Observable trends are facts about the direction of AI development, more reliable than speculation about the distant future. Let us cover the main ones, honestly noting the uncertainty: 1) MODELS GROW MORE CAPABLE: — Over time AI models better understand language, reason, solve more complex and diverse tasks. What models can do now noticeably exceeds the recent past. — This is real, observable progress (not hype). — BUT honestly: it is unclear HOW LONG and HOW MUCH this will continue. Whether progress will accelerate, slow down, hit limits is debated even among experts. Do not project past growth infinitely into the future (as with trends in forecasting — growth is not necessarily eternal). 2) MULTIMODALITY: — From working with text only — to understanding and creating images, sound, video, and their combinations (from the multimodality course). AI works ever better with different data types together, closer to how a human perceives the world. — The trend expands applications (image analysis, voice, video). 3) FROM ANSWERS TO ACTIONS (AGENTS): — AI moves from "answers a question" to "performs tasks": uses tools, takes actions, plans steps (agents, from that course). — This is powerful (automation of the complex) but also raises risks (actions have consequences — hence the importance of safety, action confirmation, from the agents/LLMOps courses). — The trend toward greater autonomy is one of the most significant and discussed. 4) ACCESSIBILITY AND SPREAD: — AI is getting cheaper, more accessible: open and local models (from that course), ready tools, AI embedded in ever more products. — From a "big-lab technology" AI is becoming a tool for everyone — companies, specialists, ordinary people. — This democratizes the benefit but also spreads the risks (powerful tools available to a wide circle, including abuse). 5) INTEGRATION INTO LIFE AND WORK: — AI is ever more deeply embedded in workflows, products, everyday life (assistants, automation, help with tasks). — It changes how people work (AI as a helper amplifying the human — the topic of many courses). 6) What to honestly say about the trends (balance): — The trends are REAL and observable — these are facts about the direction. — But their LIMITS and SPEED are UNCERTAIN: it is unknown how far and fast each trend will go, where it will hit a wall, what will change the trajectory. Extrapolating trends into the distant future with confidence is a mistake (as in forecasting: a trend may slow down/change). — Trends do not lead AUTOMATICALLY to a specific future (e.g. "models are more capable" is not "AGI/an omnipotent AI soon" — that is speculation, the next module). — Different experts differently assess where the trends will lead. The practical meaning: understanding the trends helps see the DIRECTION of AI development (more capable, more multimodal, more agentic, more accessible, more integrated) — this is useful for decisions and expectations. But hold the uncertainty of the limits: trends are about "where it is moving now", not "where it will exactly end up". Practical rule: look at the observable TRENDS of AI development as facts about the direction: models grow more capable, more multimodal, move from answers to actions (agents), get cheaper and spread, integrate more deeply into life and work. These trends are real (you saw them in other courses). But honestly hold the UNCERTAINTY of their limits and speed: it is unknown how far and fast they will go, they cannot be extrapolated into the distant future with confidence, and trends do not lead automatically to specific scenarios (e.g. AGI). Understanding the trends helps see the direction and prepare, but do not turn an observable direction into a confident prediction of the distant future. Trends are a reliable basis to think about AI; their boundless extrapolation is already speculation.