i will present the connection between two topics which seem completely unrelated, other than by a fortuitous common name. the original shuffle conjecture equates two symmetric polynomials $phi_n[x;q,t]$ and $pi_n[x;q,t]$. the first was conjectured by garsia-haiman in the early 1990's to produce the bigraded frobenius characters of diagonal harmonic polynomials. the second was defined by haglund et al around 2002 as a weighted enumeration of parking functions in the $n \times n$ lattice square. in a recent paper, hikita constructs a new polynomial $pi_{m,n}[x;q,t]$ for any pair of relatively prime natural numbers $m$ and $n$, by extending the notion of parking function to the case of an $m \times n$ lattice rectangle. it follows from hikita's construction that $pi_n[x;q,t] = pi_{n+1,n}[x;q,t]$. the shuffle algebra is a representation-theoretic object constructed by feigin and odesskii, which acts on the k-theory of the hilbert scheme of points in the plane. the k-theory is isomorphic to the ring of symmetric functions in infinitely many variables, so the shuffle algebra action produces many interesting symmetric functions. in particular we are able to construct a new polynomial $phi_{m,n}[x;q,t]$ which reduces to $phi_n[x;q,t] for m=n+1$, and conjecture that $phi_{m,n}[x;q,t] = pi_{m,n}[x;q,t]$ for general coprime $m,n$. in this talk, we will explain the ample connection that led to this extension of the shuffle conjecture. this is joint work with eugene gorsky.