In 2016, the AI pioneer Geoffrey Hinton declared that 'people should stop training radiologists now' because 'it's just completely obvious that within five years, deep learning is going to do better than radiologists.' He was half right. Today, the FDA has approved more than 1,000 AI radiology tools, some capable of analyzing medical images to detect injuries or diseases with greater accuracy than human specialists. Yet radiologists'human ones'are in more demand than ever. Since 2016, the number of radiologists has risen by 17 percent, the field's vacancy rates are near all-time highs, and the average salary has increased from about $350,000 to $570,000, making radiology the third-highest-paid medical speciality in the United States. Many people now fear that AI will make a huge number of careers obsolete. Last year, Anthropic CEO Dario Amodei claimed that AI would soon 'wipe out half of all entry-level white-collar jobs.' But the radiologist story suggests that whether AI will replace a given profession is not so straightforward to predict. Answering the following three questions can help you determine how endangered a job really is....
The Transformer is currently the reference architecture for serious AI. Not because it is obviously the most brain-like, elegant, or efficient design, but because it has the best scaling story. You add data, parameters, compute, context length, better training recipes, better post-training, and the model gets better in a surprisingly smooth way. That is rare. In deep learning, many ideas are clever. Few are industrial. The Transformer's superpower is attention. Every token can look at every other token and decide what matters. This is an incredibly general operation. It works for language, code, images, audio, video, protein sequences, robotics tokens, and tool traces. The architecture is simple enough to scale, parallel enough to train efficiently, and expressive enough to absorb huge datasets. But it has an obvious tax: attention is expensive. Full self-attention scales badly with sequence length. In autoregressive generation, the model accumulates a key-value cache, which grows with context. A Transformer remembers by keeping a large, explicit, token-indexed memory. That is powerful, but it is not how you would design every intelligent system from first principles....
' 2026 Conde Nast. All rights reserved. WIRED may earn a portion of sales from products that are purchased through our site as part of our Affiliate Partnerships with retailers. The material on this site may not be reproduced, distributed, transmitted, cached or otherwise used, except with the prior written permission of Conde Nast. Ad Choices...
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time. Karpathy started this week at Anthropic, where he is working on pre-training under team lead Nick Joseph. Pre-training is responsible for the large-scale training runs that give Claude its core knowledge and capabilities, according to the company. It's also one of the most expensive, compute-intensive phases of building a frontier model. Karpathy is one of the few researchers who can bridge the gap between LLM theory and large-scale training practice. Tapping him to build such a team is a clear sign from Anthropic that it believes AI-assisted research, rather than pure compute, is how it stays competitive with OpenAI and Google. While at OpenAI, Karpathy focused on deep learning and computer vision until he departed in 2017 to join Tesla. He led Tesla's Full Self-Driving (FSD) and Autopilot programs before leaving in 2022. He then went back to OpenAI for one year before leaving again in 2024 to start Eureka Labs, a startup dedicated to applying AI assistants to education....
