Raw LLM Responses

I wouldn’t put too much stock into this. 1. This study relied on data provider that aggregates job postings. We don’t fully understand how representative this data is or what kind of blind spots it has. For example, what about positions filled via referrals or people switching roles inside a company (lots of devs calling themselves AI engineers after integrating LLM into their companies product)? 2. Are job postings even a good proxy for how many people have a certain role? Let’s say a company fires all it’s support agents to replace them with AI, notices 6 month later that that was a bad idea and hires 20% back. Now if these 20% make it into the dataset it looks like the company just went on a hiring spree, but doesn’t capture that they still reduced the positions for that role by 80%. 3. Even if you had perfect data into job growth and reductions, it doesn’t tell you anything about who that is happening. The study makes some guesses based on priors, but it’s still just reading tea leaves. Lots of other things affecting the job market are happening not just AI. There may very well be jobs that are being automated with AI but are still growing for other reasons, that may or may not be related to AI.