AGI fantasy is a blocker to actual engineering
8 days ago
- #Ethics in AI
- #OpenAI
- #AGI
- OpenAI and its associates genuinely believe in the creation of AGI (Artificial General Intelligence), viewing it as either humanity's salvation or doom.
- Elon Musk founded OpenAI partly to counter what he saw as the threat posed by DeepMind's Demis Hassabis, whom he likened to a supervillain.
- Ilya Sutskever, OpenAI's co-founder, dramatized the concept of AGI with a symbolic burning of an effigy representing a deceitful AGI, emphasizing the need to destroy it.
- The 'pure language' hypothesis, which posits that AGI can emerge from language models alone, gained traction at OpenAI following the success of scaling from GPT to GPT-2.
- The pursuit of AGI justifies massive resource consumption, including significant environmental impacts and exploitation of data workers, under the assumption of its enormous potential value.
- Expected value (EV) arguments for AGI are criticized for being based on speculative probabilities and ignoring tangible negative externalities like environmental damage.
- The author advocates for a pragmatic approach to technology, focusing on solving specific problems efficiently and ethically, rather than chasing the AGI fantasy.