It failed because the intent was to cut costs and they got what they paid for. The successful ones were genuinely trying to expand their engineering talent pool and quickly figured out that the cost savings were a marginal benefit that made up for some of the extra overhead of international accounting and management. Quality engineers are one visa lottery away from Western salaries so the local median salary is often completely disconnected from what a FAANG might pay for a decent engineer, which is a rude awakening for anyone trying to cut costs without destroying the quality of their output. On top of that, the people most likely to make it a smooth transition are also the people most in demand (arms race!) and competition for them helps evaporate any savings for the business.
The guy in Argentina is in a great position to get hired to work remotely for an American company, but I don't seem him as competition regardless of how good he is. It's an arms race so my employer's competitor can't replace their team with Argentinians because that down time will give my employer time to crush them (which we learned in the 2000s). They can hire an extra team of Argentinians on top of their existing head count but if they do that my employer will be pressured to hire a team of Brazilians. Before you know it, both companies are hiring even more local teams to help manage the flow of work between their existing local teams and the outsourced ones.
There's more work and money to pay for it than there are people to do it. Until that changes, we're not the ones competing, the employers are.
This has worked far better than the 'Tata' chapter which exported to very cheap Asia where staff tech skills and the support model were too often haphazard and frequently underperformed. Hiring remote staff in new-growth US cities with lower CoL (e.g. Charlotte NC or Austin TX) has delivered capable staff and nearly synchronous operating hours to the main office. However in the past 2-3 years, IT labor costs have risen greatly in 'secondary' US cities, reducing those savings substantially.
In my experience, remote staff in eastern Europe have skills equal to US folks, but the 5 hour mismatch in daytime hours and their inability to visit US sites where non-IT R&D work is done and data is acquired does hamstring this model.
I wouldn't be surprised if a 4th era of outsourcing arises soon: remote work in small towns across the US and in western Europe. European IT wages often seem to be 50% or more below US large cities. That differential is likely to diminish as remote work increasingly is established as a new norm.
You can do bigger differences and many of us do on occasion. But on a multiple times a week basis, both 5am calls and 11pm calls get old.
My intuition is that a max of 3 or 4 time zones is the sweet spot, but it's cool to hear that you've had success up to 6.
I would argue that outsourcing failed primarily due to companies trying to farm out the coding to uninterested entities whose incentives did not align well. Not necessarily because the foreign workers doing the coding were bad.
Back to the guy in Argentina -- I imagine that he/she is actually in a reasonably good position as the amount of remote work increases. Indians not so much, because they are 12.5 hours ahead (of Pacific time). Argentina is only 4 hours ahead, which makes for a lot more overlap.
So I think the field has leveled a little bit in that sense, because if you have a remote developer in another state, they are not very different than one in another country who happens to be in a similar time zone. The other big barrier IMO is communication, so someone in Argentina who is not merely fluent in English but speaks it very clearly could be in a really strong position.