For a start, the test for the container material was not at the advised temperature for the system of 4350F or 2400C. Then add the requirements for overrun and one might need to verify the system with temperatures as high as 20% above those cited herein. Now they describe passing the liquid Silicon through "tubes" ... and do not describe the material from which the tubes are formed; let alone the material forming the heating elements.
I for one would want to see a very detailed analysis of the design parameters before I would believe that the proposed energy storage system is a viable proposition; particularly for very long term use. It is one thing to test below temperature for an hour; quite another to satisfy a design that must need to operate for decades without intervention.

So solar energy is converted to electrical energy at %18 eff The Electrical energy is used to melt silicon at %95 eff Melted silicon is pumped through transparent tubes that can withstand 4000+deg heat Light from these tubes is converted to electricity at %45 eff. There are enough of these tubes and multijunction photovoltaics to generate about 150 megawatts of power (roughly enough for 100,000 homes). sounds great esp after the rigorous testing of building a mini tank and heatin it ot about %75 of the target temperature for an hour.
It sounds like some idea a first year student "team" with a limited budget and mechanical skills came up with.
I admit I have not read all the attached literature yet so maybe I am wrong, but it sounds pretty pie in the sky to me.

IMHO, best way to store energy is probably flywheels! What really needed is, R&D on flywheel energy storage devices of different sizes, from small battery sizes to very large grid sizes! Imagine, if they used magnetic levitation/bearing, in vacuum, to reach truly extreme rotation speeds (to store energy, far more of any other kind of battery)!

I love how we keep coming up with giant leap innovations. I did not know Silicon glowed at temperature.

I think they will need a gravity system with graphite lifts and boats and under vacuum. The PV's can be placed over the top of the boats. They will need massive cooling. The silicon could be the liquid first wall of a nuclear fusion reactor.

You have to wonder why the team at MIT fail to cite nor even acknowledge in passing the prior work done by the researchers at the Solar Energy Institute of the Polytechnic University of Madrid years ago. This really isn't an MIT thing. https://www.energynews.es/en/to-store-solar-thermal-energy-molten-silicon/

What is the theoretical end efficiency and projected life span of a unit that would power 100k homes?

I had a friend, many years ago, who came up with an idea that sounds much safer than a thousand degree tank of silicone. He used sand in aluminum pouches. It could have scaled up or down, for home use or grid use. Our federal agencies would not consider it back in the early 90's. I could never understand why not, other than it didn't involve the usual cast of big businesses and corrupt politicians. He had all the technical details laid out and formulas worked out. He even had a small stove unit in his house. Cheap and safe but no cigar.

Silicon melts at just over 1400C so providing they could come up with some sort of heat exchanger, that's more than enough to drive a turbine directly at about 40% efficiency. That's better than you'd get from a PV system.

If this silicon energy storage proves viable, I see no reason not to also use it in solar thermal power plants, instead of molten salt. This would greatly improve their efficiency. The MIT concept can be used in more places, but in very sunny areas, why not CSP, aka solar thermal, base load solar plants. And there is not the energy loss of conversion from electricity to heat or light, and back to electricity, that Captain Danger talked about.