This is one of those patents that absolutely in no way signifies that the company who files it is working on the thing that the patent is patenting. Here’s the basic idea: you’ve got a cavity that’s a sort of hemisphere shape, kind of like the business end of a rocket engine. You toss a pellet of fuel into that cavity, and then lasers blast the fuel pellet, causing it to release a bunch of energy (by exploding, fissioning, fusing, or whatever). That energy pushes against the walls of the cavity, and the cavity moves forward. At the same time, the explosion heats the walls of the cavity, and this heat is harvested to drive the lasers.
Let’s just go through this really quick. First, free-electron lasers. That’s a thing. We have those. They’re like regular lasers, except they use the wiggling of electrons to generate light, which is nice because you can tune a free-electron laser through a huge range of wavelengths, all the way from microwaves to X-rays. In what is almost certainly not a coincidence, Boeing has been working on with the goal of putting it on a boat, because you can’t fit the required electron accelerator (plus a whole bunch of amplifiers) into anything much smaller than that. (I’m looking at you, jet engine.)
Now, about this idea of using lasers to cause fusion inside a pellet of fuel: we’re working on that, too. At the National Ignition Facility, they’ve managed to fuse small pellets of deuterium and tritium using 192 lasers, reaching breakeven in terms of the amount of energy deposited into the fuel and the amount of energy that the fuel releases. Note that “the amount of energy deposited into the fuel” is only a tiny fraction of the amount of energy sucked down by the lasers, but in principle, it works.
As for the specific impulse that this hypothetical laser fusion engine would produce, that’s an absolutely bananas number for efficiency. I discuss what specific impulse means in this article about Hall effect thrusters, but essentially, it’s an efficiency measurement, and the most efficient engine that we’ve ever produced (even as a prototype) has a specific impulse of just under 20,000 seconds. So we’re talking an entire order of magnitude more efficient.