"AWS" and it's "GovCloud" and other related cloud partitions are very far ahead of Azure in many regards. They are the market leader. It's hard to make a strong comparison, but AWS is likely 5x-7x bigger than Azure, possibly closer to 10x in terms of revenue and resources. There seems to be some support in the record that AWS was the favorite and that Microsoft winning was seen as something of an upset. But that's not fact (Yet).

> That presumes the people writing and testing the code work for free. And even if they work for pennies on the other side of the world, the CXO "managing" them doesn't come cheap either. That's what "marginal cost" means. After you invest in the initial development, the additional per unit cost is very small. > But more to the point...industrial cameras ain't cheap. The cameras in the front and back of cars with all these gizmos (color, industrial spec temperature range, etc) probably add a couple hundred to the cost of the car. The SoC that does the video processing is probably close to a hundred if the cost breakdown of comparable parts in smartphones is a guide. The median cost of a new car is $41,000. If the electronic vision systems and related electronics cost $400 as you suggest, it's about 1% of the total cost of the vehicle. The backup system saves lives and injuries - in the ballpark of 1,000 accidents and 50-100 fatalities per year (Somewhat old data, but seems reasonable: https://www.washingtonpost.com...).

There is nothing wrong with scientific nomenclature, but the problem with AI is it hasn't been trained to understand the long history of the language that it is emulating. Calling black men, for example, primates or other animal related names is part of the deeply racist history of the language and those who use it. AI doesn't know that. AI knows what it sees, and what it's training model has allowed it to digest and use to predict future events.

It is a little bit of oversimplification to say that it's the drivers fault that they don't understand the differences in the system when Tesla's people are out there flogging the capabilities of the system, calling it things which imply it's more capable than it is, and generally overpromising every single chance they get. You can't put on shocked Pikachu face when drivers parrot back Tesla's hyped promises about the quality of their driver-assist systems. I agree however that software, marketing, or driver traning short-comings are not grounds for a recall. They are grounds for regulation and/or restrictions on some aspects of the product offering.

No way. The rule is that employers can't prevent employees from organizing for better working conditions. This isn't 1910. Employees have the right to organize and press for better collective working conditions. Remote work is a perfect example of that.

Everyone - all big enterprises - use Slack. I've heard rumors that inside Microsoft some teams use Slack. It's the new corporate death star running and taking 2.5GB of RAM at all times.

You can't prevent people from organizing or discussing improvements to their collective work environment or rights - either at work or privately. And you can't constructively create rules that have that effect. So shutting down "off topic" discussion rooms is a fairly naked attempt to constructively ban organizing for better working conditions. Google was sued by the NRLB over this, after they cracked down on internal dissent, and they were forced to settle and publish specific guidelines and employee rights: https://www.cnbc.com/2019/09/1.... Essentially, NRLB has taken the "watercooler or lunch room" rules - if you were employer, and people were organizing and agitating for better working conditions in the company lunch room, the company can't just shutdown the lunch room in response - and applied it to digital spaces. Which seems entirely fair and reasonable.