The brisk progress of artificial intelligence (AI) is fanning fears of job displacement yet again. At the ongoing World Economic Forum (WEF) in Davos, while interacting with moderator Zanny Minton Beddoes and Google DeepMind CEO Demis Hassabis, Anthropic CEO Dario Amodio touched upon how his staff have forgone writing code and are entrusting AI models to do the same.
The CEO told the host that the world is now in a phase where engineers are increasingly using AI and have stopped writing code. He shared that his staff at Anthropic use AI to write code, which they later edit, and do things around it. Amodei further added that we may be six to 12 months away from when AI models will be doing all that is done by software engineering services.
“We are now, in terms of the model that writes code; I have engineers within Anthropic who say I don’t write any code any more. I just let the model write the code, I edit it, and I do the things around it. I think we might be six to 12 months away from when models may be doing all that SWEs (software engineering services) do end-to-end. And then it’s the question of how fast that loop closes,” Amodei said during the ‘World After AGI’ session at the WEF.
Coding has been seemingly one of the tasks that has seen an increase in the use of AI. Once the bastion of programmers, with AI, coding has become more accessible. Last year ‘Vibe Coding’, a software development method where users rely heavily on large language models or AI agents to create applications, dominated the news feeds. The trend showcased how applications can be built using simple instructions in natural language, and that too in a conversational format.
At the beginning of the session, Beddoes reminded Amodei of his comments at WEF 2025, where he had said that there will be models that can do everything a human can do at the level of a Nobel laureate across many fields by 2026-27. When asked if he continued to abide by that timeline, the Anthropic boss said that ‘it was always hard to know exactly when something will happen. Amodei added that regardless of the timeline, his company will continue to produce the next generation of model and speed it up to create a loop that would increase the speed of model development.
Amodei also warned that not every part of the loop can be sped up by AI, as there are chips, the manufacture of chips, and training time for the model. Even in the wake of uncertainty, the CEO said that there is no reason for him to believe that it could take a few years. He said all of this is moving much faster than people would imagine. The key is the element of code, and increasing research capabilities are the key drivers. “It is really hard to predict how much…but something fast is going to happen.”
Anthropic’s Claude has been one of the standout AI tools that has consistently pushed its coding capabilities. From chatbot to agentic coding partner that is capable of managing the entire development workflow, Claude has emerged as one of the best AI tools for coding. Last year, Anthropic introduced Claude Code, which is a terminal-based agent, Computer Use (autonomous desktop interaction), and upgraded Artefacts for real-time code rendering. The company has demonstrated consistent advancement in its model capabilities and has expanded its API and ecosystem.
Expand © IE Online Media Services Pvt Ltd