Historical note: x86 is a bastadised rip-off of the PDP11 instruction set.
And as with most technological descendents, the folks who did the job botched it. Incredibly obtuse instruction decoding, special instructions that do five things at a time (most of which are not useful), and horribly slow to interrupt and restore.
The PDP11 was built as a "hardware Fortran machine" ie one instruction represents one Fort[r]an instruction as far as was achievable in 1970.
Uh, not really. The PDP-11 was designed as a general-purpose ISA, used as much for assembly code as Fortran. In addition, it hosted four OSes (RT-11, RSX-11/M, and RSTS/E from DEC and UNIX from an odd place called Bell Labs). The different OS'es used different tools. A lot of RT-11 code was used for industrial control and was done in Assembler (did some of that), RSX-11/M was their mainline OS for applications and was programmed in COBOL (the implementation here sort of sucked) or FORTRAN (a pretty brilliant implementation) or Assembler, and RSTS/E was an odd duck that had a BASIC interpreter. UNIX had C. The best thing about programming on the -11 (besides the nice, relatively orthogonal ISA) was the FORTRAN automatic overlay feature. It let you bundle code into overlay segments that were automatically swapped in when routines in the module were called. A performance killer when used improperly, it was the only way I could fit a FORTRAN program that took 320K on an IBM\360 into the PDP-11's 64K.
C is (just one) PDP11 assembly language!
I don't think I'd go that far. There were many things (conditional branches on overflow, control of interrupts/traps, computed gotos) that, although accessible via assembler, could not be easily done in C. That's why today you still have assembly modules and/or use of inline assembly in UNIX code.
The VAX instruction set was an attempt to achieve a higher level machine code, which worked quite well - most VAX assembly instructions are actually function calls to application specific microcode.
As were most instructions in those days. As for "worked quite well"? Well, there was that whole RISC/CISC thing going on and, you know what? RISC sort of won the technical war - it may be papered over with an ugly CISC instruction set on the inside, but internally, it's all condensed onto execution on a mostly RISC core.
X86 was a poor ISA when the first 8086 chips were made (but good, given hardware capabilities at the time). That was about 40 years ago. MIPS and Sparc (and ARM) are all better than x86.
Well, yeah. They have the benefit of hindsight and much less self-inflicted baggage. On the other hand, that baggage has kept Intel in the game while they try to catch up to ARM in power consumption.