Gregory's Server
@bedast @x_cli Neural networks are still OK — there is a good technology started in 1970, which helps people. AI (I mean real AI) — will be cool and futuristic too. But current "AI" (LLM) hype-train is not cool, it is not a future we wanted (who want to live in cyberpunk with "high tech low life"?). Disadvantages we already know. Advantages — may be will be visible when hype train will stop (improved natural language processing? help in decoding old and lost languages?).
(3/3)
@evgandr
One situation changed dramatically between the 70's and now: climate change and the understanding of its anthropic source. From a purely computer/math science standpoint, I share your opinion.
From an holistic standpoint, I consider that all commercial AI are criminal, in the 2020's. I am more lenient regarding models developed by state nations for the purpose of a greater good for all living beings. Incidentally, I don't want any AI in any end consumer product, unless it is scientifically proven (before the model training starts) that it improves the situation regarding the climate crisis.
@bedast
@evgandr
One situation changed dramatically between the 70's and now: climate change and the understanding of its anthropic source. From a purely computer/math science standpoint, I share your opinion.
From an holistic standpoint, I consider that all commercial AI are criminal, in the 2020's. I am more lenient regarding models developed by state nations for the purpose of a greater good for all living...