Re: Not late.
There is a slice of up called the "ignorasphere" that is between what a balloon can do and the lowest orbit. There are plenty of researchers that are looking to put experiments into that realm.
That's what sounding rockets are for. New Shephard may be able to do it cheaper, which would be good. But it's not inaccessible territory and there are plenty of other people in a position to do it cheaply as well with the burgeoning small-sat launcher market (Rocket Labs, Pegasus, etc, etc).
There is also a metric called Technical Readiness Level (TRL) that gauges the progress of something intended to be sent into space. A lower altitude rocket can be useful for testing some items to raise their TRL before the are fielded or tested further.
Like SpaceX did with Grasshopper. Years ago. They learnt what they needed and moved on. Evidently BO think they can make some cash repurposing it into a fairground ride.
BO may fly people sooner than SpaceX.
To orbit? No. SpaceX will be flying people this year. Even with Elon time, Crew Dragon is happening this year or next. New Glenn's first flight isn't slated until 2021. If it doesn't slip. Which is actually quite likely, because to this point, BO have zero expertise in ground-handling or launch operations at Florida. They've sent NS up and down from Texas, but playing with the big boys putting up orbital-class vehicles with all the attendant paperwork and processes? That's a different ballgame.
SpaceX ran into issues just getting their Transport-Erector reconfigured from F9 to FH - what makes you think BO will have it all smooth sailing getting Complex 36 up and running?
Elon doesn't have any credentials in material science or mechanical engineering and he's lecturing ex-NASA engineers about materials for spacecraft.
He literally has some idea. He taught himself an unofficial Masters in Rocket Science getting F1 off the ground. If he didn't know an answer he went and found someone who did - ask Tim Worstall.
At the end of day, it's two ways of doing things.
SpaceX have developed a minimum viable product and built a profit-making business whilst they refine the vehicle to human-rated standards. It's a software-based move-fast-and-break-things way of doing it, and it's impossible to say it hasn't worked for them.
Blue Origin have chosen to rely almost entirely on Bezos selling a billion dollars a year in Amazon stocks to bankroll the development of his "final" rocket. It's a very old-space way of doing things, but seems to be working for them (though they have some side-income now that Vulcan is planning on using their BE-4 engines).
NASA isn't happy with their "put the astronauts on-board and they pump the fuel" approach since their last explosion happened while fueling.
1. That has been well-fettled. They're happy with the current process. It's also been well established that this is a moronic way of thinking for capsule-based vehicles. Maybe it made sense for the Shuttle where you couldn't eject/abort quickly - you had to unstrap, climb out and get into a zip-wire basket - it made sense to get it fuelled, stabilised and then load the crew. For a vehicle carrying a simple crew capsule with a launch-abort mechanism, why would you want humans walking around the vicinity of a loaded, fuelled rocket? Either way carries risks to the flight-crew, but Load-and-Go eliminates all risk to the ground-crew.
2. It wasn't an explosion, it was a fast fire. That makes a difference, because a Crew Dragon would have had time to escape (in the style of the recent Soyuz abort).
3. For AMOS-6 they were trying out a new loading procedure. Whilst we can call into question the wisdom/stupidity of doing that with a customer payload on top, the fact is that they're not going to go off-piste on crewed launches and will be sticking to the manual.
4. Falcon 9 is now 66/68 fully successful launches. That's a 97.1% success rate - which means they've overtaken Ariane 5 (widely considered to be extremely reliable) on 97/102 (95.1%). If we narrow that to the current version - F9 Block 5 vs. Ariane 5 ECA then it's 11/11 (100%) vs 67/69 (97%).
A smaller sample, but given that the first B5 flight was less than a year ago, a huge achievement nonetheless.