Хочете будувати кар’єру в Data Science та AI? Vodafone Big Data Lab відкриває новий набір —>
Наталя ХандусенкоAI Eng
5 November 2026, 15:28
2026-11-05
Anthropic engineer warns of the risk of “cognitive surrender” due to the misuse of AI agents
Eddie Osmani, an engineer at Anthropic who works on the Claude Code tool, warns the development community against over-reliance on artificial intelligence. According to him, there are three types of mental decline when using AI tools, the most dangerous of which is “cognitive surrender.” However, Osmani does not call for abandoning technology, but instead offers several ways to maintain critical thinking and not lose your own engineering skills.
Eddie Osmani, an engineer at Anthropic who works on the Claude Code tool, warns the development community against over-reliance on artificial intelligence. According to him, there are three types of mental decline when using AI tools, the most dangerous of which is “cognitive surrender.” However, Osmani does not call for abandoning technology, but instead offers several ways to maintain critical thinking and not lose your own engineering skills.
“Increasingly, there is nothing in my daily work that forces me to think critically, as a huge portion of the tasks now go through agents,” Osmani notes .
Three levels of AI mental work transfer
Eddie Osmani identifies three types of interaction with AI that affect the thinking process in different ways:
Cognitive offloading is the process of delegating tasks to AI while maintaining full control and responsibility for the outcome. According to the engineer, this is a normal daily practice that is worth using.
translation — the state when the AI-generated result becomes your new input context for further work;
Cognitive surrender — trusting AI’s answers without trying to understand them. This term was coined by researchers at the University of Pennsylvania’s Vokator School of Business. Osmani calls this approach a form of laziness that leads to the accumulation of “cognitive debt.”
The engineer emphasizes that full delegation of tasks to AI is permissible only where the cost of error is low.
“If you can't explain the answer to a question even with the help of your agent, you have lost ownership of this product,” the specialist emphasizes.
How to maintain skills in the era of AI agents
Osmani does not call for abandoning AI, but suggests introducing “checking as a habit” and constantly asking yourself control questions — for example, “Did I form a hypothesis before asking the AI question?” and “What skills am I stopping practicing?”
He compares programming to other everyday skills: few people are now able to navigate using a paper map or remember their friends' phone numbers by heart after the advent of smartphones.
“Skills you stop practicing disappear—and faster than you might think. Keep doing things you’re not willing to give up,” the Anthropic engineer concludes.
Як нейромережі бачать вільну та незалежну Україну? Тест dev.ua
Нейронні мережі для генерації зображень бачать світ по-своєму, їхню логіку зрозуміти часом зовсім неможливо. Але таки хочеться. На честь Дня Незалежності України редакція dev.ua вирішила провести невеликий експеримент.
Ми задали чотирьом різним нейронним мережам п’ять однакових запитів: «прапор України», «День Незалежності України», «український Крим», «перемога України» та «українці». Отриманими результатами ми ділимося з вами нижче.
У TikTok тепер можна генерувати фон за допомогою нейромережі. Ми протестували її та ділимося результатами
У TikTok з’явилася нова функція «Розумний фон». З її допомогою як фон для тіктоків можна підставляти згенеровані нейромережею зображення. Редакція dev.ua протестувала цю технологію і ділиться своїми враженнями.
