The way we see it, there's limited value in including predictions and future announcements in our regular newsletter. We want our readers to focus on immediately actionable content. However, maintaining an on-demand reference for future developments serves as a valuable short-term intelligence roadmap. As part of the Intelligence Almanac, you'll find:
- Reported Major AI Releases and Updates
- Planned Academic and Research Events
For inclusion in the Almanac, please email humans@wordoflore.ai
Early 2025
AI/Tech
[RELEASED] OpenAI's Deep Research in ChatGPT Pro
#Product Release
Deep Research, OpenAI's latest tool, brings sophisticated research capabilities to ChatGPT Pro users. It can analyze hundreds of sources simultaneously, synthesizing information from texts, images, and PDFs into comprehensive reports
Details
What it is: OpenAI, the company behind ChatGPT and GPT-4, develops advanced AI models and tools that have transformed how we interact with technology. Their ChatGPT service has become a widely-used AI assistant for writing, analysis, and problem-solving tasks, with over 100 million weekly active users.
What's new: Deep Research, OpenAI's latest tool, brings sophisticated research capabilities to ChatGPT Pro users. It can analyze hundreds of sources simultaneously, synthesizing information from texts, images, and PDFs into comprehensive reports. The system provides detailed citations and explains its reasoning process, powered by OpenAI's new o3 model which excels at complex analytical tasks.
Why it matters: Picture spending days on a literature review that Deep Research could help you draft in minutes. While human expertise remains essential for verification and nuanced understanding, this tool could dramatically accelerate research tasks. At $200 monthly with 100 queries included, it's an accessible way to enhance your research workflow.
Read OpenAI's announcement or related article on The Guardian or Reuters

[RELEASED] OpenAI's GPT-4.5
#Product Release
OpenAI announces GPT-4.5 (Orion) as their final non-chain-of-thought model, marking a shift toward simplified product offerings and unified intelligence. The update aims to streamline AI interactions and eliminate complex model selection
Details
OpenAI has announced GPT-4.5 (internally known as Orion) as their final non-chain-of-thought model. This release aims to simplify their product offerings and improve user experience by moving away from complex model selection interfaces. The update focuses on creating a more streamlined, unified approach to AI interactions.
Source: Sam Altman on X
Amazon's Enhanced Alexa
#Product Release
Amazon is preparing to unveil a significant evolution of its Alexa assistant at a showcase event in New York City on February 26, 2025. This enhanced version represents Amazon's most ambitious update to the platform since its initial launch, integrating advanced AI capabilities to transform how users interact with their digital environment.
Details
Amazon is preparing to unveil a significant evolution of its Alexa assistant at a showcase event in New York City on February 26, 2025. This enhanced version represents Amazon's most ambitious update to the platform since its initial launch, integrating advanced AI capabilities to transform how users interact with their digital environment.
The reimagined Alexa aims to bring more natural, context-aware conversations to homes. Users can expect the assistant to maintain conversation history, enabling more coherent multi-turn dialogues and personalized interactions. Practical enhancements include streamlined service integration for tasks like food ordering and transportation, suggesting a shift toward more seamless daily assistance.
This development marks a strategic move in the evolving AI assistant landscape as major tech companies race to integrate generative AI into their consumer products. The new version is expected to debut through a subscription model, potentially starting with a free trial period to allow users to experience the enhanced capabilities firsthand.
Apple Intelligence EU Expansion
#Product Release
Apple Intelligence will indeed roll out to iPhone and iPad users in the European Union starting in April 2025. The rollout will include core features such as Writing Tools, Genmoji, a redesigned Siri with enhanced language comprehension, and ChatGPT integration.
International Conference on Learning Representations (ICLR) 2025
#Conference/Summit
The International Conference on Learning Representations (ICLR) 2025 will be held in Singapore from April 24-28, 2025. This marks the thirteenth edition of this premier conference dedicated to deep learning and artificial intelligence research.
Details
The International Conference on Learning Representations (ICLR) 2025 will be held in Singapore from April 24-28, 2025. This marks the thirteenth edition of this premier conference dedicated to deep learning and artificial intelligence research.
Conference Details
Location: Singapore
Key Dates:
- Full Paper Submission Deadline: October 1, 2024
- Financial Assistance Application Deadline: March 2, 2025
- Early Registration Deadline: March 15, 2025
Conference Scope
ICLR focuses on various aspects of deep learning, including:
- Unsupervised, semi-supervised, and supervised representation learning
- Computer vision and natural language processing
- Reinforcement learning
- Theoretical issues in deep learning
- Societal considerations including fairness, safety, and privacy
- Applications in robotics, neuroscience, and biology
ICLR is one of the three primary conferences of high impact in machine learning and artificial intelligence research, alongside NeurIPS and ICML
Cognitive Science
COGNITIVE 2025: The Seventeenth International Conference on Advanced Cognitive Technologies and Applications
#Conference/Summit
COGNITIVE 2025 is scheduled to take place from April 6-10, 2025 in Valencia, Spain. This conference represents the seventeenth edition of the International Conference on Advanced Cognitive Technologies and Applications.
Details
COGNITIVE 2025 is scheduled to take place from April 6-10, 2025 in Valencia, Spain. This conference represents the seventeenth edition of the International Conference on Advanced Cognitive Technologies and Applications.
Conference Topics
Brain and Cognition
- Brain information processing and informatics
- Human reasoning mechanisms
- Cognitive architectures
- Neural structures and neurobiological processes
- Brain-computer interfaces
Emerging Areas
- Brain connectivity
- Neuromorphic computing
- Brain-like computing
- Neuroscience
Autonomous Systems
- Swarm intelligence and emergent behavior
- Self-organized intelligence
- Complex behavior aggregation
- Agent-based systems and modeling
Mid 2025
AI/Tech
International Joint Conference on Artificial Intelligence (IJCAI) 2025
#Conference/Summit
The 34th International Joint Conference on Artificial Intelligence (IJCAI-25) is scheduled for August 16–22, 2025, in Montreal, Canada. A satellite event will follow in Guangzhou, China, from August 28–31, 2025
Details
The International Joint Conference on Artificial Intelligence (IJCAI) 2025 is the 34th edition of this premier global AI conference, scheduled to take place in Montreal, Canada, from August 16 to August 22, 2025.
To accommodate participants facing visa challenges, a satellite event will be hosted in Guangzhou, China, from August 28 to August 31, 2025.
The conference invites submissions across various tracks, including:
- Main Track: Original research on all aspects of artificial intelligence.
- Special Tracks:
- AI for Social Good
- AI and Arts
- Human-Centred AI
- AI Enabling Critical Technologies
- Survey Track: Comprehensive overviews of specific AI topics.
Important dates for prospective authors include:
- Abstract submission deadline: January 16, 2025
- Full paper submission deadline: January 23, 2025
- Summary reject notification: March 7, 2025
- Author response period: April 6–9, 2025
- Paper notification: April 28, 2025
The conference is chaired by Professor Shlomo Zilberstein from the University of Massachusetts Amherst, with Professor James Kwok from the Hong Kong University of Science and Technology serving as the Program Chair.
Baidu's open-source Ernie 4.5
#Product Release
Baidu announces Ernie 4.5 series launching gradually with plans to open-source on June 30, 2025. Ernie Bot becomes free from April 1, featuring enhanced deep search capabilities and multi-modal support
Details
Baidu has announced significant updates to its Ernie AI platform, with the Ernie 4.5 series set to roll out gradually in the coming months. In a move toward AI democratization, the company will make this series open-source on June 30, 2025. The update coincides with Ernie Bot becoming completely free from April 1, 2025, introducing enhanced deep search features with improved reasoning capabilities and multi-modal support. The platform currently serves 430 million users, processing 1.5 billion queries daily.
Baidu's Ernie 5
#Product Release
Baidu plans to release Ernie 5.0 in late 2025, featuring next-generation AI capabilities to compete with emerging market players. The announcement comes as Baidu looks to expand beyond its current 13M user base in China's competitive AI landscape
Details
Baidu is preparing to launch Ernie 5.0 in the second half of 2025, marking a significant evolution in their AI capabilities. This next-generation model aims to compete with emerging players like DeepSeek in China's rapidly evolving AI landscape. The announcement comes amid increasing competition, with market data showing ByteDance's Doubao leading at 78.6 million users, followed by DeepSeek with 33.7 million users, while Ernie Bot currently serves 13 million users.
International Conference on Machine Learning (ICML) 2025
#Conference/Summit
The 42nd International Conference on Machine Learning (ICML 2025) will be held at the Vancouver Convention Center in Vancouver, Canada, from July 13-19, 2025
Details
The 42nd International Conference on Machine Learning (ICML 2025) will be held at the Vancouver Convention Center in Vancouver, Canada, from July 13-19, 2025
ICML is a premier gathering for machine learning professionals, featuring cutting-edge research presentations in artificial intelligence, statistics, and data science. The conference attracts a diverse range of participants, including academic and industrial researchers, entrepreneurs, engineers, graduate students, and postdocs
Schedule Highlights:
- Tutorials: July 14, 2025
- Main Conference Sessions: July 15–17, 2025
- Workshops: July 18–19, 2025
Knowledge Discovery and Data Mining (KDD) 2025
#Conference/Summit
The 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2025) will be held in Toronto, Canada from August 3-7, 2025. This data science conference features both Research and Applied Data Science tracks.
Details
The 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2025) will be held in Toronto, Canada from August 3-7, 2025. This data science conference features both Research and Applied Data Science tracks.
Cognitive Science
CogSci 2025
#Conference/Summit
The Cognitive Science Society's annual conference, CogSci 2025, will take place from July 30 to August 2, 2025, at the Marriott Marquis in San Francisco. This will be a hybrid event, offering both in-person and virtual attendance options.
Details
The Cognitive Science Society's annual conference, CogSci 2025, will take place from July 30 to August 2, 2025, at the Marriott Marquis in San Francisco. This will be a hybrid event, offering both in-person and virtual attendance options.
This conference brings together researchers from various disciplines, including Artificial Intelligence, Linguistics, Anthropology, Psychology, Neuroscience, Philosophy, and Education, to explore the nature of the human mind.
Cognitive Computational Neuroscience (CCN) 2025
#Conference/Summit
Cognitive Computational Neuroscience (CCN) 2025 is an annual scientific conference focusing on neural computations underlying complex behavior. The conference will be held at the University of Amsterdam from August 12-15, 2025.
Details
Cognitive Computational Neuroscience (CCN) 2025 is an annual scientific conference focusing on neural computations underlying complex behavior. The conference will be held at the University of Amsterdam from August 12-15, 2025.
The 8th annual CCN conference focuses on understanding the computational basis of cognition by integrating neuroscience, cognitive science, and artificial intelligence. The conference brings together researchers working on:
- Brain information processing for real-world tasks
- Brain connectivity and dynamics
- Cognitive and neural-level modeling
- Integration of machine learning with neurobiology
- Comparative studies of biological and artificial intelligence
Late 2025
AI/Tech
Conference on Neural Information Processing Systems (NeurIPS) 2025
#Conference/Summit
The Conference on Neural Information Processing Systems (NeurIPS) 2025 will be held from December 9-15, 2025, in San Diego1. NeurIPS is one of the three primary conferences of high impact in machine learning and artificial intelligence research
Details
The Conference on Neural Information Processing Systems (NeurIPS) 2025 will be held from December 9-15, 2025, in San Diego. NeurIPS is one of the three primary conferences of high impact in machine learning and artificial intelligence research.
The conference covers diverse fields including:
- Machine Learning
- Computational Neuroscience
- Cognitive Science
- Psychology
- Computer Vision
- Statistical Linguistics
- Information Theory
Unscheduled 2025
AI/Tech
OpenAI's GPT-5
#Product Release
OpenAI's GPT-5 integrates o-series technology into a unified system offering unlimited chat access for free users, with enhanced capabilities for Plus and Pro subscribers. Features include voice, canvas, search, and deep research tools
Details
OpenAI's GPT-5 announcement outlines an integrated system incorporating their o-series technology, including o3. The release plan includes three access tiers: standard intelligence setting for free users with unlimited chat access (subject to abuse thresholds), enhanced capabilities for Plus subscribers, and advanced features for Pro subscribers. The system is designed to include voice interaction, canvas functionality, search capabilities, and deep research tools.
Source: Sam Altman on X