Reddit plans to leverage OpenAI's platform to introduce new AI-powered features for both redditors and moderators. Finally, OpenAI will become a Reddit advertising partner, further strengthening their collaboration (see Reddit’s announcement).

ChatGPT Competitor Anthropic Expands to European Market

Anthropic, the AI startup behind Claude and a fierce competitor of ChatGPT, has hit the EU market, supporting multiple languages including Italian, French, Spanish, and German. This move follows the extension of its API to Europe, marking a significant push for growth as the company raises more capital and hires for a new office in Dublin (see Anthropic’s announcement).

Google Introduces LearnLM — A Family of New AI Models for Education

Google has introduced LearnLM, a new family of models fine-tuned for learning, based on Gemini AI. Collaborating across Google DeepMind, Google Research, and product teams, LearnLM aims to enhance learning experiences by incorporating principles from learning science. This includes inspiring active learning, managing cognitive load, adapting to the learner, stimulating curiosity, and deepening metacognition. The initiative emphasizes responsible collaboration with the AI and EdTech communities to maximize its positive impact (see Google’s announcement).

Microsoft ‘Places’: AI-Driven Office Coordination

Microsoft introduces "Places", a new AI-powered app for Microsoft 365 aimed at streamlining hybrid work schedules. This innovative tool helps employees coordinate office days, ensuring valuable in-person interactions and efficient use of office spaces, among other features. From suggesting optimal office days to booking meeting rooms, Places integrates seamlessly with Outlook and Teams, offering a “new connected workplace app that reimagines flexible work” (see Microsoft’s announcement).

TikTok's Experimental AI Smart Search

TikTok is testing a new ‘AI Smart Search’ feature, displaying ChatGPT-generated results for select queries. This innovative feature provides users with AI-generated snippets at the top of search results, covering topics from recipes to tech recommendations. Additionally, a ‘search highlights’ section offers concise information, though its sources remain unclear. By integrating these AI capabilities, TikTok aims to transform its platform into a more comprehensive search engine, catering to the information-seeking habits of its predominantly younger user base (read more).

Instagram Co-Founder Joins Anthropic!

Mike Krieger, Instagram's co-founder, steps into a pivotal role as Anthropic's first Chief Product Officer. Tasked with leading product engineering and design, Krieger aims to scale transformative AI products like Claude, enhancing both consumer and enterprise experiences. With a strong background in developing intuitive products, Krieger's addition to the team signals Anthropic's aggressive move to capture the consumer market, competing with major players like OpenAI and Google in the evolving AI landscape (see Anthropic’s announcement).

Sony Music Takes Firm Legal Stance on AI Training

Sony Music Group, featuring artists including AC/DC, SIA, and Beyoncé, has issued warnings to over 700 tech companies and streaming services against using its music to train AI without permission. Highlighting the potential of AI while stressing the importance of respecting artists' rights, Sony demands full disclosure of any unauthorized use of its content, threatening legal action if necessary. The letter insists that innovation must not come at the expense of songwriters and recording artists, and seeks details on which songs were used, how they were accessed, and whether copies still exist. This move underscores the growing tension between the music industry and AI developers over copyright infringement issues (see Sony Music’s announcement).

Leveling the AI Playing Field with Cheap Computing Power

Hugging Face is injecting $10 million worth of shared Graphics Processing Units (GPUs) into the AI community, offering small developers, academics, and startups the chance to create cutting-edge AI technologies without the heavy cost burden of dedicated GPUs. The initiative, called ZeroGPU, aims to democratize access to computational power and challenge the dominance of tech giants (read more).

Sona Secures $27.5M for Next-Generation Workforce Management

Workforce management platform Sona has raised $27.5 million in Series A funding to expand into the U.S. market. Leveraging AI to optimize frontline labor, Sona aims to streamline operations for large enterprises in social care and hospitality, setting its sights on new international ventures (see announcement from one of Sona’s major investors).

Voicenotes: A Game-Changer in AI-Powered Voice Note Apps

Voicenotes, created by the founders of Buy Me a Coffee, Jijo and Aleesha Sunny, offers a unique twist in the crowded AI voice-note app market. Not only can it transcribe and summarize your voice notes, but it also features an AI assistant that can answer questions based on your past recordings (try Voicenotes for free here).

And Lastly, Google Unleashes a Barrage of AI Innovations at Google I/O 2024

Google's AI technology took center stage at this year's Google I/O developer conference. You can watch the keynote here for all the latest AI innovations from Google. Highlights included:

• Gemini Pro AI: Google’s most advanced model with a 2 million token context window.

• AI Integration in Google Workspace: Enhancements for Gmail, Docs, and other apps.

• Project Astra: A vision for a multi-sensory AI future that can identify objects in real-time video recordings.

• Pixel 8a: Featuring the Tensor G3 chip, priced at a relatively affordable $499 for an AI phone.

• "Ask Photos" in Google Photos: An AI-powered feature using speech to search through your photo libraries, debuting later this year.

• Scam Call Detection: Gemini AI will be used to notify us of potential scam calls via a pop-up message.

🧰 AI TOOLBOX — RECAP

Understand AI’s Risk and Opportunities

With the rapid growth of AI, it's crucial for businesses to understand both the risks and opportunities it presents. Discover how you can utilize AI to drive success while minimizing potential pitfalls.

In the MIT Artificial Intelligence: Implications for Business Strategy online short course you’ll gain:

• Practical knowledge and a foundational understanding of AI's current state

• The ability to identify and leverage AI opportunities for organizational growth

• A focus on the managerial rather than technical aspects of AI to prepare you for strategic decision making

LEARN MORE

👋 UNTIL NEXT TIME

Thanks for joining us as we explored cutting-edge developments in AI. From OpenAI's GPT-4o to Google's barrage of innovations at Google I/O 2024, it's been an exhilarating week. Stay tuned for our next edition as we continue to unravel the latest in AI breakthroughs!

In partnership with