The proponents of no estimates, particularly Vasco, Neil and Woody use a straw man argument saying estimates are just wild ass guesses.
This WAG approach is just *bad* estimating, but due to the prevalence of this approach and the community's illiteracy on estimating approaches it gets attention. Estimates are useful and have their place.
What the No Estimates crew are arguing for behind their shock and awe hastag is a contextual, pragmatic and skilled approach to estimating.
Consider these scenarios and how estimates apply.
Scenario 1: control the environment; have a clear strategic imperative, make sensible, small clear steps forward. In this scenario estimates are not required. Just do the right thing.
Scenario 2: it's complex and there are uncertainties and you need to prioritise based on cost as well as value; use evidence based methods to determine a likely cost. Make small moves to test your theory and be ready to change plan.
Scenario 3: this is completely novel. We don't even know if this will work; Try something. Keep it in the frame of an experiment. Iterate on knowledge acquisition; you can't know what the outcome or effort will be so just keep the cycles small. Know when to quit or commit.
In scenario 2 Story Points may be valid, but they are as much guessing as time estimates.
Techniques like planning poker help normalise the guesses and consider past experience so this method helps.
Using past cycle time performance, normalising story sizes, reference models (this is like that) and paying attention to data are all important methods that will make things more reliable.
Know when estimates matter and when they don't.
Utilise the skills you can easily access to perform estimates well. Don't act like an amateur.