Summary: elevation estimates using data from SRTM [1] are too high in built-up areas, because unlike with LIDAR data it's not able to tell the difference between "this is high ground" and "this is low ground with a building on it". Which means many coastal urban areas are more vulnerable to sea level rise than people had previously estimated.
They try to get a better estimate by building a more complex model, and calibrate it using LIDAR where that's available. This lets them say not just "things are worse than we thought" but "these specific parts of these populated areas are at risk".
If ice cliffs in the Antarctic won't collapse catastrophically that decreases our estimate for how quickly the sea is likely to rise, which is certainly good news!
But we're still going to get some rise, and accurate elevation modeling is important for figuring out what effects that's likely to have.
Roughly, you can think of these as two disconnected efforts: get the best estimate you can of the distribution of possible future sea levels, and get the best estimate you can of the impact of different sea levels on humanity. Your link is the former, this link is the latter.
They try to get a better estimate by building a more complex model, and calibrate it using LIDAR where that's available. This lets them say not just "things are worse than we thought" but "these specific parts of these populated areas are at risk".
[1] https://en.wikipedia.org/wiki/Shuttle_Radar_Topography_Missi...