Of course adding enough lanes to meet demand is not worth the cost. However the induced demand argument is wrong, it is a cost benefit calculation: will the additional people who are able to get someplace be worth the cost to add the lane.
Transit does have the same problem, but transit can scale much better in a given amount of space. A bus every 5 minutes can handle more traffic than a lane filled with cars - assuming the bus doesn't compete with the cars. You can run a bus every 3 minutes, and if that isn't enough a train every 45 seconds holding 5 times as many people as the bus. If that isn't enough - okay, it is time to build another track for the train, but that is a huge number of people (ie a good number would find their life improved by a track on a slightly different route that)
Mind: shared rights-of-way modes (busses, streetcars, LRT) can be impacted by private ve traffic, but contributes little to that itself.
Fixed-track systems (Muni Metro) can have coordination delays if multiple routes join.
It is very true but the point of adding capacity (to any form of transit) isn't to reduce congestion at peak hours (that congestion is bounded by what people will tolerate) it's to allow more trips to happen with equivalent congestion thereby increasing the total number of trips and/or decreasing the duration of the congestion.
Even in SF, there's some places where public transportation is faster. But in very very limited situations because SF transportation is horrible. But an express Caltrain is often much faster than driving MTV-SF if there's any traffic. With all the (super delayed unfortunately) infrastructure projects going on (new SF subway being installed, Caltrain electrification), things will hopefully get better...