nimish

I'm not sure that skipping SPARC is a bad thing. Who even has SPARC machines sitting around to validate builds? Even Fujitsu moved to ARM for their later cores.

I think a fun exercise would be for the Times to ask their own accountants how they classify their own engineering employees for the R&D tax credit.

I remember writing a little screensaver for the NDS. Got me into programming in a way

That seems like it's within striking distance of competitive, no? You get some major advantages in size and production automation. Perhaps it's ok for it to die sooner if you can get it built now and then replace it later.

We already have had that ossified layer thing multiple times in mathematics. Formalisms change. For example, prior to vector calculus writing out multidimensional PDEs was tedious. Vector calculus has serious issues in its own right, so you get people pushing geometric algebra. In more rarefied domains you have things like categories and sheafs replacing the previous "ossified" layer.

It'll keep going on and on.

Have to say marimo is excellent and is a breath of fresh air compared to Jupyter!

Over a century old! https://en.wikipedia.org/wiki/Streetlight_effect

The UK's tier 2 and the EU's blue card are strictly better than the US H-1B to green card mechanism and have been for over a decade. You face no 7% per country capping on naturalization. Has it worked out? These visas look like O-1 competitors more than anything, not H-1B.

Are there details on what visas Hyundai needed they couldn't get? L-1A/B for sending experts or management would make sense but I'm not sure there's any real issue getting those, especially not in 2023 when the plant was started.

That consumption figure is per acre of almonds. But your point is still valid. In total almonds or other crops like alfalfa consume millions of acre-feet of water a year in a dry state like California while a single data center only consumes ~500-1000.

Why would anyone develop AI models in the EU? You have a lot of compliance requirements and there's fewer enterprises that will pay for them

A half baked programming language that isn't deterministic or reproducible or guaranteed to do what you want. Worst of all worlds unless your input and output domains are tolerant to that, which most aren't. But if they are, then it's great

Interoperability is, and always was, the hardest part of programming systems together. It's telling that the ai tooling needed sustained non ai effort to expose the interfaces via MCP (or ws-* or rest or an enterprise service bus or xml or CORBA or EJB or...)

I wish I wasn't medically barred from having a pilot's license. Not for the pay, but I just like the idea of flying. Unfortunately, I cannot. I recommend people use their salaries to learn how to fly regardless! It's maybe ~$15-20k to get a PPL which is doable for the tech crowd with some planning.

At this point it seems like just learning the library is easier than trying to cram the documentation into an LLM compatible format.

Iirc it was due to a literal computer finding it funny, back when computer was a job title and not a machine.

In the integral, the h_m(s) are chebyshev polynomials of the first kind

https://en.wikipedia.org/wiki/Pulsar-based_navigation is a fun one, if not that practical

This paragraph alone is fundamentally better than the page or so of text in the blog post.

I'm not sure what it is about the style of technical writing I've seen lately but just directly getting to the point versus trying to obfuscate the thesis on a potentially controversial topic is increasingly rare

Rent seeking is the name of the game for much of b2b SaaS.

MCP is an attempt to make that easy, but the issue here is that a lot of the companies offering integration could be disintermediated entirely by LLMs. Hard to say what that means.

Hooray, finally we are getting the geometric analysis of embedding spaces we need. Information geometry and differential geometry is finally getting its moment in the sun!

Lack of wisdom and judgment. Ironic.

Pretty much.

It requires a given state to allow it to operate and have jurisdiction. That's a political act through and through.

The US doesn't recognize it anymore so it's baffling why they didn't move to non US equipment.

No, if it was just a WIP on arxiv it would actually be a good way to conduct review, like a proper PR system with tracked comments and edits and such. It would have been an example of self regulation working in academia. But it got published in a non-trivial journal!

So the silos are in expertise and review as much as data access, because Baez points out like half a dozen results that say this one is bogus. Where were those people in the review process? What happened?

There's an issue this highlights and it's not that the original authors were stupid so much as there's clearly a lot of knowledge held in silos.

That's not a good thing if your goal is to advance everyone's knowledge. Whatever is going on in academia is failing relatively closely related fields which is not good.

Yep. Trusting a guy who made a rookie excel error is not something I think we should continue to do.

Especially when the consequences downstream of that error were so stark in terms of policy.

Interesting, I guess you could port LVGL to this and get a full GUI?

This is basically it. MVP and you can ride it!

I think the biggest issue with CAHSR is that Florida just did it anyway.

I'm happy to be the next AMD CEO to try out!