The overhead cost for error correction improves as gate error rate declines.
The Google Quantum supremacy demonstrations confirmed the quantum world has huge computing resources. We are in the noisy qubit era. We can explore heuristic quantum algorithms. We might get near-term quantum advantage for useful applications but this is not guaranteed.
Near-term algorithms should be designed with noise resilience (noisy qubits) in mind.
We will get good truly random number generation and will explore new quantum simulation of complex systems.
Lower quantum gate error rates will lower the overhead cost of quantum error correction, and also extend the reach of quantum algorithms which do not use error correction.
Dequantization: Practical uses of quantum linear algebra and of quantum-inspired classical algorithms are still unclear.