aewens

It’s also been harder than ever for scientists in the non-AI space to get more HPC compute with AI data centers causing pricing of components to skyrocket and constraining supply from manufacturers. So there will certainly be a boon in the HPC space one way or the other when the AI bubble pops if pricing and supply returns closer to pre-AI norms.

Amusingly, some of the earliest AI research was using Lisp which beget AI winter. Now we’ve come full circle with LLMs that struggles to write valid Lisp. Almost poetic.

Just be prepared near the end for refinements. When I did mine, I had to do another 6 months of refinements near the end to help get the locking of my bite grip to be where it needs to be. I was otherwise expecting that day to be when I’d be done, and later learned refinements are a fairly common occurrence when going through Invisalign.

As the sibling comment, it’s definitely worth it. Best of luck!

Not who you asked, but I think making the nuance between retail and corporate credit. With firms being corporate credit (i.e. we aren’t talking about individuals / retail).

My brother’s house is in the same boat. Unfortunately, not all ZIP codes exclusively belong to one city.

The way Git computes diffs is by more or less storing all the source code in the .git directory as objects. At first glance it looks like a bunch of hashes, but tools can pull out source code from the objects tracked within the .git directory. Not least of which, the remote URL points to their username on GitHub and the author for commits can give you their email.

Lossy compression vs lossless compression is the difference of whether you can get a 1:1 copy of the original data if you compress and then decompress it.

A simple example of this is if you have 4 bits of data and have a compression algorithm that turns it into 2 bits of data. If your dataset only contains 0000, 0011, 1100, and 1111; then this can technically be considered lossless compression because we can always reconstruct the exact original data (e.g. 0011 compresses to 01 and can decompress back to 0011, 1100 compresses to 10 and can decompress back to 1100, etc). However, if our dataset later included 1101 and got compressed to 10, this is now “lossy” because it would decompress to 1100, that last bit was “lost”.

An LLM is lossy compression because it lacks the capacity to 1:1 replicate all its input data 100% of the time. It can get quite close in some cases, sure, but it is not perfect every time. So it is considered “lossy”.

Probably since you can embed it in other languages quite trivially. For instance, in C after about a dozen lines of code you can now pass around data into Lua and back to C and thus give you access to a scripting language with little fuss. It’s also a fairly small and simple language, so adding it in won’t add much more to the overall footprint of the project.

To play the devil's advocate:

IANAL, but if I recall correctly part of having a trademark is having a legal obligation to enforce it, and if I'm reading the tweet correctly this is a name of a company and typically one cannot use a trademark in their company name.

An analogy may help, imagine the website as a door. A website using HTTP is a normal door and using HTTPS is a door with a lock, where the keyring in this analogy are the trusted CAs by your browser. A website using HTTPS with an expired certificate is a door that should have a lock, but the lock no longer latches; and a self-signed certificate is a locked door with a key left in the doorknob.

From a security perspective, a door without a lock has no expectation of protecting anything. But a door that should lock but doesn’t, or is supposed to be locked but has the key left in the latch is not providing the security expected, and should be given pause when anticipating security from the lock. This is what the browser is trying to translate with its UI.