133 MHz Arm has an order of magnitude more power than an 8 MHz 68000?
I don't think so! He's apparently forgotten that M68k takes 8 clock cycles for a 32 bit register-to-register add/sub/cmp/and/or/eor while the Arm takes 1 clock cycle for this. Shifts and rotates on the M68k take 8+2n cycles where n is the number of bits shifted by -- the Arm takes 1 cycle regardless of the shift amount. Conditional branches take 8 cycles on the M68k even if not taken, 10 if taken. The Pi Pico takes 3 cycles if taken, only 1 if not taken.
Generally speaking, the Pi Pico is going to be more than TWO orders of magnitude faster than the Mac 128k, when running native code.
Of course 68k emulation will eat up some of that. But even if it's a pure interpreter (like in the first PPC Macs) that should eat up only one of those orders of magnitude.