With trade, there's always relative competitive advantage. It's probably economically more sensible for you the software developer to hire someone to mow your lawn, since you make more money in an hour of writing software than it costs to get someone to mow your lawn. Even if you're better at mowing your lawn than the person you hire - you still make more money writing software (example: You make $100/hour writing software. It takes you 1 hour to mow your lawn, or you can hire someone for 2 hours to do it for you at $20/hour. Opportunity cost to mow the lawn: $200. Gain from hiring someone to mow the lawn: $160. So even though you can mow the lawn twice as fast as the person you hire can, it's still better to hire someone.)
So now it becomes a resourcing problem - you need people who can mow your lawn. Robots can do that, but I don't think that robots can do all jobs at least not in the near future. Which is why I think that a huge part of this is a resourcing problem - you have a mismatch between skills and demand, which again I think is solved through education. Consider software developers again - I bet that if suddenly there was no demand for software developers, any one of us could be successful in another career, since at the heart of it, we're logical thinkers. We need to get as many people to that level as possible.
Furthermore, it's very important to not make poverty generational. I think that I contribute to society through my employment (at least I pay a ton of taxes), but if I wasn't as lucky as I was, I might not be in the same position to do so. I'd still be as capable, at least in terms of potential, but I wouldn't have the opportunity.
The lawnmower example isn't what we're seeing already now - we have structural unemployment of young average people who could and would do simple jobs like mowing lawns, flipping burgers and cleaning sidewalks; but there already are millions more of them than such jobs. And with every year, millions these simple jobs are going away, and unlike all the previous tech-job-shifts, this time the new industries aren't creating new jobs for them. Software developers are hiring other people for lawns and babysitting, but they are currently replacing much more people in "obsolete" industries with high-tech low-headcount disruptive business models.
I do agree with your comments about what's important and how it should be - but it's not really relevant when estimating what will actually be; wishful thinking won't solve it, especially as it's clearly beneficial for the powerful to ignore this problem as long as possible.
Both distopia and utopia is possible, but the distopia is much more likely due to the capitalist system we have, and the rich/poor gap is only going to grow.
We'll have good, decently skilled, enthusiastic, hardworking "zero marginal product" people simply because software/robots will be able to do it cheaper than even a subsistence (calories+shelter) wage, and in most tasks employing a human for it would be a senseless luxury.
You can hire a bunch of people to carry you around in a carriage on their backs, or you can hire people to do arithmetics for you - but almost noone does. Soon (in my and your lifetime) the same will be for all the manual, exhausting jobs; for most of the driving, service, cleaning, delivery, and retail jobs; for most of the simple clerical jobs. Even if there is no progress in AI at all, we'll come to a state where, say, 10% of population can cover all major needs (agriculture, food production, service, clothing, construction, consumer goods manufacture, transportation, medical, legal, etc) of everyone. What we'll do with the other 90% (or 99%, or even 100% if AI happens) ? They will be nice, great, fellow humans, but practically speaking, the society can do with them and can do without them. That's a choice for the society - utopias or distopias are both possible.