OpenAI’s GPT-4, released in 2023, is a proprietary multimodal model (text and images) available via API or Azure and handles a few thousand input tokens by default. Anthropic’s Claude 2 offers a 100,000-token context window and is accessible via API, with prompting costs roughly half of GPT-4 per token. Meta’s LLaMA 2, released in 2023,…
Read more
ChatGPT uses OpenAI’s GPT-4 with multi-turn dialogue and voice input/output, refined through RLHF, and offers a ChatGPT Plus plan at $20 per month. Google Bard runs on PaLM/LaMDA, can generate multiple draft responses and pull web results via Google Search, and is currently free to use in the experiment phase. Soul Machines builds digital-human avatars…
Read more
In Russia, 40% of agricultural enterprises have adopted digital solutions, including agrodrones and predictive analytics, as part of a push toward 2030 productivity targets per Minister Oksana Lut. Renault’s AI-driven monitoring and generative AI cut urgent shipments and production halts by 50%, saving €260 million in inventory costs. Johns Hopkins researchers developed an AI algorithm…
Read more
AI video generators use neural networks, including GANs and diffusion models, trained on large datasets to turn text, images, or clips into realistic videos. They storyboard from a text prompt, breaking a script into scenes and automatically assembling visuals and audio with automated edits and transitions. Text-to-video generation is available on platforms like Runway Gen-2…
Read more
The EU’s AI Act is set to take effect in mid-2026, imposing strict, risk-based regulations and market access requirements. A 10-year moratorium on state-level AI regulation was removed from a major US Congressional bill, underscoring regulatory contention. The US and China are locked in a high-stakes AI race, with the US investing about $700 billion…
Read more
Google unveiled AI Mode as an experiment in March 2025 and began rolling it out to all U.S. English-speaking users by May 2025 during Google I/O. AI Mode runs on Google’s Gemini AI family, initially using Gemini 2.0 before being upgraded to Gemini 2.5 at I/O 2025. The feature uses a query fan-out technique that…
Read more
AI-powered report generators combine ML and NLP/NLG to automatically analyze data and produce visualizations plus plain-language narratives, as exemplified by Power BI Copilot and Tableau’s AI features. Gartner predicts that by 2025 data storytelling will be the most common way analytics are consumed, with about 75% of stories auto-generated. Power BI Copilot provides a chat-based…
Read more
In May 2024, Anthropic disclosed a fundamental breakthrough mapping millions of human-interpretable concepts inside its Claude large language model, offering the first detailed inside look at a modern AI. The European Union’s AI Act took effect in 2024, requiring high-risk AI systems to be transparent and explainable, with penalties up to €30 million or 6%…
Read more
PixVerse released V4 in February 2025 and V4.5 on May 15, 2025, adding cinematic camera controls and a multi-element fusion system. The mobile PixVerse app rose to #1 in the US App Store’s Photo & Video category and #4 overall within days of the V4.5 launch. In June 2025, PixVerse launched Pai I, a Chinese…
Read more
The EU’s AI Act is moving forward without delays, with general-purpose AI obligations beginning in August 2024 and high-risk models following in August 2026, with penalties up to 7% of global revenue. More than 45 European CEOs signed open letters urging a two-year pause, warning that overly complex regulation could hinder Europe’s global competitiveness and…
Read more