By any measure, 2024 was the best yr ever for artificial intelligence – at least when it involves commercializing the technology.
The large language model (LLM) boom sparked by the launch of ChatGPT in late 2022 shows no signs of slowing down, with quite a few latest LLM models introduced not only by OpenAI and tech stalwarts like Microsoft, Meta and Google, but also dozens other startups and individual developers.
Reports of a slowdown in artificial intelligence research have proven to be, if not unfounded, definitely exaggerated for now.
Additionally, latest technologies have begun to emerge beyond the Transformer architecture that underlies most large LLMs, comparable to Liquid AI’s Liquid Foundation Models.
And finally, corporations have begun to completely embrace an “agency” approach to AI – developing specific AI-powered bots, applications, and workflows that may work on specific problems independently or with less human involvement than the typical back-and-forth of LLM chatbots .
Getting this yr’s news right down to the top fourteen, much less the top ten or top 4, was a vexing effort. But I went ahead and tried, even if I cheated a little, by combining several stories into larger themes. In my opinion, here’s what’s going to have the biggest impact on moving out this yr:
1. OpenAI has expanded far beyond ChatGPT
The company arguably most responsible for ushering in the gen AI era hasn’t lost a penny this yr, despite intensifying competition from upstarts and legacy technologies, and even its own investor and partner Microsoft.
Model o1: OpenAI has released its first latest family of large general-purpose models outside of the GPT series, the o1 “reasoning” series, which allocates more time to process complex prompts, resulting in greater accuracy. It is especially effective in scientific tasks, coding and reasoning.
o3 model: It followed the o1 model in September with the year-end announcement of the blockbuster o3 model. While this solution won’t be available publicly or even to third parties until early 2025, it shows that OpenAI is not resting on its laurels.
ChatGPT Search: Initially introduced as a standalone, invitation-only product called SearchGPT and later collapsed into ChatGPT proper, this feature enables real-time searches of web information inside ChatGPT and improved presentation of search results, increasing its usefulness for timely queries and head face to face with Google, Bing and newcomer Perplexity.
Canvas: Introduced in October, Canvas extends ChatGPT beyond a conversational interface to a workstation-like panel that may dynamically update content at the user’s request, comparable to editing a document or coding project. Of course, it was hard to not see this as a response to, or at least comparable to, Anthropic’s Artifacts announced a few months earlier.
Sora: After almost a yr of teasing us with its closely guarded video generator model, OpenAI finally brought Sora to the masses in early December, quickly generating a wide selection of reactions because it sought to face out in the extremely competitive AI video space with a unique and well-thought-out interface and storyboarding feature.
2. Open source AI has taken off
Lama 3 and 3.1: Meta introduced Llama 3 in April, setting a latest performance standard in open-source AI, and then quickly introduced Llama 3.1 in July with 405 billion parameters. Llama 3.1 releases were used to power Meta AI, the company’s assistant integrated with platforms comparable to WhatsApp, Messenger, Instagram and Facebook, with the goal of becoming the most generally used AI assistant.
Lama 3.3: Released in December 2024, Llama 3.3 delivers performance comparable to larger models, but at a fraction of the compute cost, making it more accessible to enterprise applications.
Meanwhile, Chinese models comparable to Alibaba’s Qwen-2.5 family and DeepSeek’s latest V2.5 and R1-Lite Preview models emerged seemingly out of nowhere to top some benchmark rankings, with Nvidia itself going above and beyond to supply graphics cards and architectures software to release its own open source, powerful Nemotron-70B model.
Nous Research, a small San Francisco-based company that goals to supply more personalized and less restrictive AI models in an open-source format, also showed off some cool latest ideas.
And let’s not forget the French company Mistral, which has quickly expanded its own offering of open-source and proprietary AI solutions.
3. Google’s Gemini series has grow to be a serious contender for the best available
In the comeback story of the yr, Google’s Gemini series of AI models that were once derided for strange image generation and criticized for being overly “woke” have come roaring back with latest, more powerful versions that are now topping third-party performance charts and are increasingly attractive to developers and enterprises.
Google introduced Gemini 2.0 Flash, a multimodal AI model that may analyze streaming video and can see and instruct what you are doing on screen, and then introduced Gemini 2.0 Flash Thinking, which competes with OpenAI’s o1 and o3 inference models.
4. Agentic AI has taken over the enterprise
As the yr progressed, “agentic” AI went from being all the buzz of the world to a veritable series of major product announcements and initiatives from leading enterprise software vendors. Take for example:
Agentforce 2.0 by Salesforce: Salesforce a few days ago announced Agentforce 2.0, an advanced AI agent program that enhances inference, integration and customization capabilities across CRM and Sales offerings, in addition to Slack, significantly improving enterprise productivity tools.
Joule SAP: SAP transformed the Joule chatbot into an open-source Large Language Model (LLM)-based AI agent, driving innovation and efficiency in enterprise environments.
Google Astra project: As part of its Gemini 2.0 initiative, Google launched Project Astra, an AI assistant designed to supply real-time contextual responses by leveraging Google’s suite of services to reinforce user productivity and decision-making.
My big prediction for 2025: AI-generated content will reign supreme
Building on these advances, 2025 will see the proliferation of AI-generated content across business and consumer domains, especially as everyone from OpenAI to Meta, Google, Microsoft, Apple, and even Elon Musk’s xAI now have built-in AI image generators to their sacrifice.
This extension will streamline content creation, improve personalization, and increase efficiency across sectors.
Additionally, we anticipate initial large-scale deployments of Large Language Models (LLM) and AI-powered generative robotics in each business and consumer environments, which is able to revolutionize automation and human-robot interactions.
That’s all for the last #AIBeat newsletter for 2024. Thank you for reading, writing, subscribing, sharing, commenting and being here with us. I am unable to wait to share more with you and hear more from you in 2025.
Merry Christmas and New Year from all of us at VentureBeat to you and your family members.