The Waymo cars are really comfortable luxury Jaguars. For Uber and Lyft there are many price tiers, but to reliably get an equally or more comfortable car you probably need to book the black car options. I’m sure Uber / Lyft are way more expensive per mile than Waymo on that tier.
In addition to all the things people have pointed out that makes it a better experience.
Almost every Uber Black and Black SUV I’ve ordered was a Chevy Suburban or GMC Yukon.
The quality is across the board, but one thing I’ve found consistent is the terrible quality seats. The seats feel like it’s just cardboard supporting you that pops in and out as you move with the car.
It’s rare to get an actual luxury car even when paying more.
Their promise of “professional” drivers is also wild. Sometimes you get a guy who’s friendly and seems eager to please and helpful with luggage, but I’ve had plenty of downright rude drivers who feel inconvenienced by my presence.
> I’ve had plenty of downright rude drivers who feel inconvenienced by my presence
This is my general observation about life (at least in the US) these days: the seeming prevalence of people who think they're doing you a favor by doing their job.
Mostly a blog about data analysis in a niche field of biology.
In the very beginning it was a static website built in Hyde. Then after learning I could use MathJax in Tumblr I converted it to a Tumblr site with a custom URL.
Eventually after getting frustrated with squeezing things into Tumblr I got a Squarespace account and converted all the content to that.
Mostly I'm happy with Squarespace, though it involves a lot of manual typesetting to work around various unclear escaping that happens in the Markdown parsing. Still better than the hassle of hosting something custom for me.
Thanks for the write up. We're taking a similar approach at my startup to classify business reviews. The breakthrough for us came when we split the reviews into sentences and did N-gram analysis at the sentence level. The challenge is that the most significant N-grams (e.g. N > 2) have such low frequency that there isn't much data to train on. Our current approach is to try to coax patterns out of the N-grams (e.g. "salesman was rude" and "manager was mean" become "[employee]=[negative]"). I do like the top 5 approach, and I think I'll see if I can work that into our approach.
The 2 or 1 lanes are separated by a metal fence. You can only overtake at the stretches of road where your direction have 2 lanes on one side of the fence.
They did feel a bit terrifying and constrained when we first got them. But now we're used to it.
I did nearly have an accident once when I tried to overtake a truck too close to the road switching from 1+2 to 2+1. Though that was due to me being really really stupid.
