I'm curious why Microsoft has that setting on by default. What are the benefits of letting any peripheral make system level changes to that extent without popping up a user installer?

Users don’t have to laboriously track down the drivers for a new peripheral and manually install them. It enables almost any device to become plug and play. But it needs guard rails.

One of the primary purposes of batching is to release features at a regular cadence, and all releases go through the same minimum set of checks to catch any unexpected regressions.

In my experience, the workload of these checks grows much slower than linearly with batch size.

What other factors am I missing here?

I think the article is mostly just confusing throughput and latency. Releasing on a fixed schedule adds some latency, but that doesn't mean there's a bottleneck constraining throughput. And long-term, development speed is determined by throughput, not whether you shipped a feature a few days earlier.

High latency might become a problem if you need user feedback to decide what to work on next, but you don't necessarily need to push everything to production immediately to get feedback. You can have a separate beta/staging system and ask users to check out your new changes there before they get rolled out to everyone.

But you most likely signed a binding arbitration clause in the TOS

Same result and then Microsoft would be paying for arbitration

Did I sign anything? EULAs aren’t necessarily 100% enforceable.

What would a reasonable person expect out of the transaction?

Which promises and verbiage in the marketing material contradict the EULA?

Slightly misleading title: It's an Artificial Eggshell/Membrane.

They took regular eggs, broke the shell/membrane and emptied the contents into a silicone based eggshell/membrane for incubation.

Still good science, but greatly tempered implications. Allegedly a group of Japanese school students did this previously using cellophane wrap.

The article repeatedly uses the API charge to the 'customer' as an estimate of the cost value to Anthropic/OpenAI.

Is there any good data on the true cost of inference for these companies?

The mouse sensitivity on mobile though is wildly high!

Not sure I fully understood what you meant by 'the open category is done'?

Also, from the categories that you mentioned, do you compete in West Coast Swing?

Reverse the benefits != increase the net risk

It took a while going through the data in the results section to see this.

They actually do compare against a control group. This is the study that is being referenced.

https://bmjmedicine.bmj.com/content/5/1/e002150

The data on the results section shows almost parity between the control group and participants who discontinued for 2 years.

Note that while it is a well conducted study at the US VA with 300,000+ patients, it is not a randomized study so fully eliminating confounding variables and reverse causality is hard.

This is the study that the article references. (The linked one on the post is an older study)

https://bmjmedicine.bmj.com/content/5/1/e002150