Google's new AI model goes natively multimodal; Meta’s chief AI scientist doesn’t think AI super intelligence is coming soon; AMD forecasts $45 billion AI chip market this year
French AI startup Mistral raises $487m with help from a16z; Sam Altman is Time's 2023 CEO of the year; AI assistants are changing the way code gets made; McDonald's uses AI to make better burgers
November 23, 2023 marked the one year anniversary since the launch of ChatGPT - a landmark moment for the AI industry. While the last 12 months have felt like a decade of innovation on fast-forward before our eyes, the underlying large language models powering chatbots like ChatGPT, Claude or Pi still have one important limitation: they lack the reasoning and common sense that comes from a deeper understanding of the world. These models are trained on vast amounts of data, allowing them to recognize patterns and generate human-like text, images, audio or video. However, they are not natively multimodal; instead, separate components for different modalities are trained separately and then stitched together to roughly mimic how humans perceive the world. As a result, these AI models have no inherent concept of objects, causality, or the typical relationships between things in the world so their responses may sometimes defy logic or physical realities.
Enter Gemini, the largest and most capable AI model from Google. Gemini has been pre-trained from the start to be natively multimodal and therefore can make sense of complex written and visual information. Check out this video from Google to see some of the tasks Gemini can achieve in real-life scenarios!
Beyond implementing basic planning and reasoning capabilities, there’s an even more promising avenue for making machines smarter: integrating a "world model" that equips the AI model with knowledge about how the world works. Think of adding a world model to a large language model as making the large language model go to primary school - it can simulate and emulate the world around it based on scientific principles and observational data. By linking this model of the world with the pattern recognition power of a language model, an AI system could gain more contextual, causal, and factual grounding for its linguistic outputs. It could provide more logical, consistent responses while retaining its ability to handle open-ended topics and tasks.
There are still massive challenges in developing such an integrated system, but this presents a promising path toward less brittle, more broadly intelligent AI - watch my video below to learn more about world models!
And now, here are this week’s news!
❤️Computer loves
Our top news picks for the week - your essential reading from the world of AI
Google DeepMind's Demis Hassabis Says Gemini Is a New Breed of AI [Wired]
Inside the A.I. Arms Race That Changed Silicon Valley Forever [New York Times]
OpenAI COO Brad Lightcap talks about ChatGPT launch, DevDay and how Sam Altman thinks [CNBC]
Google CEO Sundar Pichai on Gemini and the coming age of AI [MIT Technology Review]
OpenAI Rival Mistral Nears $2 Billion Valuation With Andreessen Horowitz Backing [Bloomberg]
Meta and Microsoft say they will buy AMD’s new AI chip as an alternative to Nvidia’s [CNBC]
CIOs Look Past the OpenAI Drama [Wall Street Journal]
How AI assistants are already changing the way code gets made [MIT Technology Review]
From Hollywood to enterprise comms: the Synthesia story [Sifted]
How Nations Are Losing a Global Race to Tackle A.I.’s Harms [New York Times]
The OpenAI Board Member Who Clashed With Sam Altman Shares Her Side [Wall Street Journal]
Forget Sam Altman. America's greatest AI visionary is... an English professor in Illinois [Business Insider]
⚙️Computer does
AI in the wild: how artificial intelligence is used across industry, from the internet, social media, and retail to transportation, healthcare, banking, and more
Meta launches a standalone AI-powered image generator [TechCrunch]
AI is being used to catch fare-dodgers on the London Underground [Timeout]
Microsoft Copilot for Windows 11 Gets GPT-4 Turbo and Dall-E 3 [CNET]
Generative AI translation lifts overseas sales of Chinese online literature industry [South China Morning Post]
This founder is developing AI gun detection technology and uplifting his community while doing so [Fortune]
Waymo is full speed ahead as safety incidents and regulators stymie competitor Cruise [CNBC]
AI to speed up finding the children at risk of mental health conditions [BBC]
U.S. Border Patrol is using AI to crack down on fentanyl trafficking [Axios]
AI can now turn a rough sketch of a skyscraper into a detailed rendering in a matter of minutes [Fortune]
BlackRock to roll out first generative AI tools to clients next month [Financial Times]
‘Help me write’ AI is coming soon to Chrome for desktop [9to5Google]
Calm’s ‘It’s a Wonderful Sleep Story’ will star Jimmy Stewart’s AI-generated voice [The Verge]
McDonald's unveils Google deal to use AI to produce better burgers [The Street]
Bank of England to look closer at rise of AI in finance [Reuters]
With colour-changing fabric, Hong Kong AI lab aims to reduce clothing waste [Reuters]
Google’s ‘Gemini’ makes mobile breakthrough for generative AI [Financial Times]
Video Game Soundtracks Up Next for AI Disruption, Experts Say [Bloomberg]
Which Tasks Can You Delegate to A.I.? Aflac's Guardrails Might Hold the Answer [Inc]
AstraZeneca, AI biologics firm Absci tie up on cancer drug [Reuters]
Prestige or plonk? AI develops a nose to sniff out fake vintage wines [The Telegraph]
Consulting Giants See AI Shaving Years Off the Path to Partner [Bloomberg]
Pika Labs begins rollout of its new AI video model — here's why this is big [Tom’s Guide]
EY claims success in using AI to find audit frauds [Financial Times]
Applying to College? Here’s How A.I. Tools Might Hurt, or Help. [New York Times]
Gmail’s AI-powered spam detection is its biggest security upgrade in years [Ars Technica]
This faux AI chatbot will judge your music taste and make you laugh (hopefully) [ZDNet]
🧑🎓Computer learns
Interesting trends and developments from various AI fields, companies and people
Google DeepMind’s new Gemini model looks amazing—but could signal peak AI hype [MIT Technology Review]
Meta and IBM Launch AI Alliance [Wall Street Journal]
Apple joins AI fray with release of model framework [The Verge]
TikTok owner ByteDance joins generative AI frenzy with service for chatbot development, memo says [South China Morning Post]
Elon Musk seeking to raise $1 billion for his xAI firm [CNN]
Meta's top AI scientist reportedly warned Mark Zuckerberg that Facebook and Instagram could go extinct if they didn't catch up with ChatGPT [Business Insider]
Meet the AI image generator that can create pictures up to 16x higher resolution than Stable Diffusion [Tom’s Guide]
Microsoft readies 'groundbreaking' AI-focused Windows release as new leadership takes the helm [Windows Central]
IBM was early to AI, then lost its way. CEO Arvind Krishna explains what’s next [CNBC]
AI is the new UI for enterprise customers according to Clara Shih, the CEO of Salesforce AI [No Priors on YouTube]
Stability AI goes ‘smol’ with StableLM Zephyr 3B [VentureBeat]
AI could mean free doctors and lawyers for everybody in 10 years, OpenAI investor Vinod Khosla believes [Business Insider]
Highlights from Semafor’s Finding Common Ground on AI event [Semafor]
OpenAI Cofounder Reid Hoffman Gives Sam Altman a Vote of Confidence [Wired]
The Rise of AI in Alternative Browsers—and What’s Next [Wired]
Google teases AlphaCode 2 – a code-generating AI revamped with Gemini [The Register]
X begins rolling out Grok, its ‘rebellious’ chatbot, to subscribers [TechCrunch]
Nvidia Sees Huawei as Formidable AI Chipmaking Rival, CEO Says [Bloomberg]
Runway incorporates Getty Images into its AI generated video [Axios]
OpenAI tender offer is on track for January despite leadership fracas, sources say [CNBC]
Cisco unveils AI assistant for enhanced cybersecurity in Security Cloud platform [Silicon Angle]
What Is Holding Back Neuromorphic Computing? [EE Times]
Databricks launches new tools for building high-quality RAG apps [VentureBeat]
Alibaba Cloud boosts open-source community with enhanced AI and more open-sourced LLMs [Tech Wire Asia]
There’s a smarter way to consume renewable energy [Wired Middle East]
STMicro Unifies AI Toolchain Across Product Lines [EE Times]
Fireside Chat with Scott Belsky, Chief Strategy Officer at Adobe, on AI and creativity [Data Driven NYC on YouTube]
Keep reading with a 7-day free trial
Subscribe to Computerspeak by Alexandru Voica to keep reading this post and get 7 days of free access to the full post archives.