In this talk, I will present several unconventional analog IC design techniques. First, I will talk about how we can make use of noise, which is usually deemed as an undesirable thing, to estimate the conversion residue and increase the SNR of a SAR ADC. It is an interesting example of stochastic resonance, in which the increase in noise can lead to not SNR degradation but SNR enhancement. Second, I will talk about how we can perform data conversion below the Nyquist rate by exploiting the sparsity of the input signal. I will show two examples compressive sensing ADCs and how the effective ADC conversion rate can be reduced by 4 times but without losing information. Third, I will discuss the challenges of the classic opamp-based analog signal processing (ASP) framework in advanced nanometer CMOS processes. I will then present novel phase-domain ASP techniques and show how to use them to build scaling-friendly and power efficient mixed-signal circuits.