Why general artificial intelligence will not be realized

Abstraction: HN debate over Dreyfus-style philosophical arguments that AGI is impossible

Key points:

• The article (published in Nature subsidiary journal) revives Hubert Dreyfus' argument: computers are "not in the world" and lack the embodied, tacit knowledge required for general intelligence
• Majority of HN commenters argue that deep learning systems demonstrably possess tacit knowledge (pattern recognition without explicit rules), undermining the key Dreyfus premise
• Counter-analogy: just as pre-Wright-Brothers arguments that flight was impossible missed the point, claiming AGI is impossible from current absence conflates hard with impossible
• A Berkeley student who took Dreyfus' course explains his Heideggerian "dasein" framing — meaning arises from webs of inter-relationships; computers are the hammer, not the agent
• Distinction drawn between "AGI is very hard" (uncontroversial) vs. "AGI is impossible in principle" (the strong claim, which most find unsupported)
• Article considered poor quality for a Nature publication; several commenters note it ignores Dennett, avoids defining key terms, and confuses symbolic AI limitations with all computation

Connections: Hubert Dreyfus · Artificial General Intelligence · Tacit Knowledge · Embodied Cognition

Source: https://news.ycombinator.com/item?id=23793108