subharmonicon

Hotels in Dallas for at least 15 years have disuaded people from walking even a few blocks downtown because they equate all the homelessness with crime.

I found this funny because by far the biggest danger I have seen there are endless electric scooters littering the sidewalks.

Second strongest I remember in my 16 years here, with the 2014 Napa Quake being notably more shaking.

The blog post is about using an NVIDIA-specific tensor core API that they have built to get good performance.

Modular has been pushing the notion that they are building technology that allows writing HW-vendor neutral solutions so that users can break free of NVIDIA's hold on high performance kernels.

From their own writing:

We want a unified, programmable system (one small binary!) that can scale across architectures from multiple vendors—while providing industry-leading performance on the most widely used GPUs (and CPUs).

Saw an exhibit with some of her work, I think in Albuquerque. Was surprised/delighted to see weavings of circuits.

I’ve also been curious to see actual users compare/contrast their experiences with other options, but so far haven’t seen that.

There seem to be enthusiasts who have experimented a bit and like what they see but I haven’t seen much else.

TLDR: In order to get good performance you need to use vendor-specific extensions that result in the same lock-in Modular has been claiming they will enable you to avoid.

The market tends to be pretty efficient for things like these. We’ve seen significant rapid adoption of several different ML solutions over the last decade, yet Mojo languishes. I think that’s a clear sign they aren’t solving the real-world pain points that users are hitting, and are building a rather niche solution that only appeals to a small number of people, no matter how good their execution may be.

Cancelled my membership many years ago over their refusal to support open access.

Windows machine with NVIDIA CPU & 5070 class GPU?

Love this paper and read it several times, most recently around 10 years ago when thinking about whether there were looping constructs missing from popular programming languages.

I have made the same point several times online and in person that the famous quote is misunderstood and often suggest people take the time to go back to the source and read it since it’s a wonderful read.

I recall the original post about Lungy.

Having had an incentive spirometer prescribed for post-surgical use after being on bypass, my experience was that it seemed boring and like a waste of time, so anything that makes breathing exercises more engaging and feel more worthwhile is a win.

Off topic, but the typesetting here looks top notch and I am curious if anyone can elaborate on the tooling used to render this for the web?

My dad grew up in 1930’s Detroit and would tell us about how when he was five his mother would give him a little money and have him walk a few blocks to the baker to get a loaf of bread.

I grew up in the 70’s where after maybe eight we had pretty much free rein. I rode my bike several blocks away and crossed two busy four lane roads to get candy at 7-11.

The world is no more dangerous now than it was then, and yet here we are, with parents being treated like criminals for letting their kids play in their own front yard.

I’ve spent 30 years working on compilers.

They have bugs. Lots of them.

With that in mind, the article is correct that the vast majority of issues people think might be a compiler bug are in fact user errors and misunderstanding.

My experience actually working with users has been somewhat humorous in the past, including multiple instances of people completely freaking out when they report something that turns out to be a miscompile. I’ve seen people completely freaking out, to the point that they no longer felt that any code could be trusted since it could have been miscompiled in some way.

But it is absolutely true that the federal government is endlessly wasteful; it’s insane watching everyone around me get gaslit into thinking the government is actually efficient.

It's also absolutely true that private companies are endlessly wasteful. I've worked for four large companies now, and the waste is mind boggling. I think what upsets people about government is that tax dollars are used to fund it, but I would claim from my own observation that beyond a certain size any kind of organization is filled with waste.

I'll point out, though, that there are areas of government that have been studied and found to be very efficient, and have high levels of satisfaction. It's been quite a while now, but I recall around ~2006-2007 an academic study came out which was originally intended to look only at private insurers. As they designed the study they realized that given the size of Medicare they should also include Medicare as part of the study.

What they found:

- A much much higher percentage of the money going into Medicare goes toward patient care than in any of the private insurers. Like low single digit percentages of overhead vs. 10-25% overhead in the case of the private insurers.

- Customer satisfaction from dealing with the bureaucracy (claim processing) of Medicare was much higher than customer satisfaction with the private insurers.

- Patient satisfaction with the care they were receiving from Medicare was as high or higher than the private insurers.

Making fewer pennies means needing more nickels. Which have an even larger loss per unit: https://www.cnn.com/2025/02/10/business/cost-to-make-penny-n...

Why did the select subcommittee not report that Fauci lied if in fact he did?

Why did Fauci raise the possibility of a lab leak on Feb 1, 2020, if he was trying to cover that possibility up?

Time and time again we see conspiratorial claims with nothing to back them up.

Thanks!

I have known about slang for some time but hadn’t heard about these commercial products.

Can you elaborate?

I’m honestly a bit surprised if that’s the first time you’ve seen something like that.

I’ve been working on compilers 30 years, primarily on optimizers (although some frontend work as well). Learned C from Borland C and C++ from…Zortech C++ 1.0, so thank you for that!

Within a few years of working on compilers I came across my first examples of 50k+ line functions. These were often (but not always) the result of source code generators that were translating some kind of problem description to code. It taught me very early on that you really need to focus on scalability and compile time in compilers, whether it’s the amount of code within a function, or across functions (for IPO / LTO).

And yes, working on compilers is never dull. 25 years ago I thought we’d end up in a monolithic world with x86, C++, and Java being the focus of all work. Instead, there’s been an absolute explosion of programming models, languages, and architectures, as well as entirely new problem spaces like graph compilers for ML.

The latest standard is from 2019: https://ieeexplore.ieee.org/document/8766229

There is still some flexibility in implementation, for example how and whether FMAs are formed for a given compiler when FP_CONTRACT is ON, and in the standard itself in things like when tininess is detected.

But to your point if you stick to the basic operations in the standard, and don’t enable FP_CONTRACT and FENV_ACCESS in C/C++, have a bug free compiler and don’t use fast-math, you’re good to go.

[edit to add a caveat about compile-time constant folding which is a whole can of worms]

[edit again to point out that the C/C++ standards allow for implementations to compute intermediate results at higher precision, so a compliant implementation can use all 80 bits on x87 when computing expressions]

Fight Club.

It came out at a time in my life when I was struggling to avoid the pull of consumerism and had a strong sense of isolation in the modern world. I used to watch this quite often.

When it came out I was also at a point in my life when I was interested in screenwriting and was watching sometimes as many as eight movies a week, in part to study screenplays and stories. I would sometimes watch this a couple times a month during that time.

The “Personal Voice” feature on iOS is a fantastic way to capture your own voice in case you are eventually in a position that you cannot speak naturally.

I set it up when it first came out. It’s imperfect, but very very good.

Specifically a lot of younger people are using cannabis, rather than alcohol, so alcohol use has declined in those populations.

The article doesn’t even attempt to explain this, but this is talking about Carmel-by-the-Sea, a small part of the greater Carmel area.

I’ve not seen any street addresses, just house names, in my few dozen trips where I have walked through Carmel-by-the-Sea.

Can you point to any Tegra data showing that it’s actually competitive with Apple?

I just searched and don’t see that.

I’m totally on board with what you’re saying.

Unfortunately a lot of tech is far too biased towards a small set of learnable skills that are sometimes hard to demonstrate in an interview setting, and the result is that a lot of good people don’t get offers for jobs that they could totally crush in.

In the 90’s when I was at Microsoft it was common for people to ask brain teasers and algorithm questions and expect people to be able to reason through the problems even if they walked in not knowing the particular algorithm or data structures involved. The interview was graded more on your thought process and ability to make forward progress with hints and less on getting to a correct answer in 25 minutes.

People were super harsh about attention to detail, e.g. you had to write code that would compile on the whiteboard and you would get penalized to some degree for making mistakes, especially if there were enough to make it clear that you hadn’t really been using the language you claimed to be proficient in.

I don’t think NVIDIA stock is a bubble, but I do think AI hype is overheated and unless we start seeing companies turning substantial profits from all the NVIDIA hardware they’ve purchased, the purchases will slow down.

To that point, what percentage of their data center revenue came from just Meta, Microsoft, Amazon, and Google this past year? And will those companies, who can afford to drop 10s of billions on hardware, refresh everything with Blackwell?

If enough inference moves on-device or to other hardware platforms, I do think NVIDIAs growth will slow substantially, and sales may even drop substantially over the course of a couple years.

For years I would eat either a half or whole avocado a day. Why would you need to be a vegan/vegetarian to do that?

They are fiber and potassium rich with a lot of unsaturated fat, which adds to making them a great “dessert” if you eat them with a bit of olive oil, salt, and lemon.