New AI Models: Why Experts Are Raising the Alarm

AI expert details how new AI models have already changed the work of professionals in the tech industry and will soon affect all professions.

Matt Shumer, CEO of HyperWriteAI

February 2020 seemed like an ordinary month. If you followed the news about the virus, you may have heard alarming signals from abroad. Most people did not attach much importance to this: the stock market felt great, children went to school, restaurants were open, people met, shook hands, and planned trips. If someone talked about toilet paper supplies, it looked strange and unfounded. And three weeks later, the world changed forever.

Today, many experts see a similar situation with artificial intelligence. What some consider an exaggeration is actually just a harbinger of grand changes.

New models released in early February 2026 — GPT-5.3 Codex from OpenAI and Opus 4.6 from Anthropic — have changed the rules of the game. Previously, AI performed individual tasks and required constant human control. Now it is capable of creating applications, writing hundreds of thousands of lines of code, testing them, and refining them independently, making decisions almost like a human.

In the past, one had to interact with AI, guide it, and correct the results. Now it is enough to give a task and leave for a few hours, and upon returning, see the completed work.

Moreover, models such as GPT-5.3 Codex demonstrate not just the execution of instructions. They exhibit intellectual judgment, which was previously considered unattainable for machines. This "ability to make decisions" is a breakthrough that changes the perception of AI capabilities.

How this changes work

AI labs have intentionally made machines strong in programming because AI capable of creating code can help create the next version of itself. This opens the way for rapid improvement of systems and automation of almost all cognitive tasks.

The experience of specialists who have observed these changes shows that AI is already handling tasks that were previously considered exclusively human. Lawyers use it to analyze contracts and prepare procedural documents. Financial analysts use it to build models, analyze data, and create investment memorandums. Programmers use it to write multi-day projects with minimal human involvement. Medical professionals test models for reading scans and analyzing laboratory data. Even customer service is already partially automated with the help of AI.

The pace of change

AI progress is going at an exponential speed. In 2022, it made mistakes in basic arithmetic. In 2023, it passed the bar exam. In 2024, it wrote working software and explained science at the master's level. By the end of 2025, AI was performing most of the engineers' work. And since February 5, 2026, new models are capable not only of creating applications but also of participating in the creation of themselves, improving the next versions of AI without human involvement.

The METR organization, which measures the ability of models to perform tasks from start to finish without human assistance, shows a doubling of efficiency every seven months, and the latest data indicate the possibility of acceleration to four months. This means that within the next one to two years, AI will be able to work autonomously on complex projects that previously took days and weeks.

Why this is important for everyone

The technology is already affecting professionals in the tech industry, and soon it will affect all professions. This is not about gradual automation of individual tasks, but about the complete replacement of cognitive labor. Legal work, finance, journalism, programming, analytics — almost all intellectual professions will come under the influence of AI.

Early involvement becomes a key advantage. It is best to use paid versions of modern models, integrate them into real work, and spend at least one hour every day experimenting with AI. This allows you to understand the capabilities of the technology earlier than most people.

Opportunities and risks

AI opens up incredible opportunities: writing applications, books, learning new skills, access to knowledge that was previously expensive or unavailable. But there are also risks: AI can be used for manipulation, creating weapons, or strengthening authoritarian control. The expert notes that the future depends on how well humanity can control technology.

The technology is already here, and the future is arriving faster than many expect. Those who begin to use AI actively and adaptively will gain an advantage. And for the rest, it will be a sharp and unpredictable blow.