What if fusion is impossible though, from an engineering perspective? Why not continue to pursue fission using safer, modern reactor designs? It works. The timelines for fusion, if it ever works, are too long.

Most polluting regions are also the least politically stable. Today at any moment disaster could happen on largest nuclear power plant in Europe because madman decide to play a conqueror.

People can hardly be trusted with a dangerous nuclear power.

We already have a conveniently-placed fusion reactor that's boing to be operational for a few billion years.

Still need to be concerned with the amount of latent heat generated on the planet surface

Eventually yes, but that's not the immediate problem.

400 years to boil the oceans, but that kind of timescale means we can have 300 years of business as usual followed by using a von Neumann probe to rearrange Mercury into a Dyson swarm, at which point the Earth's temperature is under total external control because we can switch the sun off when we don't like the heat.

OK, so more like "hiding" the sun — switching it off is still possible with a Dyson swarm but I don't know how long that takes — but from the point of view of Earth it might as well be switching it off.

Latent heat can be radiated into space, and offset with a relatively small, invisible shade at the earth-sun Lagrange point. 10,000 1km aluminum foil and Mylar sails surfing that point would make earths temperatures completely controllable and would be measurable only as a dimming of the sun imperceptible without sophisticated measuring apparatus.

Starship can carry 2 per trip. Inflatable, polymer hardening, weigh-shift controlled kites gliding just beyond L1 on the solar wind. We would have to launch 3 a day indefinitely, so we would still be encouraged to sort out our CO2 situation.