Short Answer: Your friend is wrong because our models of friction (within certain parameters) are derived from Newton's laws. Tidal forces, for example, are are a form of friction.
The argument that Newton's "laws" are "invalid" in an pure philosophical sense, better rest on Newton simply pulling the concepts of gravity, force and inertia out of thin air in order to give meaningless names to parts of his highly predictive geometric and mathematical models.
One can also argue that since they remain accurate only within a specific range of measurements, they capture any real truth about reality but just approximate it to some degree.
It's very, very important to remember that scientific "laws" or "models" do not predict concrete reality, they predict measurements. New scientific laws/rules/models arise when the old laws/rules/models failed to be able to reproduce new measurements.
TL;DR unless you've got time to kill
The core scientific method is take a large number of measurements and create a huge data set. Then create a geometric, mathematical (or as some now argue, computational) system that will, given one part of the set, be able to reproduce another part of the set. Scientific laws are really mere convinces to compress vast number of observations down to mathematical equation, geometric ratios or (as some argue today) a computation.
That scientific laws reproduce sets of measurement data and not reality is seen easiest in statistics, which is really a second layer of math/modeling that tells us how much off from reality or measurements might be e.g. no one has 2.1 kids.
But... the equations don't always reproduce the observed measurement sets unless we cheat a bit an cram in some arbitrary numbers i.e. numbers not derived from measurement, don't seem to represent any observable or measurable phenomena, but which nevertheless, make the equation predictive of its data set.
The great-grandday of them all is Newton's gravitational constant. Cavendish just measure fine gravitational deflections, then started punching in numbers till he found one reproduced the deflection in the data.
There are many constants crammed into scientific equations for which we have no explanation of why the constants are needed or what they represent. A large number of constants in the math of a system is strong predictor that the model will fail once superior measurements become available.
One of the sources of the great friction between Leibniz and Newton was that Newton just made things up. Things like "gravity", "force", "inertia". Leibniz following the French Cartesian idea that all motion in the universe resulted from the collisions of particles and that positing a mysterious "force" called "gravity" instantly began pulling to objects towards each other, or that objects kept moving unless acted upon by another object or "force" all bordered on outright mysticism.
Newton replied that his geometric and mathematical rule-sets/models made vastly better predictions than anything Leibniz had come up with and that naming the parts of the model was just a conceptual connivence and that he would even try to guess what inertia, forces or gravity were, or were caused by, he merely asserted that using the concepts worked very well in making predictions.
But the idea that forces, gravity and inertia were just arbitrary names, didn't catch on and most people still talk about them as if they exist and have some concrete reality we understand.
Because scientific models reproduce measurements and not reality, we often evoke metaphors, call them "as-if" fantasies we use to explain what phenomena we can't observe and measure causes the phenomena we can.
Thus Newton modeled gravity between two objects "as-if" they were instantly connected by a contracting elastic band ( or a non-streching string on the axis of an ellipse. Likewise, Einstein modeled gravity "as-if" nothingness was elastic, bent and curved and caused objects follow the bends and curves.
Newton modeled gravity as a "force" that transmitted instantly between objects. Later, scientist showed that it propagated at the speed of light. Einstein, then modeled gravity not as a "force" like an elastic band between objects, but as each object "bending" space around it such that any object traveling through the bent space found itself deflected toward the initial object.
We're still doing it. The Higg's particle isn't a particle but a convenient way to model changes in the Higg's gauge field (I think.)
But it bears repeating that scientific laws/rules/models reproduce/predict measurements, not the reality that scientist attempt to measure. Since we can't measure with perfect accuracy, our scientific models will always be off a bit from reality.
But "a little bit off" is just fine when your calculating the tossing of a horseshoe blowing something up with nuke.