Microchip's Beloved MCP2221 USB-I2C Datasheet Has Mistake?!

[MicroBiggerThanMM]

Edit: No it doesn't. Murphy caused the PCB to be wrong. That being said, it's annoying that things can't be routed this simply.

Hello all,
The MCP2221 is a nifty alternative to FTDI with UART's, DAC's, ADC's, GPIO's, and I2C for $2. 

But, the datasheet says V_DD is the power input and V_USB is the output of the internal LDO.

Okay, counterintuitive, but I designed the circuit, tossed my $100 of components on the thing, and when I measure it, it turns out it was counter-counterintuitive. 
Applying 5V to VDD results in 5V on VUSB, and 5V on VUSB results in 3.3V on VDD.

Did Microchip mess up in figure 1-2? I feel like they should have been more explicit about using the internal LDO to get a 3.3V supply.
Maybe I just designed everything really late at night. 

[ataradov]

You are doing something wrong. The datasheet mentions in multiple places that V_DD is an input and V_USB is an output of the regulator.

Electrical characteristics mention:
V_DD 3.0 - 5.5 V
V_USB 3.0 - 3.6 V

It is possible that something is broken after you applied 5 V to V_USB. Absolute maximum ratings mention "Voltage on V USB pin with respect to V SS ... –0.3V to +4.0V"

[MicroBiggerThanMM]

Doh! The PCB was switched around