✨ Sora, Gemini 2.0, and the Quantum Chip Willow: AI Closes 2024 Transforming Technology 🚀
🌟 IBM accelerates data center sustainability, Meta launches Llama 3.3 with 100 million users, Apple integrates ChatGPT for free with Siri, and Brussels chooses Barcelona for its first AI factory.
Few weeks have been as productive as this one in the world of Artificial Intelligence. As 2024 comes to a close, Google and OpenAI have made a series of impactful announcements, but IBM, Meta, xAI, and Apple haven’t been idle either. We end the year with an unprecedented sense of democratization in generative AI, with Sora now available to all premium ChatGPT users, except for Europeans and citizens of some authoritarian regimes. Users of OpenAI’s long-awaited video generation tool—after a wait during which several competitive alternatives emerged—have found that Sora doesn’t quite live up to the demos from a few months ago, struggling with movement quality and hyper-realistic visuals. It remains to be seen whether this is due to OpenAI's deliberate censorship to comply with regulations or actual limitations of the tool. Sora features visible watermarks, embedded metadata for traceability, and strict prohibitions against generating copyrighted content, depictions of minors, and explicit material.
Sora's arrival to the AI mainstream, after weeks and months of its competitors gaining the upper hand, highlights an aspect that seemed to have faded with generative AI in other formats: the importance of prompting. While it once seemed like skill in giving instructions or iterating was becoming less critical, generating video with tools like Sora, Runway’s Gen-3 Alpha, or Tencent’s Hunyuan Video shows how vital textual commands are for success. The time will come when creating high-quality, spectacular video becomes effortless and requires little skill. Video generation lags behind text and image generation. With Sora, we’re likely to see a boom in AI-generated clips, despite restrictions to prevent deepfakes. Sora arrives ten months late, with OpenAI holding back its potential and excluding the EU and certain other nations, but Sam Altman's company remains a trendsetter. Just look at their advent calendar campaign, with nightly OpenAI live announcements keeping everyone engaged.
This week, Google made a significant move. They unveiled Willow, a quantum chip capable of solving a task in 5 minutes that would take supercomputers quadrillions of years. They also presented Trillium, a powerful chip for their Gemini 2.0 AI, adding multimodal capabilities to compete with its rivals. Google is making strides in the era of AI agents. Gemini 2.0 aims to play hardball. Additionally, Google has transitioned the Jarvis project to Mariner, making Jules a lifesaver for developers. This week’s developments at Sundar Pichai’s company are monumental. Pichai has hinted at significant changes to Google’s search engine starting in early 2025—brace for impact.
If you, dear reader, are one of the fortunate subscribers to Transparent Algorithm and previously stayed distant from AI developments, know that now is the perfect time to catch up. This week marks a turning point: Sora’s launch, Google’s announcements, OpenAI's introduction of a $200-per-month ChatGPT Pro with unparalleled reasoning, productivity, and robustness, along with new moves from other tech players and the implications of Donald Trump’s second term as President, including his influence on figures like Javier Milei and, of course, Elon Musk.
Musk must be a Transparent Algorithm reader because, just last week, we mentioned how Grok on X (formerly Twitter) is an exciting and stimulating integration. Now, he has announced that it will be free. Grok’s image-generation tool, producing hyper-realistic, uncensored, punk-inspired results, is already making waves on X, shaping a freer vision of AI’s possibilities. Dear reader, take note of these developments and don’t hesitate to try out new applications and tools. You’ve arrived at the right moment and haven’t missed anything crucial from the past two years: things are getting serious now, both for users and in the business world. Any edge others may have over you is cultural or habitual, but embracing the latest tools gives generative AI, multimodality, and countless applications a whole new dimension, enabling you to hop on the accelerating and popularizing AI train.
IBM, Meta, and Apple haven’t been idle either. It seems impossible for a week to go by without the tech giants making moves in AI. Following X accounts, YouTube live streams, and specialized news feeds has become a mix of thrill, passion, and stress. IBM is accelerating the race to make data centers more sustainable with an innovation that reduces energy consumption for AI training fivefold. Meta has released the Llama 3.3 70B text model, adding nearly 100 million new Meta AI users in the past month. Apple has enabled free ChatGPT access via Siri on iPhones, allowing OpenAI’s AI to answer complex questions. With iOS 18.2 now available, ChatGPT and Apple Intelligence’s most advanced features have arrived on iPhones.
Before wrapping up, let me highlight one of the week’s most strategically significant announcements for Barcelona, Catalonia, and Spain: Brussels has chosen the Barcelona Supercomputing Center as the site for a new AI factory, with Spain investing €62 million. In the coming weeks, expect profiles and insights into Trump’s inner circle, particularly David Sacks as the ‘czar’ of AI and cryptocurrencies. I sense a growing presence of AI coverage in mainstream media, presented more kindly and generously. Call me optimistic, but it feels like the catastrophizing has started to recede.