Is it undemocratic for the candidate with the most support from the majority of voters to win?

This is a very strange attack on ranked-choice voting from The Atlantic:
Cuomo is likely to get more first-choice votes than any other candidate. But he’s not projected to win an outright majority, meaning that the ranked-choice system would kick in. Candidate after candidate would get knocked out, and their supporters’ votes reapportioned. In the end, the political scion with a multimillion-dollar war chest and blanket name recognition could lose to the young Millennial whom few New Yorkers had heard of as of last year. One new survey, by Data for Progress, shows Cuomo ultimately defeating Mamdani by two points, within the margin of error. Another poll shows Mamdani with more support than Cuomo.
Seeing a no-name upstart attempt to upset a brand-name heavyweight is thrilling. But the system has warped the political calculus of the mayoral campaign. Candidates who might have dropped out are staying in. Candidates who might be attacking one another on their platforms or records are instead considering cross-endorsing. Voters used to choosing one contender are plotting out how to rank their choices. Moreover, they are doing so in a closed primary held in the June of an odd year, meaning most city residents will not show up at the polls anyway. If this is democracy, it’s a funny form of it.
Needless to say, the fact that ranked-choice voting doesn’t throw away critical information about a voter’s preferences is exactly what’s good about it — it would in fact be bad for Cuomo to win as a plurality candidate if he would lose head-to-head matchups to the other frontrunners because a majority of voters (as Lowrey goes on to concede) don’t like him. Having candidates cross-endorse rather than dropping out is different than what happens in first-past-the-post but it’s unclear and unexplained how this is undemocratic. Indeed, the argument against ranked-choice is so weak and unexplained that she has to smuggle in a non0-sequitur about low-turnout off-year elecions, which is an accurate critique but one that applies irrespective of the voting system.
The arguments don’t improve:
Ranked-choice voting might better reflect voter preferences, but it is chaotic, requiring extra strategizing by both candidates and voters. To keep Cuomo out of Gracie Mansion, some candidates have said that they are contemplating cross-endorsing Mamdani, telling their supporters to rank them first and him second. Unions and political groups are endorsing multiple candidates; many are pushing a simple “Don’t rank Cuomo” message. (Ramos, an exception, has thrown her support behind Cuomo while remaining in the race, saying he has “experience, toughness, and the knowledge to lead New York.”)[…]
Whether Cuomo or Mamdani wins this month, New Yorkers might have another chance to decide between them. After this annoyingly chaotic primary, we could have an annoyingly chaotic election: If Mamdani loses, he might run in the general on the Working Families Party ticket. If Cuomo loses, he might run in the general as an independent, as will the disgraced incumbent, Eric Adams. At least, in that election, voters won’t be asked to rank their favorite, just to pick one.
When it comes to the democratic legitimacy of a system, I think we should stop with “better reflect voter preferences.” I don’t think the point about extra strategizing by voters is necessarily accurate — it actually allows voters to vote their sincere preferences rather than try to figure out ex ante which candidates in a crowded field will appeal to other voters — but even if it did, so what? Accurately conveying the preferences of voters seems obviously more important to democratic legitimacy than simplicity. (Would we rather the 2000 presidential election system in Florida be simple, or convey the accurate preferences of the state’s electorate? Many hundreds of thousands of people are unavailable for comment.)
The real impetus of the article seems to be frustration that some of the more experienced non-Cuomo candidates aren’t getting more traction than Mamdani. Knowing what I know now, I would tentatively rank Lander first if I had a vote. But it’s also neither here nor there — the also-rans are not winning under any electoral system if the polls are remotely accurate. If a majority of voters would prefer Mamdani to Cuomo in a head-to-head matchup — well, I think they’re right, but either way he should win. The fact that in the general a multi-candidate race could lead to a candidate opposed by a majority of voters winning is bad, not deserving of a relieved “at least” because ranked-choice is too much work or something.