Prompting ChatGPT for analysis, H1-B edition
One of the cool things about ChatGPT is the ability to gather and aggregate data about various things, and present such data in a relatively objective format. It's not always totally correct (see: "stochastic parrot"), but it does a reasonable job presenting well-stated conclusions with supporting data, particularly when the conclusion is already pretty well-known. Below is the prompt result from asking the LLM if the H1-B visa program is being used as-intended, or if it's being abused to effectively underpay US workers (with an ancillary focus on Zscaler, because that's my current employer, and I didn't know if they used H1-B visas at all before querying for this).