Is 12 months a long time? In the AI industry, it’s like 5 years. Technology giants are leading the race that drives innovation and stimulates the development of the entire sector. The year 2024 brought many innovations, changes, new tools and promising announcements. What will the development of the artificial intelligence industry look like in 2024?
More, faster, better – the perpetual motion machine of the AI industry
There were so many AI-related events in 2024 that I decided to divide this text into 4 quarters. In each of these 3-month periods, something important happened that clearly shows the dynamics of this sector. Such rapid progress is driven by the competition between two giants, OpenAI and Google. Yes, the creators of the legendary search engine have caught up and caught up with OpenAI in the development of their own AI solutions.
What else besides this race? First of all, a multitude of tools for generating graphics and video content. But as Mr. Bogusław Wołoszański said, “let’s not anticipate the facts.”
1. January-March – development of ChatGPT, Sora and increasing competition
It might seem that the first months of 2024 in the artificial intelligence industry passed rather calmly – after all, 2023 was a real rollercoaster. None of these things, because in January OpenAI reveals ChatGPT 4-Turbo, an improvement of its flagship language model. As the next months will show – the 4Turbo model was only a warm-up.
AI industry leaders didn’t stop at January announcements, because in February (February 15 to be precise), Sam Altman reveals Sora, an artificial intelligence model capable of generating realistic-looking video clips. After Sora’s presentation, the Internet went crazy – a revolution was announced, although Sam Altman and his team did not confirm when the Sora generator would be released (spoiler alert – much, much later).
Although Sora caused a lot of confusion, such a distant release date (basically unannounced) made all sorts of competitors decide to act.
Haiper is a tool created by two former Google DeepMind employees, which appeared in March 2024. At the beginning of its existence, it was able to create 2-second animations, both text-to-video and image-to-video. The trend of 2024, i.e. video generators, also slowly began to emerge. It is worth adding that in March, the European Parliament adopted the AI Act, a 144-page regulation on artificial intelligence.
2. April-June – giants flex their muscles at conferences
The second quarter of 2024 will see AI giants flexing their muscles and lots of announcements. April passed under the sign of X’s and entries on social media revealing the upcoming events. But not only that, because at the end of April Apple presented OpenELM, an open-source framework for training and deploying AI models on mobile devices. This tool aims to increase the availability of advanced AI features for application developers.
May turned out to be a month of conferences. OpenAI Spring Update introduced the new GPT-4o model (“o” for “omni”). It is multimodal – it supports sound and image, so it can be said that ChatGPT has eyes and ears. It is worth adding that a new advanced voice mode was presented during the conference, which responded to OpenAI presenters without delay. During the presentation, ChatuGPT’s voice itself closely resembled the sound of actress Scarlett Johanson’s voice, which became famous a few weeks later.
As for the remaining models, OpenAI has expanded support for more languages, improving the quality of translations and text generation. The API has also been updated, allowing for more flexible integration options and better performance when working with large data sets. Sora, announced in February, still hasn’t arrived.
Google and Apple
Google also presented its new products in May during the I/O 2024 conference. Interestingly, the conference took place the day after the one from OpenAI.
The giant showed a number of innovations, but focused on the development of solutions based on artificial intelligence. Updates have been introduced to Gemini models, including Gemini 1.5 Pro and Gemini Nano, which offer advanced multimodality features (a clear trend for 2024). Google also announced Project Astra, a new AI agent developed by DeepMind that features advanced spatial understanding. New generative tools were introduced, such as Imagen 3 for image creation, Veo for video generation, and Music AI Sandbox for experimenting with music. Google Search has been enhanced with generative features, and Android 15 has gained privacy and security improvements by integrating advanced AI. Phew… that’s a lot, and there’s still Apple left!
During its (June) conference, Apple, in addition to new operating systems, also presented its own version of AI… although not its own version at all. Apple Intelligence was announced as an AI system from Apple that will be based on ChatGPT.
3. July-September – holidays in the AI industry? None of these things!
Contrary to appearances, the holiday period did not bring any slowdown in the AI industry. In July, we saw a real surge and boom of AI tools, both for generating video and music. Udio and Suno (AI music generators) have become so popular that they have attracted the attention of giants. Music publishers sued tool makers for unlawfully using their artists’ songs to train their models.
August is the peak of models for generating videos and graphics. The Flux.1 graphics generation model from Black Forest Lab enters the market and is immediately called the “Midjourney Killer”. China is also entering the videoAI game, releasing Kling AI and Hailuo AI (MiniMax) generators.
On August 1, the AI Act also comes into force in EU countries.
Little? In August, OpenAI presents SearchGPT, its own search engine. This month is also marked by increased hype about the first open-source LLM model, Bielik. In September, giants like Google, Microsoft and Amazon are starting to think more and more seriously about powering their AI centers with nuclear energy.
4. October-December – the speeding AI train does not intend to stop
The last quarter of 2024 brought many long-awaited (and unexpected) solutions and novelties. In October, there was a lot of talk about the Nobel Prize in Physics for groundbreaking research on AI (for John J. Hopfield and Geoffrey E. Hinton) and China’s desire to become independent from Nvidia’s H100 chips (which are the heart of modern AI systems). The Chinese announced that, together with Huawei, they have developed AI Ascend, i.e. their own AI chips whose results match Nvidia’s hardware. As for OpenAI, in October it presented new tools, i.e. Canvas – an editor of content generated in ChataGPT.
November saw Elon Musk continue his work on his own chatbot based on artificial intelligence – Grok. As of November 15, all users agreed to train with their own tweets (X). It is worth adding that two weeks later Grok was made available to all users. On November 11, ElevenLabs (AI tools for voice generation) announced an investment of USD 11 million in Poland. The founders of ElevenLabs are two Poles – Mati Staniszewski and Piotr Dąbrowski.
December was mainly marked by the Advent calendar from OpenAI. Over the course of 12 days, OpenAI presented a number of new features, such as the full version of the o1 model (unveiled in September), the launch of Sora (finally, but not in the EU), and a $200 per month ChatGPT Pro subscription. Google also showed its new products in the pre-Christmas period, such as Gemini 2.0 and the VEO 2 video generator, which outclassed Sora, expected all year round, in terms of quality and physics mapping.
What will 2025 bring?
As you can see, there was a lot going on in the AI industry in 2024 – and this is only a superficial glance of the most important events and topics. So what will 2025 bring?
First of all, we can expect significant development of AI agents, which will be more and more advanced and integrated with everyday tasks in both business and private life. With progress in the field of artificial intelligence, AI agents will become tools capable of making independent decisions, optimizing processes and predicting user needs in real time.
These types of agents, built on advanced models, will be able to understand context and adapt in dynamic environments, as well as integrate with various systems and platforms. Companies will begin to implement such solutions in the areas of customer management, product development, and automation processes. At the same time, there will be challenges in ensuring transparency and accountability of the decisions made by these agents, which will necessitate the development of more advanced audit systems and ethical frameworks for artificial intelligence. In addition, of course, a lot of new tools and even more legal frameworks 😉
You can buy Bitcoin and other cryptocurrencies in a simple and safe way on the zondacrypto exchange.
