wbraun

Where are your weight figures coming from, they are not in the linked article.

Packaged semiconductors are going to be more metal interconnect / plastic encapsulation / ceramic insulation than silicon by weight.

These systems will also have a significant weight fraction in magnetic materials, either ferrite ceramics or amorphous metals.

Still a huge weight savings, but the weight fractions you are giving see off and are missing some important materials.

I was a volunteer for the rescue program in 2018. I am not sure what the exact criteria were, but dogs from the Chernobyl exclusion zone have been re-homed before.

A few dogs which otherwise would have been good candidates were too internally contaminated to re-home. Not too bad, it would have been something equivalent to a few x-rays a year if the dog slept on your bed every night. But the modal rescue dog has no radiation risk at all to worry about.

Beyond that, the workers in the exclusion zone are quite fond and protective of the stray dogs, so I think there would have been significant blowback if a significant portion of the dogs were re-homed.

As a STEM PhD student I felt a bit jealous reading this. Then I remembered that law students pay tuition while many PhD projects are funded by companies in industry, often at a somewhat dubious value proposition. There is at least a mild expectation you would consider working for the project sponsor after graduation.

There is a lot of money being thrown around towards recruitment in many fields. Tuition is worth a whole lot of free food and branded swag.

The article states that only first party data collected by NYT is used for the model. They also use no demographic data.

It seems that the only data that leaves is your previously read articles.

Amusingly, this implies that there are articles that actually increase your free article limit when accessed.

Are there different variants of DALL-E Mini? Running prompts through both this version and the one hosted on huggingface gives noticeably different results. The one on huggingface seems to give more accurate responses.

The MIT 5th year masters program for EECS requires a thesis and for you to find a faculty thesis advisor that is separate from your academic advisor.

I took some glass blowing classes in undergrad and wanted to make something with uranium glass. I bought uranium glass frit from Gaffer glass and used it, none of the instructors or students seemed to mind. The main concern was the dust when pouring out the frit, once it is melted into something its pretty inert.

I made some cool stuff with it (only limited by my glass blowing skills).

Does this behavior anger anyone else on a deep level? I get that its hard to buy GPUs right now, but this seems like such an attack on general purpose computing.

Hardware manufactures already segment features between consumer and busness grade parts that the silicon itself is capable of, such as virtualization, but restricting what algorithms one can run is a whole new level.

I am pretty sure my next GPU is going to be AMD due to this behavior by Nvidia.

The article is wrong. If you read the report its $30 - $50 million dollars, not billion.

I was a bit perplexed as to why the total could be so large....

One of my friends got jabbed by a used needle in a pair of pants he bought from Dollar a Pound. It was not a fun experience for him...

I liked the clothes I got from there though! I found a cool jacket once.

Reading through the paper, I think the real buried lead here is that the reviewers discriminated in favor of selecting black and hispanic applicants, with a larger effect size than anything analyzed but test scores and attractive vs unattractive (not neutral vs unattractive). This includes obesity, grades / class rank, research publications, and honor society membership.

At least your appearance and test scores are mutable characteristics.

Its a material that basically gives electrons "inertia" which acts like an inductance derived from magnetic fields. There is a time constant to it, which determines at what frequency range it starts contributing to the inductance value.

To make it useful for power electronics they would have to push out this time constant by at least 4 orders of magnitude, I don't believe the theoretical material properties support that.

You might see this shaving off a few mm^2 from RF ICs and providing better RF performance. Applications in power electronics, where inductors take up the most volume, seems unlikely.