mohaine

Well, they kind of suck these days, at least compared to how we remember they used to taste.

Who know if they are actually the same, plain genetic drift, or they have been selected to be more shelf stable at the cost of taste.

Well, we'll know to worry/celebrate when they start laying of lawyers.

The software side can probably go many years without new features. Nobody picks Oracle in 2026 for the features.

I would be more worried about the weird white stuff squeezing out of the lower right of the screen.

1. NAT and a firewall are 2 different things 2. With IPV6 you can have so many IPS that unwanted users can't guess your IP. This isn't true security but see 1 for that.

pretty sure it is a way of fingerprinting your machine/network

Agreed. I’m on my 4th Garmin and the charging cable last about a year or two at most. That said, for me it has always just been the cable and the watch is self last forever. All 4 of my old Garmin’s still work if I get them out of a drawer and charge them up. Now I just small USBC -> Garmin adapter's that normally get lost before they wear out.

I would love a standards based wireless option here.

I'm pretty sure the software side of the project is a direct descendent from the WRT54G stack.

LinkSys got sued to release the firmware as it was GPL linked. This dump got modified to make the WRT54G way more powerful than LinkSys ever planned but they got to sell the hardware for years more than would have been expected at the time.

I love by Dream Machine Pro. Seems to just work and keep everything up to date. I have it running my security cameras as well and it has been pretty much bullet proof.

What needs do you have for a router that the Cloud Gateway is missing or is bad at? A PiHole equivalent is about all I can think I'm missing.

Pretty sure phased array LASERs are not yet a thing.

To bad they used red and green that look exactly the same to me

It seems like releasing more would have probably broken the exact same rules they are claiming AdaFruit broke.

Ah, most the problem in programming is writing the tests. Once you know what you need the rest is just typing.

I can see an argument where you can get none programers to create the input and output of said tests but if the can do that, they are basically programmers.

This is of course leaving aside that half the stated use cases I hear for AI are that it can 'write the tests for you'. If it is writing the code and the tests it is pointless.

But data science usually isn't an island.

Use whatever you want on your one off personal projects but use something more non-data science friendly if you ever want your model to run directly in a production workflow.

Productionizing R models is quite painful. The normal way is to just rewrite it not in R.

Been awhile since I've worked at one but it is usually grounded in trying to achieve 100% MS usage.

It is rarish to find a partial MS shop. Most of this is how hard MS makes it to use other tools. Even in 2025 they have good interop with external tools hamstrung.

Example: SQL Servers JDBC driver will convert an entire table's of data from ASCII to UTF and a full table scan instead of convertering your UTF bind to ASCII and using the ASCII based index. This doesn't break interop but does make it painful to code and one more reason to just use .Net.