Once a machine can rewrite its own code, the leap from human-level to superhuman may be far shorter than the leap to human-level.
If you have a human-level AGI in 2029, it is going to be able to optimize its own code and analyze its own foundations and exceed the human level in much less than the sixteen-year gap Kurzweil drew between AGI and superintelligence. This is recursive self-improvement: the AI makes a smarter AI, makes a smarter AI. We’re actually designing Hyperon precisely for recursive self-improvement, which the human brain never evolved to do.
Source: Rethinking the AGI Race, with Benjamin Goertzel (eCornell Keynote)