I expect that you lose efficiencies with respect to security. I.e., it's not easy to secure one large nuclear reactor (fission or fusion), but it's a lot easier than securing dozens of small reactors.
Drop a chunk of U238 in/near the fusion reactor somewhere, let it absorb stray neutrons, and you'll soon have some plutonium. That's bad for control of nuclear weapons proliferation, which relies on high-energy neutron sources like breeder reactors being difficult to build and expensive.
Not sure about the radiation profile from this system though.
The energy output of HB11 consists of fast-moving alpha particles, which is good because that's the energy you capture, along with x-rays. None of this would activate reactor components. The waste product is non-radioactive helium.
https://en.wikipedia.org/wiki/Aneutronic_fusion#Boron
Neutron energy would be a bit under 3 MeV, comparable to fast fission neutrons but in much lower quantity for the power output.
https://en.wikipedia.org/wiki/Neutron_temperature
Interestingly, that page says the thermal neutrons in a conventional nuclear reactor are better at activating materials than fast neutrons.
According to a presentation I saw by the leader of MIT's fusion program, the inner wall of a D-T tokamak would be activated by the neutrons, but would only need to be stored for a few decades. D-T releases 80% of its energy as 14 MeV neutrons.
According to LPP, which is working on a different boron fusion design, a 20MW reactor would activate the walls with short-lived radiation, but it could be safely opened for maintenance after about twelve hours (iirc).
This is something different. With every laser shot there will be a burst of alpha particles. Then they'll slow down and just be plain old helium. If the direct energy conversion works, they'll get slowed down very quickly.
Only if you eat or inhale them. A sheet of paper will stop alpha particles.
It might also be worth pointing out that "It must be safe! They even put it on wristwatch dials!" is not, historically, a sound argument.
This depends on the threat model to some degree, but there's security in knowing that it's vastly harder for a threat to affect multiple targets at once, no matter the threat.
That might be different for this fusion device, but for outdated looking fission reactors it’s artificial dilution for NPT purposes and by no way a technological limitation.