rrss

By your criteria, the Central Valley was never a great climate for growing things. A century ago there was not enough water in the San Joaquin valley, and people were talking about getting fresh water from hundreds of miles away.

I assume your opinion is probably that the water projects in California were all mistakes, but it’s not accurate to say that at some point the Central Valley was naturally ideal for growing things without artificial irrigation from far away.

but “does not require artificial irrigation” was not what I meant by “ideal climate.”

There should be a page history with a comment describing why changes were made. The latest revision can describe the current best knowledge “without fluff”, and the revision history can show why and how the content changed over time.

The CDC already has a list of when updates were made to all the pages, but AFAICT it doesn’t say what changed or why.

https://www.cdc.gov/coronavirus/2019-ncov/whats-new-all.html

P.S. this part about needing irrigation is not unique to the Central Valley.

Irrigation agriculture, rather than depending on natural water for growing, was an innovation humanity developed 6000+ years ago.

No. More water arriving directly in the Central Valley would make it worse for agriculture, not better. Rain or snow are bad for crops. Controlled irrigation using water from somewhere else plus the mediterranean climate combine to make the Central Valley very productive.

long growing season

sun

good soil, almost all capability class 1

ideal mediterranean climate

what is the relevance of needing to have a lot of pumps? Do you think the engineering involved is beyond the capability of the US, and your argument now is that it cannot be done?

The idea that californias water projects only direct water downhill is invalid anyway - the California aqueduct flows uphill quite a lot, and over mountains.

https://en.m.wikipedia.org/wiki/Edmonston_Pumping_Plant is one of at least 5 pumping plants along the aqueduct.

The Central Valley is not a desert, I do not like this meme.

Farmers in California were running out of water 100 years ago too, and people responded with massive engineering projects to regulate water throughout the year and move it from places that were at the time water-rich to places where more water would allow using the extremely fertile land with favorable weather for agriculture.

It’s not insane to consider doing the same again.

Stallman’s hard line is pretty incoherent / absurd when it comes to firmware.

The FSF does not care about proprietary firmware, provided that the firmware “installation is not intended after the user obtains the product.”

https://ryf.fsf.org/about/criteria

So if you have a product that needs a ton of proprietary software to run, RMS will go berserk if you run that software on the CPU, but if you shift it to run on another processor in the product and make sure users can’t change it then all is good, have a RYF certification. This does not make sense. Maybe it made sense 30 years ago.

If you’re stallman, proprietary firmware isn’t bad so long as it’s suitably hidden so it can’t offend your sensibilities. So pretend CPU microcode updates don’t exist and add secondary processors to deal with blobs (purism). Does ensuring users cannot upgrade firmware make them more free?

This does not explain why the IRS needs $80 billion.

Do each of these problems take billions of dollars to fix?

$1B ~= 1000 engineers working for 10 years.

$1B ~= buying 1.6 ORNL frontier supercomputers.

Related March 2021 preprint: https://arxiv.org/abs/2103.03074

Jan 2022 paper: https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.12...

There are still a ton of systems which have firmware that does not support large or resizable BAR. Support only became common in the last year or two (after AMD marketed as SAM and pushed to get it supported).

So requiring the “easy default” to be available means prohibiting any one with a system from ~2020 or earlier from upgrading to these GPUs.

Refer to “Practical uses of synchronized clocks in distributed systems” for some info on algorithms like this.

https://dl.acm.org/doi/10.1145/112600.112601

Google spanner uses stuff like this with a clock called TrueTime. I don’t know how leap seconds / smearing is handled in that system.

The charge controller is surprisingly large (to me) relative to the rest of the system. I don’t have much experience - is there a reason it needs to be, or is the charge controller significantly overspec’d for this 50W panel?

BOM (linked from github page): https://docs.google.com/spreadsheets/d/1X0RDVBC6ifIS6Qov1loz...

panel short-circuit current is 3 A, but is paired with a 20A charge controller.

No.

https://scholarship.law.cornell.edu/cgi/viewcontent.cgi?http...

also, the waypoint-based restart protection used in AGC software to handle power failures and the decision to reuse that mechanism for handling running out of core sets or VAC areas was due to charley muntz, not hal laning.

But regardless I agree Laning deserves more recognition for his numerous contributions.

I don’t think it’s a particularly accurate or fair simplification to say that overload during Apollo 11 was due to operator error.

The rendezvous radar was intended to be enabled during landing to be prepared for an abort. It only caused problems due to a hardware bug in the interface between the radar and the LGC which caused something similar to an interrupt storm. Grumman and MIT had both discovered this bug prior to Apollo 11 but decided not to fix it.

source: various publications by don eyles

> Apple Coalition of Organized Retail Employees – AppleCore – that will be part of the International Association of Machinists and Aerospace Worker

Why are apple store employees going to be in a machinists union?

My primary laptop is a 2011 MacBook Pro, running Ubuntu most of the time.

I added RAM (to 8 GB) in 2013 or so and swapped the hard drive out for an SSD in 2015.

I have been contemplating an upgrade for about 7 years but haven’t been able to decide on something :)

How does one distinguish a build quality that is 100x better from a build quality that is only 10x better?

I think the point is that it seems silly to say “two orders or magnitude” when there’s no quantitative metric. Is it more informative than “way better”

I’d like a job where I used more math - can you share your job & industry?

> The parent comment linked to a tweet where he says everyone else working

https://twitter.com/elonmusk/status/1531867103854317568

> hey elon a lot of people are talking about this leaked email, any additional comment to people who think coming into work is an antiquated concept?

> They should pretend to work somewhere else

I think it’s a reasonable interpretation that Musk is saying people working remotely are only pretending to work. For people who believe they have been working hard remotely, IMO it’s easy to see why this could be viewed as insulting.

consider an engineer hired at Tesla in 2019.

They probably have a 4 year RSU schedule determined based on the TSLA share price in 2019, so $50.

If their equity compensation target was $30k per year, then they would get stock grants in 2021 and 2022 worth $400-600k.

That works out to a lot of millionaires.

Many reasonable CS degrees have no electronics or hardware in the curriculum.

Last one was mid march, so every few months. The current plan (last I checked) is 2 more in this google-sponsored free program.

One supplier: skywater.

In 1990, Cypress Semiconductor acquired a Control Data fab, which became Cypress Fab 4. At some point they developed a 130nm process. In 2017, cypress sold the fab to skywater, which was formed specifically to operate this fab. Skywater, google, and efabless collaborated to publish the PDK for the 130nm process.

“Old” processes like this have lots of uses, even before the recent supply chain issues (which were/are often worse for mature processes than more leading edge stuff). They are commonly used in automotive / embedded stuff, and the US government / military require US-based semiconductor fabrication for some contracts.

Metal is apple’s api for writing software that uses their GPUs.

https://en.m.wikipedia.org/wiki/Metal_(API)

Speaking of drivers on the GPU: https://github.com/enfiskutensykkel/ssd-gpu-dma

> Build userspace NVMe drivers and storage applications with CUDA support

> libnvm can be linked with CUDA programs, enabling high-performance storage access directly from your CUDA kernels. This is achieved by placing IO queues and data buffers directly in GPU memory, eliminating the need to involve the CPU in the IO path entirely.

No, just because general relativity and quantum field theory use linear algebra does not mean they are linear algebra.

When you say the “whole imaging stack,” which one do you mean?

EHT imaged M87 with four teams working independently using different methods. Each of those teams used algorithms that were published and validated before the EHT results were published.

> Scientific outreach has a real “draw the rest of the owl” problem

It seems like this is just the “you need to study the problem and it’s dependencies for many years to understand the physics and mathematical details” problem. Which does not actually seem like much of a problem to be solved, just a reality.

Your example in the other comment depends on linear algebra, so ~first year undergraduate of most technical fields. IMO that’s fundamentally more accessible than the black hole information paradox, with its dependencies on general relativity and quantum field theory.

I’m sure more technical videos and posts could be made about this, but how small would the target audience be?