>>509689984Okay there’s an hour long video I just linked you of a robot doing a mail sorting job previously done by humans.
They don’t go on break, don’t take vacation, don’t sleep, don’t need employee benefits, don’t need HR, don’t have “off days” ie don’t waver in efficiency.
Take a look at the capabilities of current, publicly released SOTA models and tell me they can’t do a cashiers job (we already have self-checkout…).
GPQA benchmark (PhD level STEM questions that are designed to not be “googleable”): https://arxiv.org/abs/2311.12022)
SOTA models now scoring >85%, actual human PhDs only scored ~74%.
Updated LLM scores: https://www.vellum.ai/llm-leaderboard
USMLE (Physician licensing exam)
https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0000198
Bar exam (Lawyer licensing exam) https://www.abajournal.com/web/article/latest-version-of-chatgpt-aces-the-bar-exam-with-score-in-90th-percentile
Turing test https://www.livescience.com/technology/artificial-intelligence/open-ai-gpt-4-5-is-the-first-ai-model-to-pass-an-authentic-turing-test-scientists-say
Remember. ChatGPT came out in Nov 2022, less than 3 years ago and this is where we stand currently.
I’m not going to play your gay little “predict the future” game and give you an exact timeline. But it doesn’t take a genius to see the trend.
Also, if you think these companies are just going to copy paste their current, publicly facing, general purpose LLMs into their enterprise applications and are thus basing your perspective of job threat on that then you’re beyond retarded.