The cancer stem cell (CSC) hypothesis has been disproved in many cancers. CSCs may exist in blood cancer, while many epithelial cancers may not have CSCs but tumor-initiating cells (TICs). Several independent studies have provided strong evidence for existence of CSCs in brain, skin, and colon cancers (Mani et al. in Cell 133:704-715, 2008, Joseph et al. in Cancer Cell 13:129-140, 2008, Reya et al. in Nature 414:105-111, 2001), while the CSC hypothesis remains controversial (Magee et al. in Cancer Cell 21:283-296, 2012). Liver TICs have bipotential to give rise to two different lineage types: hepatocellular carcinoma (HCC) and cholangiocarcinoma (CC). In the liver cancer field, the origin of HCC and CC is extensively debated. Several groups have validated that TICs gave rise to HCC and CC. Hepatocytes gave rise to HCC. Several groups have demonstrated that oval cells (or liver progenitor cells) give rise to TICs. However, CSCs may be a myth in gastrointestinal cancer, while many groups have validated liver TICs. The definition of CSCs includes pluripotency, while TICs do not have to have pluripotency and only need to have bi- or multipotential to give rise to diverse tumor types and tumor initiation potential in mouse models. The CSC hypothesis therefore controversial (Magee et al. in Cancer Cell 21:283-296, 2012). Cancer tissues contain subpopulations of cells known as tumor-initiating stem-like cells (TICs, so-called CSCs) that have been identified as key drivers of tumor growth and malignant progression with drug resistance. Stem cells proliferate via self-renewing division in which the two daughter cells differ in proliferative potential, with one displaying differentiated phenotype and the other retaining self-renewing activity.