Reddit filed a lawsuit on 4 June 2025 against Anthropic, alleging the AI startup unlawfully used the platform's data to train its Claude chatbot without a proper licensing agreement. This marks the first time a Big Tech company has legally challenged an AI model provider over its training
Chinese startup DeepSeek announced DeepSeek-R1-0528 on 28 May 2025, delivering significant performance improvements in complex reasoning tasks and achieving near-parity capabilities with paid models OpenAI o3 and Google Gemini 2.5 Pro. The update increased accuracy on the AIME 2025 test from 70% to 87.5%, whilst improving coding performance
Persona-based approaches contextualise AI responses within specific roles, expertise domains, or cognitive frameworks, thereby improving both relevance and quality of generated outputs (Kong et al. 2024). These techniques encompass mega prompts providing extensive contextual information, expert prompts assigning specific professional roles, and advanced reasoning frameworks such as Tree of Thoughts
Tokyo-based Sakana AI, co-founded by former top Google AI scientists, unveiled the Continuous Thought Machine (CTM) on May 12, 2025. The CTM is the first artificial neural network that uses neuron synchronisation as its core reasoning mechanism, enabling AI to "think" through problems step-by-step. This represents a significant
OpenAI announced on May 8, 2025, its first "connector" for ChatGPT deep research, bringing GitHub integration to the popular AI assistant. The new feature lets developers connect their code repositories directly to ChatGPT, enabling the AI to sift through codebases and engineering documentation to generate comprehensive, cited research
On 23rd May 2025, Perplexity officially launched Perplexity Labs, an early-access environment specifically developed for business users that enables them to create customised AI solutions with their own organisational data. The system is built on a proprietary 100 billion parameter language model called Perplexity Labs LLM, which offers functionality significantly
On May 16, 2025, OpenAI announced its latest development, Codex, a cloud-based software engineering assistant capable of working on multiple tasks simultaneously. Codex is powered by codex-1, a version of OpenAI's o3 model specifically optimized for software development. This development marks an important milestone in the field of
Meta officially launched the Llama Startup Program on 21st May 2025, providing 50 selected ventures with access to the complete Llama 3.1 model family, including the 405 billion parameter Llama 3.1 405B model, alongside technical support and up to $150,000 in Meta AI cloud credits. The programme
On April 28, 2025, the European Commission announced that the EU AI Office is launching a €9.08 million tender for monitoring compliance and assessing systemic risks of General-Purpose AI (GPAI) models. The call for tenders relates to the EU Artificial Intelligence Act, which entered into force on August 1,
Opera announced Opera Neon on May 28th, 2025, a new AI browser designed to perform tasks on users' behalf, including shopping, filling out forms, and coding. The Oslo-based Norwegian company describes it as the first "AI agentic browser" that transforms user intent into action, representing years of
Atomic Canyon, a startup founded by Trey Lauderdale eighteen months ago, closed a $7 million seed funding round in May 2025 led by Energy Impact Partners to provide AI-powered document search solutions for the nuclear industry. The company aims to accelerate the traditionally slow nuclear sector whilst tech companies urgently
Example-driven prompts leverage demonstrations to guide large language models (LLMs) in producing precise and contextually relevant outputs, forming a critical component of prompt engineering. This category includes zero-shot prompting, one-shot prompting, and few-shot prompting, each offering varying degrees of guidance for research tasks. These techniques enable researchers to tailor LLM
