All in all, 2024 was the most important 12 months yet for artificial intelligence – at the least by way of commercialization of the technology.
The Large Language Model (LLM) boom sparked by the launch of ChatGPT in late 2022 showed no signs of slowing down, with quite a few latest LLMs coming not only from OpenAI and established tech giants like Microsoft, Meta and Google, but additionally from quite a few others Startups were introduced and individual developers.
Reports of a slowdown in AI research, while not unfounded, definitely proved exaggerated for now.
Additionally, latest technologies began to emerge that went beyond the Transformer architecture that underlies most major LLMs, akin to Liquid AI's Liquid Foundation Models.
And finally, corporations began to completely embrace the “agentic” AI approach – developing specific AI-powered bots, applications, and workflows that may work on specific problems on their very own or with less human responsibility than the everyday back-and-forth of LLM Chatbots.
It was an arduous task to narrow down the 12 months's news stories to the highest 14, let alone the highest 10 or top 4. But I attempted, if cheated a bit, by combining several stories into larger themes. Here's what I believe could have the most important impact this 12 months:
1. OpenAI has expanded far beyond ChatGPT
The company arguably most liable for ushering within the Gen AI era hasn't slowed down this 12 months, despite increasing competition from latest entrants and legacy technologies, even from its own investor and partner Microsoft.
o1 model: OpenAI has released its first latest family of huge general-purpose models beyond its GPT series, the o1 “Reasoning” series, which allows more time to process complex prompts, leading to greater accuracy. It is especially effective for science, coding and reasoning tasks.
o3 model: It followed September's o1 model with the groundbreaking announcement of a balanced o3 model at the top of the 12 months. Although this won't be publicly available and even available to 3rd parties until early 2025, it shows that OpenAI just isn’t resting on its laurels.
ChatGPT Search: This feature was originally introduced as a standalone, invite-only product called SearchGPT before being integrated into ChatGPT. It enables higher real-time retrieval of web information inside ChatGPT and a more refined display of search results, improving utility for current queries and a head-to-head race against Google, Bing and newcomer Perplexity.
canvas: Launched in October, Canvas extends the ChatGPT interface beyond that of a conversational interface to a workstation-like area that may dynamically update content on the user's request, akin to when editing a document or coding project. Of course, it was hard to not see it as a response to, or at the least a comparable feature to, Anthropic's Artifacts announced just a few months earlier.
Sora: After teasing us for nearly a 12 months with its closely guarded video generator model, OpenAI finally released Sora to the masses in early December and quickly sparked a wide selection of reactions because it sought to ascertain itself in a highly competitive AI video space with a singular and unique design to distinguish well thought out user interface and storyboarding feature.
2. Open source AI took off
Lama 3 and three.1: Meta introduced Llama 3 in April, setting a brand new standard for performance in open source AI. Llama 3.1 quickly followed in July with 405 billion parameters. Versions of Llama 3.1 were used to power Meta AI, the corporate's assistant integrated into platforms akin to WhatsApp, Messenger, Instagram and Facebook, with the aim of becoming essentially the most widely used AI assistant.
Llama 3.3: Llama 3.3 was released in December 2024 and delivered performance comparable to larger models but at a fraction of the computational cost, making it more accessible to enterprise applications.
Meanwhile, Chinese models like Alibaba's Qwen 2.5 family and DeepSeek's latest V2.5 and R1 Lite preview appeared seemingly out of nowhere and topped among the benchmark charts, and Nvidia itself went above and beyond Providing graphics cards and software architectures, it launched its own open-source, powerful Nemotron 70B model.
Nous Research, a small company in San Francisco that goals to supply more personalized and fewer restrictive open-source AI models, also presented some cool latest ideas.
And let's not forget French company Mistral, which has been rapidly expanding its own open source and proprietary AI offerings.
3. Google's Gemini series became a serious competitor for the very best series available
In the comeback story of the 12 months, Google's Gemini line of AI models, once mocked for his or her strange image generations and criticized for being overly “woke,” got here back with latest, more powerful versions that now top third-party performance benchmark charts and are increasingly attractive for developers and firms.
Google introduced Gemini 2.0 Flash, a multimodal AI model that supports streaming video analytics and might see and control what you do in your screen, and followed with Gemini 2.0 Flash Thinking, which competes with OpenAI's o1 and o3 reasoning models.
4. Agentic AI took over the enterprise
Over the course of the 12 months, “agentic” AI went from being a fad to an actual series of major product announcements and initiatives from leading enterprise software providers. Take for instance:
Agentforce 2.0 from Salesforce: A couple of days ago, Salesforce introduced Agentforce 2.0, a sophisticated AI agent program to enhance reasoning, integration and customization capabilities in its CRM and sales offerings in addition to Slack, significantly improving business productivity tools.
Joules from SAP: SAP has transformed its Joule chatbot into an AI agent based on open source LLMs (Large Language Models), driving innovation and efficiency within the enterprise environment.
Google's Project Astra: As a part of the Gemini 2.0 initiative, Google launched Project Astra, an AI assistant designed to offer real-time, contextual answers using Google's services, with the aim of improving user productivity and decision-making.
My big prediction for 2025: AI-generated content will take over
Building on these advances, 2025 is anticipated to see the proliferation of AI-generated content across business and consumer spaces, especially as everyone from OpenAI to Meta, Google, Microsoft, Apple, and even Elon Musk's xAI now have AI image generators built into their offerings .
This expansion will streamline content creation, improve personalization, and increase efficiency across various sectors.
Additionally, we expect the primary large-scale deployment of huge language models (LLMs) and generative AI-powered robotics in each business and consumer sectors, revolutionizing automation and human-robot interactions.
It's all in the ultimate #AIBeat newsletter for 2024. Thank you for reading, writing, subscribing, sharing, commenting and for being here with us. I sit up for sharing more with you and hearing more from you in 2025.
From all of us at VentureBeat, we wish you and your family members a completely satisfied holiday and a brand new 12 months.
